WEBVTT Kind: captions Language: en 00:00:01.070 --> 00:00:14.949 align:start position:0% [Applause] 00:00:14.949 --> 00:00:14.959 align:start position:0% 00:00:14.959 --> 00:00:17.830 align:start position:0% so<00:00:15.959> today<00:00:16.860> I'm<00:00:17.010> here<00:00:17.039> to<00:00:17.250> talk<00:00:17.550> about 00:00:17.830 --> 00:00:17.840 align:start position:0% so today I'm here to talk about 00:00:17.840 --> 00:00:21.800 align:start position:0% so today I'm here to talk about validating<00:00:19.610> verticals<00:00:20.610> and<00:00:20.849> then<00:00:21.150> a<00:00:21.180> neat<00:00:21.390> one 00:00:21.800 --> 00:00:21.810 align:start position:0% validating verticals and then a neat one 00:00:21.810 --> 00:00:24.830 align:start position:0% validating verticals and then a neat one and<00:00:22.250> when<00:00:23.250> they<00:00:23.400> need<00:00:23.520> to<00:00:23.580> level<00:00:23.939> using<00:00:24.300> AI 00:00:24.830 --> 00:00:24.840 align:start position:0% and when they need to level using AI 00:00:24.840 --> 00:00:30.650 align:start position:0% and when they need to level using AI agents<00:00:26.539> so<00:00:27.539> we<00:00:28.289> focused<00:00:28.710> on<00:00:28.890> the<00:00:29.189> ant<00:00:29.699> no 00:00:30.650 --> 00:00:30.660 align:start position:0% agents so we focused on the ant no 00:00:30.660 --> 00:00:32.390 align:start position:0% agents so we focused on the ant no matter<00:00:30.869> what<00:00:31.289> I'm<00:00:31.500> learning<00:00:31.590> and<00:00:31.949> you<00:00:32.220> may<00:00:32.340> be 00:00:32.390 --> 00:00:32.400 align:start position:0% matter what I'm learning and you may be 00:00:32.400 --> 00:00:36.170 align:start position:0% matter what I'm learning and you may be wondering<00:00:32.699> why<00:00:33.690> are<00:00:34.170> we<00:00:34.440> looking<00:00:34.530> at<00:00:34.769> this<00:00:35.180> how 00:00:36.170 --> 00:00:36.180 align:start position:0% wondering why are we looking at this how 00:00:36.180 --> 00:00:36.920 align:start position:0% wondering why are we looking at this how does<00:00:36.239> it<00:00:36.420> fit 00:00:36.920 --> 00:00:36.930 align:start position:0% does it fit 00:00:36.930 --> 00:00:40.490 align:start position:0% does it fit I<00:00:37.070> am<00:00:38.070> Shila<00:00:38.489> me<00:00:38.730> and<00:00:38.940> I'm<00:00:39.900> gonna<00:00:40.019> explain<00:00:40.320> all 00:00:40.490 --> 00:00:40.500 align:start position:0% I am Shila me and I'm gonna explain all 00:00:40.500 --> 00:00:47.240 align:start position:0% I am Shila me and I'm gonna explain all of<00:00:40.620> us<00:00:40.769> and<00:00:41.070> talk<00:00:41.610> you<00:00:41.790> through<00:00:41.970> that<00:00:45.829> so<00:00:46.829> in 00:00:47.240 --> 00:00:47.250 align:start position:0% of us and talk you through that so in 00:00:47.250 --> 00:00:48.410 align:start position:0% of us and talk you through that so in the<00:00:47.309> past<00:00:47.520> we've<00:00:47.760> looked<00:00:47.969> at<00:00:48.090> something 00:00:48.410 --> 00:00:48.420 align:start position:0% the past we've looked at something 00:00:48.420 --> 00:00:50.420 align:start position:0% the past we've looked at something called<00:00:48.570> selfish<00:00:49.140> money<00:00:49.379> which<00:00:49.680> is<00:00:49.860> a<00:00:49.890> strategy 00:00:50.420 --> 00:00:50.430 align:start position:0% called selfish money which is a strategy 00:00:50.430 --> 00:00:53.590 align:start position:0% called selfish money which is a strategy attack<00:00:50.940> on<00:00:51.320> IBM<00:00:52.320> through<00:00:52.530> one<00:00:53.070> boy<00:00:53.250> next 00:00:53.590 --> 00:00:53.600 align:start position:0% attack on IBM through one boy next 00:00:53.600 --> 00:00:56.630 align:start position:0% attack on IBM through one boy next vertical<00:00:54.600> where<00:00:55.350> essentially<00:00:56.129> you're<00:00:56.370> trying 00:00:56.630 --> 00:00:56.640 align:start position:0% vertical where essentially you're trying 00:00:56.640 --> 00:00:59.389 align:start position:0% vertical where essentially you're trying to<00:00:56.910> attack<00:00:57.449> the<00:00:57.600> network<00:00:57.899> that<00:00:58.230> employees<00:00:58.620> the 00:00:59.389 --> 00:00:59.399 align:start position:0% to attack the network that employees the 00:00:59.399 --> 00:01:02.510 align:start position:0% to attack the network that employees the proposal<00:00:59.760> calculation<00:01:00.329> and<00:01:01.010> you<00:01:02.010> do<00:01:02.190> this<00:01:02.370> by 00:01:02.510 --> 00:01:02.520 align:start position:0% proposal calculation and you do this by 00:01:02.520 --> 00:01:05.630 align:start position:0% proposal calculation and you do this by essentially<00:01:03.180> keeping<00:01:04.070> like<00:01:05.070> a<00:01:05.100> chain<00:01:05.399> that 00:01:05.630 --> 00:01:05.640 align:start position:0% essentially keeping like a chain that 00:01:05.640 --> 00:01:08.960 align:start position:0% essentially keeping like a chain that it's<00:01:05.789> a<00:01:06.119> head<00:01:06.860> so<00:01:07.860> essentially<00:01:08.190> you<00:01:08.759> don't 00:01:08.960 --> 00:01:08.970 align:start position:0% it's a head so essentially you don't 00:01:08.970 --> 00:01:10.550 align:start position:0% it's a head so essentially you don't send<00:01:09.360> your<00:01:09.420> blocks<00:01:09.720> as<00:01:10.050> soon<00:01:10.170> as<00:01:10.320> you<00:01:10.410> receive 00:01:10.550 --> 00:01:10.560 align:start position:0% send your blocks as soon as you receive 00:01:10.560 --> 00:01:12.320 align:start position:0% send your blocks as soon as you receive them<00:01:10.890> but<00:01:11.340> you<00:01:11.430> will<00:01:11.610> hold<00:01:11.850> them<00:01:12.060> and 00:01:12.320 --> 00:01:12.330 align:start position:0% them but you will hold them and 00:01:12.330 --> 00:01:14.660 align:start position:0% them but you will hold them and sometimes<00:01:12.539> repeatedly<00:01:13.259> as<00:01:13.440> you<00:01:13.560> keep<00:01:13.830> it 00:01:14.660 --> 00:01:14.670 align:start position:0% sometimes repeatedly as you keep it 00:01:14.670 --> 00:01:19.400 align:start position:0% sometimes repeatedly as you keep it heads<00:01:15.000> from<00:01:15.840> other<00:01:16.110> -<00:01:16.590> and<00:01:17.130> so<00:01:17.539> the<00:01:18.539> end<00:01:18.899> you're 00:01:19.400 --> 00:01:19.410 align:start position:0% heads from other - and so the end you're 00:01:19.410 --> 00:01:22.100 align:start position:0% heads from other - and so the end you're gonna<00:01:19.560> make<00:01:19.920> the<00:01:20.039> oyster<00:01:20.459> time<00:01:20.670> by<00:01:21.450> looking<00:01:21.840> at 00:01:22.100 --> 00:01:22.110 align:start position:0% gonna make the oyster time by looking at 00:01:22.110 --> 00:01:23.540 align:start position:0% gonna make the oyster time by looking at a<00:01:22.140> loss<00:01:22.349> that<00:01:22.560> you<00:01:22.649> just<00:01:22.860> sent<00:01:23.099> and<00:01:23.429> you're 00:01:23.540 --> 00:01:23.550 align:start position:0% a loss that you just sent and you're 00:01:23.550 --> 00:01:25.400 align:start position:0% a loss that you just sent and you're always<00:01:23.729> keep<00:01:24.060> ahead<00:01:24.330> so<00:01:24.750> they<00:01:24.840> do<00:01:24.899> ensure<00:01:25.289> that 00:01:25.400 --> 00:01:25.410 align:start position:0% always keep ahead so they do ensure that 00:01:25.410 --> 00:01:28.940 align:start position:0% always keep ahead so they do ensure that you're<00:01:25.590> winning<00:01:25.800> and<00:01:27.140> so<00:01:28.140> this<00:01:28.380> is<00:01:28.500> a<00:01:28.530> problem 00:01:28.940 --> 00:01:28.950 align:start position:0% you're winning and so this is a problem 00:01:28.950 --> 00:01:31.190 align:start position:0% you're winning and so this is a problem that's<00:01:29.039> they've<00:01:29.280> studied<00:01:29.550> and<00:01:30.330> different 00:01:31.190 --> 00:01:31.200 align:start position:0% that's they've studied and different 00:01:31.200 --> 00:01:32.870 align:start position:0% that's they've studied and different papers<00:01:31.530> and<00:01:31.860> there's<00:01:32.160> different<00:01:32.490> algorithms 00:01:32.870 --> 00:01:32.880 align:start position:0% papers and there's different algorithms 00:01:32.880 --> 00:01:34.460 align:start position:0% papers and there's different algorithms that<00:01:32.940> you<00:01:33.060> can<00:01:33.209> follow<00:01:33.450> as<00:01:34.020> I<00:01:34.140> described 00:01:34.460 --> 00:01:34.470 align:start position:0% that you can follow as I described 00:01:34.470 --> 00:01:36.650 align:start position:0% that you can follow as I described before<00:01:34.500> you<00:01:34.920> can<00:01:35.130> hold<00:01:35.849> a<00:01:36.030> number<00:01:36.149> of<00:01:36.360> Bluffs 00:01:36.650 --> 00:01:36.660 align:start position:0% before you can hold a number of Bluffs 00:01:36.660 --> 00:01:39.320 align:start position:0% before you can hold a number of Bluffs and<00:01:37.020> whenever<00:01:37.860> you<00:01:38.099> how<00:01:38.250> elite<00:01:38.550> if<00:01:38.940> someone 00:01:39.320 --> 00:01:39.330 align:start position:0% and whenever you how elite if someone 00:01:39.330 --> 00:01:41.719 align:start position:0% and whenever you how elite if someone sends<00:01:39.630> you<00:01:39.660> a<00:01:39.720> block<00:01:40.289> then<00:01:40.950> increasing<00:01:41.610> the 00:01:41.719 --> 00:01:41.729 align:start position:0% sends you a block then increasing the 00:01:41.729 --> 00:01:45.219 align:start position:0% sends you a block then increasing the general<00:01:42.509> heights<00:01:42.840> you<00:01:43.110> publish<00:01:43.800> a<00:01:43.830> walk<00:01:44.130> that 00:01:45.219 --> 00:01:45.229 align:start position:0% general heights you publish a walk that 00:01:45.229 --> 00:01:48.190 align:start position:0% general heights you publish a walk that and<00:01:46.030> essentially<00:01:47.030> always<00:01:47.210> keeping<00:01:47.750> a<00:01:47.960> heads 00:01:48.190 --> 00:01:48.200 align:start position:0% and essentially always keeping a heads 00:01:48.200 --> 00:01:49.600 align:start position:0% and essentially always keeping a heads up<00:01:48.320> in<00:01:48.439> to<00:01:48.530> ensure<00:01:48.979> that<00:01:49.009> you<00:01:49.250> can<00:01:49.430> maximize 00:01:49.600 --> 00:01:49.610 align:start position:0% up in to ensure that you can maximize 00:01:49.610 --> 00:01:52.450 align:start position:0% up in to ensure that you can maximize aware<00:01:50.210> that<00:01:50.299> you<00:01:50.390> getting<00:01:50.740> because<00:01:51.740> when<00:01:52.340> you 00:01:52.450 --> 00:01:52.460 align:start position:0% aware that you getting because when you 00:01:52.460 --> 00:01:54.399 align:start position:0% aware that you getting because when you something<00:01:52.759> you<00:01:52.970> either<00:01:53.330> get<00:01:53.600> a<00:01:53.659> maintained 00:01:54.399 --> 00:01:54.409 align:start position:0% something you either get a maintained 00:01:54.409 --> 00:01:56.980 align:start position:0% something you either get a maintained reward<00:01:54.890> or<00:01:55.100> an<00:01:55.250> awful<00:01:55.790> reward<00:01:56.540> which<00:01:56.810> are 00:01:56.980 --> 00:01:56.990 align:start position:0% reward or an awful reward which are 00:01:56.990 --> 00:02:00.340 align:start position:0% reward or an awful reward which are almost<00:01:57.830> essentially<00:01:58.280> the<00:01:58.369> same<00:01:59.110> and<00:02:00.110> so<00:02:00.320> I 00:02:00.340 --> 00:02:00.350 align:start position:0% almost essentially the same and so I 00:02:00.350 --> 00:02:01.779 align:start position:0% almost essentially the same and so I learned<00:02:00.860> with<00:02:00.979> the<00:02:01.040> Pegasus<00:02:01.430> team<00:02:01.640> at 00:02:01.779 --> 00:02:01.789 align:start position:0% learned with the Pegasus team at 00:02:01.789 --> 00:02:04.389 align:start position:0% learned with the Pegasus team at consensus<00:02:02.390> and<00:02:02.780> we<00:02:03.229> focus<00:02:03.530> on<00:02:03.680> research<00:02:04.039> so 00:02:04.389 --> 00:02:04.399 align:start position:0% consensus and we focus on research so 00:02:04.399 --> 00:02:06.430 align:start position:0% consensus and we focus on research so make<00:02:04.640> develop<00:02:05.060> a<00:02:05.210> framework<00:02:05.539> to<00:02:05.899> build 00:02:06.430 --> 00:02:06.440 align:start position:0% make develop a framework to build 00:02:06.440 --> 00:02:08.080 align:start position:0% make develop a framework to build simulations<00:02:07.190> for<00:02:07.399> different<00:02:07.789> types<00:02:07.939> of 00:02:08.080 --> 00:02:08.090 align:start position:0% simulations for different types of 00:02:08.090 --> 00:02:09.999 align:start position:0% simulations for different types of particles<00:02:08.330> so<00:02:09.110> we<00:02:09.319> look<00:02:09.500> them<00:02:09.649> through<00:02:09.830> from 00:02:09.999 --> 00:02:10.009 align:start position:0% particles so we look them through from 00:02:10.009 --> 00:02:12.790 align:start position:0% particles so we look them through from work<00:02:10.160> for<00:02:10.429> the<00:02:10.580> state<00:02:10.819> and<00:02:11.090> so<00:02:11.959> we<00:02:12.319> decided<00:02:12.650> it 00:02:12.790 --> 00:02:12.800 align:start position:0% work for the state and so we decided it 00:02:12.800 --> 00:02:14.770 align:start position:0% work for the state and so we decided it would<00:02:12.890> be<00:02:12.920> pretty<00:02:13.190> interesting<00:02:14.060> to<00:02:14.630> move 00:02:14.770 --> 00:02:14.780 align:start position:0% would be pretty interesting to move 00:02:14.780 --> 00:02:17.530 align:start position:0% would be pretty interesting to move implements<00:02:15.230> a<00:02:15.290> selfish<00:02:15.709> attack<00:02:16.250> and<00:02:16.640> look<00:02:17.450> at 00:02:17.530 --> 00:02:17.540 align:start position:0% implements a selfish attack and look at 00:02:17.540 --> 00:02:21.970 align:start position:0% implements a selfish attack and look at the<00:02:17.600> data<00:02:17.750> that<00:02:17.840> wouldn't<00:02:18.560> return<00:02:19.540> so<00:02:20.830> this<00:02:21.830> is 00:02:21.970 --> 00:02:21.980 align:start position:0% the data that wouldn't return so this is 00:02:21.980 --> 00:02:23.229 align:start position:0% the data that wouldn't return so this is a<00:02:22.010> bit<00:02:22.190> like<00:02:22.400> oh<00:02:22.519> there<00:02:22.700> looks<00:02:22.910> like<00:02:22.970> we 00:02:23.229 --> 00:02:23.239 align:start position:0% a bit like oh there looks like we 00:02:23.239 --> 00:02:24.789 align:start position:0% a bit like oh there looks like we actually<00:02:23.630> took<00:02:23.810> the<00:02:24.110> combination<00:02:24.620> of<00:02:24.680> the 00:02:24.789 --> 00:02:24.799 align:start position:0% actually took the combination of the 00:02:24.799 --> 00:02:26.770 align:start position:0% actually took the combination of the difficulty<00:02:25.340> we're<00:02:25.910> looking<00:02:26.269> at<00:02:26.360> how<00:02:26.569> it's 00:02:26.770 --> 00:02:26.780 align:start position:0% difficulty we're looking at how it's 00:02:26.780 --> 00:02:28.360 align:start position:0% difficulty we're looking at how it's done<00:02:26.959> if<00:02:27.200> you're<00:02:27.470> really<00:02:27.620> just<00:02:27.769> a<00:02:28.069> kind<00:02:28.220> of<00:02:28.310> a 00:02:28.360 --> 00:02:28.370 align:start position:0% done if you're really just a kind of a 00:02:28.370 --> 00:02:30.340 align:start position:0% done if you're really just a kind of a sense<00:02:28.610> of<00:02:28.670> what<00:02:28.730> their<00:02:29.000> reward<00:02:29.299> time<00:02:29.480> and<00:02:29.720> so 00:02:30.340 --> 00:02:30.350 align:start position:0% sense of what their reward time and so 00:02:30.350 --> 00:02:32.110 align:start position:0% sense of what their reward time and so the<00:02:30.680> only<00:02:30.890> what<00:02:31.040> you're<00:02:31.220> doing<00:02:31.549> as<00:02:31.790> I 00:02:32.110 --> 00:02:32.120 align:start position:0% the only what you're doing as I 00:02:32.120 --> 00:02:33.430 align:start position:0% the only what you're doing as I mentioned<00:02:32.480> before<00:02:32.540> you<00:02:32.870> can<00:02:32.930> get<00:02:33.170> a<00:02:33.230> reward 00:02:33.430 --> 00:02:33.440 align:start position:0% mentioned before you can get a reward 00:02:33.440 --> 00:02:36.039 align:start position:0% mentioned before you can get a reward from<00:02:34.090> mining<00:02:35.090> a<00:02:35.150> block<00:02:35.360> which<00:02:35.600> is<00:02:35.780> right<00:02:35.959> now 00:02:36.039 --> 00:02:36.049 align:start position:0% from mining a block which is right now 00:02:36.049 --> 00:02:38.199 align:start position:0% from mining a block which is right now to<00:02:36.110> eat<00:02:36.530> when<00:02:37.040> the<00:02:37.190> transaction<00:02:37.700> is<00:02:37.760> included 00:02:38.199 --> 00:02:38.209 align:start position:0% to eat when the transaction is included 00:02:38.209 --> 00:02:40.630 align:start position:0% to eat when the transaction is included and<00:02:38.360> if<00:02:38.720> you<00:02:38.900> include<00:02:39.019> any<00:02:39.769> animals<00:02:40.280> that<00:02:40.370> you 00:02:40.630 --> 00:02:40.640 align:start position:0% and if you include any animals that you 00:02:40.640 --> 00:02:45.580 align:start position:0% and if you include any animals that you receive<00:02:41.030> so<00:02:41.540> your<00:02:41.930> ordinary<00:02:43.090> -<00:02:44.410> 175 00:02:45.580 --> 00:02:45.590 align:start position:0% receive so your ordinary - 175 00:02:45.590 --> 00:02:50.229 align:start position:0% receive so your ordinary - 175 125<00:02:46.310> or<00:02:46.760> -<00:02:47.470> and<00:02:48.470> since<00:02:49.160> the<00:02:49.340> transaction<00:02:49.970> fees 00:02:50.229 --> 00:02:50.239 align:start position:0% 125 or - and since the transaction fees 00:02:50.239 --> 00:02:52.449 align:start position:0% 125 or - and since the transaction fees that<00:02:50.390> you<00:02:50.510> receive<00:02:50.840> this<00:02:51.160> quite<00:02:52.160> small 00:02:52.449 --> 00:02:52.459 align:start position:0% that you receive this quite small 00:02:52.459 --> 00:02:54.550 align:start position:0% that you receive this quite small sometimes<00:02:52.940> we<00:02:53.120> decide<00:02:53.359> to<00:02:53.540> not<00:02:54.019> include<00:02:54.320> this 00:02:54.550 --> 00:02:54.560 align:start position:0% sometimes we decide to not include this 00:02:54.560 --> 00:02:58.599 align:start position:0% sometimes we decide to not include this in<00:02:55.370> our<00:02:55.519> own<00:02:55.549> simulation<00:02:56.299> and<00:02:56.769> so<00:02:57.769> we<00:02:58.280> decided 00:02:58.599 --> 00:02:58.609 align:start position:0% in our own simulation and so we decided 00:02:58.609 --> 00:03:00.190 align:start position:0% in our own simulation and so we decided to<00:02:58.700> build<00:02:58.850> a<00:02:59.000> test<00:02:59.269> case<00:02:59.420> where<00:02:59.780> you<00:02:59.870> only<00:03:00.079> have 00:03:00.190 --> 00:03:00.200 align:start position:0% to build a test case where you only have 00:03:00.200 --> 00:03:03.640 align:start position:0% to build a test case where you only have a<00:03:00.410> network<00:03:00.859> of<00:03:01.180> nine<00:03:02.180> miners<00:03:02.510> were<00:03:03.109> always<00:03:03.470> fun 00:03:03.640 --> 00:03:03.650 align:start position:0% a network of nine miners were always fun 00:03:03.650 --> 00:03:04.540 align:start position:0% a network of nine miners were always fun and 00:03:04.540 --> 00:03:04.550 align:start position:0% and 00:03:04.550 --> 00:03:08.670 align:start position:0% and liner<00:03:05.060> that<00:03:05.390> is<00:03:06.200> selfishly<00:03:06.770> mining<00:03:07.400> and<00:03:08.240> so 00:03:08.670 --> 00:03:08.680 align:start position:0% liner that is selfishly mining and so 00:03:08.680 --> 00:03:10.720 align:start position:0% liner that is selfishly mining and so what<00:03:09.680> happens<00:03:10.010> if<00:03:10.250> that<00:03:10.400> there's<00:03:10.550> some 00:03:10.720 --> 00:03:10.730 align:start position:0% what happens if that there's some 00:03:10.730 --> 00:03:13.810 align:start position:0% what happens if that there's some variation<00:03:11.360> and<00:03:11.660> we<00:03:12.290> decided<00:03:12.890> not<00:03:13.070> to<00:03:13.130> internal 00:03:13.810 --> 00:03:13.820 align:start position:0% variation and we decided not to internal 00:03:13.820 --> 00:03:15.700 align:start position:0% variation and we decided not to internal warrants<00:03:14.120> of<00:03:14.240> transactions<00:03:14.810> and<00:03:15.170> some<00:03:15.560> of 00:03:15.700 --> 00:03:15.710 align:start position:0% warrants of transactions and some of 00:03:15.710 --> 00:03:18.490 align:start position:0% warrants of transactions and some of their<00:03:15.830> results<00:03:16.160> on<00:03:16.610> murder<00:03:16.910> this<00:03:17.500> so 00:03:18.490 --> 00:03:18.500 align:start position:0% their results on murder this so 00:03:18.500 --> 00:03:20.590 align:start position:0% their results on murder this so essentially<00:03:19.430> what<00:03:19.610> we<00:03:19.700> found<00:03:19.880> is<00:03:20.150> that<00:03:20.180> once 00:03:20.590 --> 00:03:20.600 align:start position:0% essentially what we found is that once 00:03:20.600 --> 00:03:22.360 align:start position:0% essentially what we found is that once you<00:03:20.780> have<00:03:20.810> to<00:03:21.170> like<00:03:21.650> this<00:03:21.890> time<00:03:22.100> of<00:03:22.250> the 00:03:22.360 --> 00:03:22.370 align:start position:0% you have to like this time of the 00:03:22.370 --> 00:03:24.700 align:start position:0% you have to like this time of the hashing<00:03:22.820> power<00:03:22.850> in<00:03:23.210> the<00:03:23.300> network<00:03:23.660> you<00:03:24.320> can 00:03:24.700 --> 00:03:24.710 align:start position:0% hashing power in the network you can 00:03:24.710 --> 00:03:26.380 align:start position:0% hashing power in the network you can start<00:03:24.950> attacking<00:03:25.280> it<00:03:25.550> and<00:03:25.820> doing<00:03:26.030> selfish 00:03:26.380 --> 00:03:26.390 align:start position:0% start attacking it and doing selfish 00:03:26.390 --> 00:03:27.970 align:start position:0% start attacking it and doing selfish mining<00:03:26.750> and<00:03:27.020> this<00:03:27.170> is<00:03:27.320> more<00:03:27.470> beneficial<00:03:27.860> for 00:03:27.970 --> 00:03:27.980 align:start position:0% mining and this is more beneficial for 00:03:27.980 --> 00:03:30.790 align:start position:0% mining and this is more beneficial for you<00:03:28.660> because<00:03:29.660> this<00:03:29.900> manipulates<00:03:30.500> the 00:03:30.790 --> 00:03:30.800 align:start position:0% you because this manipulates the 00:03:30.800 --> 00:03:33.370 align:start position:0% you because this manipulates the difficulty<00:03:31.400> calculation<00:03:32.000> and<00:03:32.240> it<00:03:32.870> increases 00:03:33.370 --> 00:03:33.380 align:start position:0% difficulty calculation and it increases 00:03:33.380 --> 00:03:34.630 align:start position:0% difficulty calculation and it increases your<00:03:33.410> own<00:03:33.680> work<00:03:33.830> so<00:03:34.010> they're<00:03:34.160> more<00:03:34.280> hashing 00:03:34.630 --> 00:03:34.640 align:start position:0% your own work so they're more hashing 00:03:34.640 --> 00:03:37.240 align:start position:0% your own work so they're more hashing power<00:03:34.670> you<00:03:35.090> have<00:03:35.410> the<00:03:36.410> better<00:03:36.620> for<00:03:36.890> you 00:03:37.240 --> 00:03:37.250 align:start position:0% power you have the better for you 00:03:37.250 --> 00:03:41.080 align:start position:0% power you have the better for you and<00:03:38.060> so<00:03:38.270> if<00:03:38.420> you<00:03:38.510> wanted<00:03:38.690> to<00:03:38.750> test<00:03:39.520> how<00:03:40.520> would 00:03:41.080 --> 00:03:41.090 align:start position:0% and so if you wanted to test how would 00:03:41.090 --> 00:03:42.850 align:start position:0% and so if you wanted to test how would this<00:03:41.270> prove<00:03:41.720> if<00:03:42.020> you<00:03:42.140> are<00:03:42.290> giving<00:03:42.620> it<00:03:42.710> to<00:03:42.740> a 00:03:42.850 --> 00:03:42.860 align:start position:0% this prove if you are giving it to a 00:03:42.860 --> 00:03:45.250 align:start position:0% this prove if you are giving it to a reinforced<00:03:43.460> not<00:03:43.580> an<00:03:43.730> agent<00:03:44.270> and<00:03:44.390> since<00:03:45.140> we 00:03:45.250 --> 00:03:45.260 align:start position:0% reinforced not an agent and since we 00:03:45.260 --> 00:03:47.140 align:start position:0% reinforced not an agent and since we have<00:03:45.470> actually<00:03:46.100> active<00:03:46.430> back<00:03:46.580> it<00:03:46.730> up<00:03:46.760> we<00:03:47.060> can 00:03:47.140 --> 00:03:47.150 align:start position:0% have actually active back it up we can 00:03:47.150 --> 00:03:48.670 align:start position:0% have actually active back it up we can compare<00:03:47.480> the<00:03:47.630> results<00:03:47.660> and<00:03:48.230> see<00:03:48.440> what<00:03:48.590> we 00:03:48.670 --> 00:03:48.680 align:start position:0% compare the results and see what we 00:03:48.680 --> 00:03:52.210 align:start position:0% compare the results and see what we would<00:03:48.800> get<00:03:48.950> with<00:03:49.130> it<00:03:50.350> and<00:03:51.350> so<00:03:51.560> that's<00:03:52.010> why<00:03:52.160> you 00:03:52.210 --> 00:03:52.220 align:start position:0% would get with it and so that's why you 00:03:52.220 --> 00:03:53.440 align:start position:0% would get with it and so that's why you might<00:03:52.400> be<00:03:52.520> wondering<00:03:52.700> why<00:03:52.940> are<00:03:52.970> we<00:03:53.180> doing 00:03:53.440 --> 00:03:53.450 align:start position:0% might be wondering why are we doing 00:03:53.450 --> 00:03:54.730 align:start position:0% might be wondering why are we doing reinforcement<00:03:54.080> learning<00:03:54.170> when<00:03:54.470> we<00:03:54.590> were 00:03:54.730 --> 00:03:54.740 align:start position:0% reinforcement learning when we were 00:03:54.740 --> 00:03:57.790 align:start position:0% reinforcement learning when we were focused<00:03:55.100> on<00:03:55.160> watching<00:03:56.410> but<00:03:57.410> if<00:03:57.530> this<00:03:57.620> will 00:03:57.790 --> 00:03:57.800 align:start position:0% focused on watching but if this will 00:03:57.800 --> 00:03:58.840 align:start position:0% focused on watching but if this will give<00:03:57.920> you<00:03:58.010> like<00:03:58.130> a<00:03:58.160> sense<00:03:58.460> of<00:03:58.640> what 00:03:58.840 --> 00:03:58.850 align:start position:0% give you like a sense of what 00:03:58.850 --> 00:04:00.580 align:start position:0% give you like a sense of what reinforcement<00:03:59.690> learning<00:03:59.750> is<00:04:00.020> in<00:04:00.290> case<00:04:00.470> you 00:04:00.580 --> 00:04:00.590 align:start position:0% reinforcement learning is in case you 00:04:00.590 --> 00:04:02.260 align:start position:0% reinforcement learning is in case you haven't<00:04:00.770> heard<00:04:00.920> it<00:04:01.190> is<00:04:01.400> the<00:04:01.580> concept<00:04:02.150> where 00:04:02.260 --> 00:04:02.270 align:start position:0% haven't heard it is the concept where 00:04:02.270 --> 00:04:04.600 align:start position:0% haven't heard it is the concept where you<00:04:02.300> have<00:04:02.570> a<00:04:02.600> piece<00:04:03.560> of<00:04:03.590> software<00:04:04.070> it's<00:04:04.280> an<00:04:04.430> a 00:04:04.600 --> 00:04:04.610 align:start position:0% you have a piece of software it's an a 00:04:04.610 --> 00:04:06.250 align:start position:0% you have a piece of software it's an a gendered<00:04:04.970> construct<00:04:05.450> for<00:04:05.600> an<00:04:05.750> environment 00:04:06.250 --> 00:04:06.260 align:start position:0% gendered construct for an environment 00:04:06.260 --> 00:04:09.490 align:start position:0% gendered construct for an environment one<00:04:06.590> of<00:04:06.680> the<00:04:06.770> most<00:04:06.950> famous<00:04:07.160> ones<00:04:07.550> is<00:04:08.200> deep<00:04:09.200> mind 00:04:09.490 --> 00:04:09.500 align:start position:0% one of the most famous ones is deep mind 00:04:09.500 --> 00:04:13.180 align:start position:0% one of the most famous ones is deep mind where<00:04:10.340> they<00:04:10.520> implement<00:04:11.420> in<00:04:11.540> a<00:04:11.630> deep<00:04:12.200> Q<00:04:12.770> network 00:04:13.180 --> 00:04:13.190 align:start position:0% where they implement in a deep Q network 00:04:13.190 --> 00:04:15.310 align:start position:0% where they implement in a deep Q network and<00:04:13.600> essentially<00:04:14.600> interacting<00:04:15.170> with 00:04:15.310 --> 00:04:15.320 align:start position:0% and essentially interacting with 00:04:15.320 --> 00:04:17.110 align:start position:0% and essentially interacting with different<00:04:15.440> attendings<00:04:16.190> and<00:04:16.400> it<00:04:16.580> learns<00:04:16.790> how 00:04:17.110 --> 00:04:17.120 align:start position:0% different attendings and it learns how 00:04:17.120 --> 00:04:19.270 align:start position:0% different attendings and it learns how to<00:04:17.150> solve<00:04:17.510> it<00:04:17.690> so<00:04:17.840> we<00:04:17.959> thought<00:04:18.140> okay<00:04:18.950> we<00:04:18.980> can 00:04:19.270 --> 00:04:19.280 align:start position:0% to solve it so we thought okay we can 00:04:19.280 --> 00:04:21.070 align:start position:0% to solve it so we thought okay we can take<00:04:19.400> this<00:04:19.520> concept<00:04:19.910> and<00:04:20.060> implemented<00:04:20.840> into 00:04:21.070 --> 00:04:21.080 align:start position:0% take this concept and implemented into 00:04:21.080 --> 00:04:23.410 align:start position:0% take this concept and implemented into different<00:04:21.470> protocols<00:04:21.890> and<00:04:22.130> just<00:04:22.940> see<00:04:23.270> what 00:04:23.410 --> 00:04:23.420 align:start position:0% different protocols and just see what 00:04:23.420 --> 00:04:23.970 align:start position:0% different protocols and just see what happens 00:04:23.970 --> 00:04:23.980 align:start position:0% happens 00:04:23.980 --> 00:04:28.000 align:start position:0% happens and<00:04:25.090> then<00:04:26.090> just<00:04:26.360> like<00:04:26.780> make<00:04:27.170> a<00:04:27.290> sense<00:04:27.530> of<00:04:27.650> why 00:04:28.000 --> 00:04:28.010 align:start position:0% and then just like make a sense of why 00:04:28.010 --> 00:04:30.340 align:start position:0% and then just like make a sense of why we<00:04:28.190> chose<00:04:28.340> that<00:04:28.400> too<00:04:28.820> is<00:04:29.060> because<00:04:29.540> it's<00:04:30.170> an 00:04:30.340 --> 00:04:30.350 align:start position:0% we chose that too is because it's an 00:04:30.350 --> 00:04:32.020 align:start position:0% we chose that too is because it's an automated<00:04:30.770> system<00:04:31.010> that<00:04:31.430> essentially<00:04:31.730> has 00:04:32.020 --> 00:04:32.030 align:start position:0% automated system that essentially has 00:04:32.030 --> 00:04:34.120 align:start position:0% automated system that essentially has proven<00:04:32.390> to<00:04:32.540> beat<00:04:32.690> humans<00:04:33.590> and<00:04:33.830> many<00:04:33.920> other 00:04:34.120 --> 00:04:34.130 align:start position:0% proven to beat humans and many other 00:04:34.130 --> 00:04:36.400 align:start position:0% proven to beat humans and many other elements<00:04:34.340> and<00:04:34.910> strategies<00:04:35.300> so<00:04:35.600> it<00:04:35.900> just<00:04:36.080> made 00:04:36.400 --> 00:04:36.410 align:start position:0% elements and strategies so it just made 00:04:36.410 --> 00:04:39.520 align:start position:0% elements and strategies so it just made sense<00:04:36.440> to<00:04:36.710> go<00:04:36.980> that<00:04:37.280> route<00:04:37.430> so<00:04:38.320> how<00:04:39.320> does<00:04:39.380> that 00:04:39.520 --> 00:04:39.530 align:start position:0% sense to go that route so how does that 00:04:39.530 --> 00:04:42.880 align:start position:0% sense to go that route so how does that work<00:04:39.680> you<00:04:40.010> know<00:04:40.790> this<00:04:40.960> room<00:04:41.960> or<00:04:42.440> you<00:04:42.560> have<00:04:42.710> an 00:04:42.880 --> 00:04:42.890 align:start position:0% work you know this room or you have an 00:04:42.890 --> 00:04:45.370 align:start position:0% work you know this room or you have an agent<00:04:43.340> that<00:04:43.550> is<00:04:44.000> collecting<00:04:44.600> ice<00:04:44.720> box 00:04:45.370 --> 00:04:45.380 align:start position:0% agent that is collecting ice box 00:04:45.380 --> 00:04:47.440 align:start position:0% agent that is collecting ice box the<00:04:45.740> environment<00:04:46.280> so<00:04:46.850> it's<00:04:47.000> partial 00:04:47.440 --> 00:04:47.450 align:start position:0% the environment so it's partial 00:04:47.450 --> 00:04:49.000 align:start position:0% the environment so it's partial information<00:04:47.570> on<00:04:48.080> the<00:04:48.170> current<00:04:48.290> state<00:04:48.680> and 00:04:49.000 --> 00:04:49.010 align:start position:0% information on the current state and 00:04:49.010 --> 00:04:51.450 align:start position:0% information on the current state and then<00:04:49.730> you<00:04:50.240> have<00:04:50.450> actions<00:04:50.990> that<00:04:51.140> you<00:04:51.260> can<00:04:51.290> take 00:04:51.450 --> 00:04:51.460 align:start position:0% then you have actions that you can take 00:04:51.460 --> 00:04:54.340 align:start position:0% then you have actions that you can take which<00:04:52.460> will<00:04:52.700> return<00:04:52.880> a<00:04:53.090> reward<00:04:53.480> based<00:04:54.110> on 00:04:54.340 --> 00:04:54.350 align:start position:0% which will return a reward based on 00:04:54.350 --> 00:04:56.920 align:start position:0% which will return a reward based on those<00:04:54.500> actions<00:04:54.680> and<00:04:55.310> so<00:04:55.550> the<00:04:55.850> process<00:04:56.570> into 00:04:56.920 --> 00:04:56.930 align:start position:0% those actions and so the process into 00:04:56.930 --> 00:04:59.620 align:start position:0% those actions and so the process into using<00:04:57.260> some<00:04:57.290> loop<00:04:57.950> where<00:04:58.850> you<00:04:58.880> just<00:04:59.240> keep<00:04:59.420> it's 00:04:59.620 --> 00:04:59.630 align:start position:0% using some loop where you just keep it's 00:04:59.630 --> 00:05:01.240 align:start position:0% using some loop where you just keep it's raining<00:04:59.900> two<00:05:00.020> different<00:05:00.380> actions<00:05:00.560> in<00:05:01.010> the<00:05:01.100> ten 00:05:01.240 --> 00:05:01.250 align:start position:0% raining two different actions in the ten 00:05:01.250 --> 00:05:05.170 align:start position:0% raining two different actions in the ten states<00:05:01.600> and<00:05:03.250> so<00:05:04.250> what<00:05:04.700> are<00:05:04.790> the<00:05:04.880> different 00:05:05.170 --> 00:05:05.180 align:start position:0% states and so what are the different 00:05:05.180 --> 00:05:06.910 align:start position:0% states and so what are the different observations<00:05:05.690> that<00:05:05.810> you<00:05:05.900> can<00:05:06.050> get<00:05:06.200> in<00:05:06.350> terms 00:05:06.910 --> 00:05:06.920 align:start position:0% observations that you can get in terms 00:05:06.920 --> 00:05:09.400 align:start position:0% observations that you can get in terms of<00:05:07.100> our<00:05:07.820> product<00:05:08.180> over<00:05:08.480> looking<00:05:09.230> at<00:05:09.350> the 00:05:09.400 --> 00:05:09.410 align:start position:0% of our product over looking at the 00:05:09.410 --> 00:05:11.560 align:start position:0% of our product over looking at the number<00:05:09.800> of<00:05:09.830> blocks<00:05:10.040> mine<00:05:10.520> the<00:05:11.060> number<00:05:11.300> of<00:05:11.390> luck 00:05:11.560 --> 00:05:11.570 align:start position:0% number of blocks mine the number of luck 00:05:11.570 --> 00:05:13.540 align:start position:0% number of blocks mine the number of luck receive<00:05:11.930> or<00:05:12.200> something<00:05:12.560> all<00:05:12.920> those<00:05:13.130> different 00:05:13.540 --> 00:05:13.550 align:start position:0% receive or something all those different 00:05:13.550 --> 00:05:15.850 align:start position:0% receive or something all those different things<00:05:13.610> that<00:05:13.850> you're<00:05:14.000> actually<00:05:14.590> doing<00:05:15.590> in<00:05:15.830> the 00:05:15.850 --> 00:05:15.860 align:start position:0% things that you're actually doing in the 00:05:15.860 --> 00:05:17.470 align:start position:0% things that you're actually doing in the network<00:05:16.310> and<00:05:16.610> feeding<00:05:17.090> that<00:05:17.300> information 00:05:17.470 --> 00:05:17.480 align:start position:0% network and feeding that information 00:05:17.480 --> 00:05:20.470 align:start position:0% network and feeding that information into<00:05:17.960> the<00:05:18.110> agent<00:05:18.530> so<00:05:18.710> they<00:05:18.830> can<00:05:19.040> learn<00:05:19.480> the 00:05:20.470 --> 00:05:20.480 align:start position:0% into the agent so they can learn the 00:05:20.480 --> 00:05:22.000 align:start position:0% into the agent so they can learn the impact<00:05:20.870> of<00:05:20.990> their<00:05:21.140> actions 00:05:22.000 --> 00:05:22.010 align:start position:0% impact of their actions 00:05:22.010 --> 00:05:25.480 align:start position:0% impact of their actions and<00:05:22.990> so<00:05:23.990> it's<00:05:24.410> really<00:05:24.470> still<00:05:24.890> blessing<00:05:25.310> but 00:05:25.480 --> 00:05:25.490 align:start position:0% and so it's really still blessing but 00:05:25.490 --> 00:05:27.250 align:start position:0% and so it's really still blessing but they<00:05:25.580> don't<00:05:25.730> want<00:05:25.850> to<00:05:25.940> be<00:05:26.450> too<00:05:26.750> broad<00:05:26.990> in<00:05:27.230> terms 00:05:27.250 --> 00:05:27.260 align:start position:0% they don't want to be too broad in terms 00:05:27.260 --> 00:05:29.740 align:start position:0% they don't want to be too broad in terms of<00:05:27.590> wider<00:05:27.800> minor<00:05:28.160> can<00:05:28.400> do<00:05:28.550> so<00:05:29.000> we<00:05:29.300> decided<00:05:29.660> that 00:05:29.740 --> 00:05:29.750 align:start position:0% of wider minor can do so we decided that 00:05:29.750 --> 00:05:34.060 align:start position:0% of wider minor can do so we decided that you<00:05:30.140> can<00:05:30.290> only<00:05:30.410> combine<00:05:32.080> up<00:05:33.080> to<00:05:33.350> ten<00:05:33.680> blocks 00:05:34.060 --> 00:05:34.070 align:start position:0% you can only combine up to ten blocks 00:05:34.070 --> 00:05:38.020 align:start position:0% you can only combine up to ten blocks hidden<00:05:34.940> in<00:05:35.210> your<00:05:35.450> own<00:05:35.630> private<00:05:36.380> chain<00:05:36.800> and<00:05:37.130> you 00:05:38.020 --> 00:05:38.030 align:start position:0% hidden in your own private chain and you 00:05:38.030 --> 00:05:40.060 align:start position:0% hidden in your own private chain and you should<00:05:38.210> be<00:05:38.330> able<00:05:38.450> to<00:05:38.630> publish<00:05:38.960> up<00:05:39.440> to<00:05:39.710> three 00:05:40.060 --> 00:05:40.070 align:start position:0% should be able to publish up to three 00:05:40.070 --> 00:05:42.550 align:start position:0% should be able to publish up to three blocks<00:05:40.340> in<00:05:40.520> a<00:05:40.580> row<00:05:40.730> so<00:05:40.790> you<00:05:41.030> could<00:05:41.350> publish<00:05:42.350> one 00:05:42.550 --> 00:05:42.560 align:start position:0% blocks in a row so you could publish one 00:05:42.560 --> 00:05:43.960 align:start position:0% blocks in a row so you could publish one two<00:05:42.800> or<00:05:42.950> three<00:05:43.130> based<00:05:43.370> on<00:05:43.550> whatever<00:05:43.730> it's 00:05:43.960 --> 00:05:43.970 align:start position:0% two or three based on whatever it's 00:05:43.970 --> 00:05:46.600 align:start position:0% two or three based on whatever it's rented<00:05:44.330> al<00:05:44.660> seems<00:05:45.050> the<00:05:45.260> best<00:05:45.440> and<00:05:45.800> then<00:05:46.460> you 00:05:46.600 --> 00:05:46.610 align:start position:0% rented al seems the best and then you 00:05:46.610 --> 00:05:49.350 align:start position:0% rented al seems the best and then you would<00:05:46.760> get<00:05:46.880> some<00:05:47.240> reward<00:05:47.600> based<00:05:47.900> on<00:05:48.080> that<00:05:48.260> and 00:05:49.350 --> 00:05:49.360 align:start position:0% would get some reward based on that and 00:05:49.360 --> 00:05:52.750 align:start position:0% would get some reward based on that and so<00:05:50.360> I<00:05:50.750> think<00:05:51.410> I'm<00:05:51.650> also<00:05:51.860> making<00:05:52.190> a<00:05:52.370> part<00:05:52.640> of 00:05:52.750 --> 00:05:52.760 align:start position:0% so I think I'm also making a part of 00:05:52.760 --> 00:05:54.550 align:start position:0% so I think I'm also making a part of setting<00:05:53.000> this<00:05:53.210> up<00:05:53.390> and<00:05:53.690> strength<00:05:54.020> understand 00:05:54.550 --> 00:05:54.560 align:start position:0% setting this up and strength understand 00:05:54.560 --> 00:05:56.620 align:start position:0% setting this up and strength understand what<00:05:55.310> is<00:05:55.490> the<00:05:55.610> best<00:05:55.760> promoter<00:05:56.030> can<00:05:56.270> you<00:05:56.420> go<00:05:56.480> to 00:05:56.620 --> 00:05:56.630 align:start position:0% what is the best promoter can you go to 00:05:56.630 --> 00:05:58.630 align:start position:0% what is the best promoter can you go to an<00:05:56.750> agent<00:05:57.140> because<00:05:57.290> you<00:05:57.470> wanna<00:05:57.800> give<00:05:58.340> her<00:05:58.460> more 00:05:58.630 --> 00:05:58.640 align:start position:0% an agent because you wanna give her more 00:05:58.640 --> 00:06:01.600 align:start position:0% an agent because you wanna give her more such<00:05:58.940> that<00:05:59.210> it<00:05:59.330> understands<00:06:00.520> what's<00:06:01.520> the 00:06:01.600 --> 00:06:01.610 align:start position:0% such that it understands what's the 00:06:01.610 --> 00:06:03.820 align:start position:0% such that it understands what's the optimal<00:06:02.030> strategy<00:06:02.240> to<00:06:02.540> think<00:06:02.750> it<00:06:02.900> also<00:06:03.050> cause 00:06:03.820 --> 00:06:03.830 align:start position:0% optimal strategy to think it also cause 00:06:03.830 --> 00:06:05.280 align:start position:0% optimal strategy to think it also cause it<00:06:04.010> was<00:06:04.100> golly<00:06:04.280> selfish<00:06:04.820> mining<00:06:05.150> without 00:06:05.280 --> 00:06:05.290 align:start position:0% it was golly selfish mining without 00:06:05.290 --> 00:06:07.900 align:start position:0% it was golly selfish mining without necessarily<00:06:06.290> implementing<00:06:07.100> the<00:06:07.550> entire 00:06:07.900 --> 00:06:07.910 align:start position:0% necessarily implementing the entire 00:06:07.910 --> 00:06:08.780 align:start position:0% necessarily implementing the entire Center 00:06:08.780 --> 00:06:08.790 align:start position:0% Center 00:06:08.790 --> 00:06:11.720 align:start position:0% Center Institute<00:06:09.480> and<00:06:10.410> so<00:06:10.650> what<00:06:11.070> we<00:06:11.160> decided<00:06:11.250> to<00:06:11.580> do 00:06:11.720 --> 00:06:11.730 align:start position:0% Institute and so what we decided to do 00:06:11.730 --> 00:06:14.390 align:start position:0% Institute and so what we decided to do in<00:06:11.850> our<00:06:11.940> case<00:06:12.150> is<00:06:12.390> to<00:06:12.570> incentivize<00:06:13.400> staying 00:06:14.390 --> 00:06:14.400 align:start position:0% in our case is to incentivize staying 00:06:14.400 --> 00:06:16.790 align:start position:0% in our case is to incentivize staying ahead<00:06:14.570> just<00:06:15.570> minding<00:06:15.900> as<00:06:16.020> many<00:06:16.230> boss<00:06:16.470> as 00:06:16.790 --> 00:06:16.800 align:start position:0% ahead just minding as many boss as 00:06:16.800 --> 00:06:20.020 align:start position:0% ahead just minding as many boss as possible<00:06:17.570> faster<00:06:18.570> than<00:06:19.050> the<00:06:19.380> other<00:06:19.500> miners 00:06:20.020 --> 00:06:20.030 align:start position:0% possible faster than the other miners 00:06:20.030 --> 00:06:22.160 align:start position:0% possible faster than the other miners but<00:06:21.030> there's<00:06:21.210> other<00:06:21.450> stretches<00:06:21.870> that<00:06:21.990> you<00:06:22.050> can 00:06:22.160 --> 00:06:22.170 align:start position:0% but there's other stretches that you can 00:06:22.170 --> 00:06:24.800 align:start position:0% but there's other stretches that you can take<00:06:22.530> we<00:06:23.040> try<00:06:23.520> to<00:06:23.580> test<00:06:23.820> in<00:06:24.000> a<00:06:24.030> few<00:06:24.090> to<00:06:24.510> see<00:06:24.660> what 00:06:24.800 --> 00:06:24.810 align:start position:0% take we try to test in a few to see what 00:06:24.810 --> 00:06:26.840 align:start position:0% take we try to test in a few to see what is<00:06:24.930> the<00:06:25.050> most<00:06:25.170> optimal<00:06:25.680> result<00:06:26.010> but 00:06:26.840 --> 00:06:26.850 align:start position:0% is the most optimal result but 00:06:26.850 --> 00:06:29.210 align:start position:0% is the most optimal result but eventually<00:06:27.240> justice<00:06:27.750> intervention<00:06:28.220> using 00:06:29.210 --> 00:06:29.220 align:start position:0% eventually justice intervention using 00:06:29.220 --> 00:06:31.670 align:start position:0% eventually justice intervention using the<00:06:29.810> metric<00:06:30.810> of<00:06:30.930> staying<00:06:31.170> ahead<00:06:31.290> that's 00:06:31.670 --> 00:06:31.680 align:start position:0% the metric of staying ahead that's 00:06:31.680 --> 00:06:35.540 align:start position:0% the metric of staying ahead that's probably<00:06:32.130> the<00:06:32.310> best<00:06:33.560> and<00:06:34.560> so<00:06:34.830> going<00:06:35.160> back<00:06:35.370> see 00:06:35.540 --> 00:06:35.550 align:start position:0% probably the best and so going back see 00:06:35.550 --> 00:06:37.610 align:start position:0% probably the best and so going back see what<00:06:35.880> this<00:06:36.030> looks<00:06:36.450> like<00:06:36.630> we<00:06:37.170> have<00:06:37.350> on<00:06:37.500> this 00:06:37.610 --> 00:06:37.620 align:start position:0% what this looks like we have on this 00:06:37.620 --> 00:06:40.040 align:start position:0% what this looks like we have on this agent<00:06:38.190> that<00:06:38.850> has<00:06:39.240> been<00:06:39.480> implemented<00:06:39.780> in 00:06:40.040 --> 00:06:40.050 align:start position:0% agent that has been implemented in 00:06:40.050 --> 00:06:42.350 align:start position:0% agent that has been implemented in Python<00:06:40.200> and<00:06:40.680> his<00:06:40.920> colleague<00:06:41.490> code<00:06:41.790> that<00:06:42.270> is 00:06:42.350 --> 00:06:42.360 align:start position:0% Python and his colleague code that is 00:06:42.360 --> 00:06:43.760 align:start position:0% Python and his colleague code that is implemented<00:06:42.570> in<00:06:42.870> Java<00:06:43.140> this<00:06:43.440> mountain 00:06:43.760 --> 00:06:43.770 align:start position:0% implemented in Java this mountain 00:06:43.770 --> 00:06:46.220 align:start position:0% implemented in Java this mountain simulator<00:06:44.130> and<00:06:44.510> it's<00:06:45.510> essentially<00:06:46.110> going 00:06:46.220 --> 00:06:46.230 align:start position:0% simulator and it's essentially going 00:06:46.230 --> 00:06:48.140 align:start position:0% simulator and it's essentially going through<00:06:46.620> this<00:06:46.740> loop<00:06:46.920> over<00:06:47.280> and<00:06:47.400> over<00:06:47.640> again<00:06:47.730> so 00:06:48.140 --> 00:06:48.150 align:start position:0% through this loop over and over again so 00:06:48.150 --> 00:06:51.170 align:start position:0% through this loop over and over again so I<00:06:48.870> think<00:06:49.140> actions<00:06:49.320> and<00:06:49.800> then<00:06:50.550> collected<00:06:51.030> the 00:06:51.170 --> 00:06:51.180 align:start position:0% I think actions and then collected the 00:06:51.180 --> 00:06:52.670 align:start position:0% I think actions and then collected the information<00:06:51.510> that<00:06:52.080> he<00:06:52.140> gets<00:06:52.350> from<00:06:52.470> the 00:06:52.670 --> 00:06:52.680 align:start position:0% information that he gets from the 00:06:52.680 --> 00:06:56.870 align:start position:0% information that he gets from the simulation<00:06:54.920> and<00:06:55.920> so<00:06:56.130> there's<00:06:56.400> a<00:06:56.460> number<00:06:56.700> of 00:06:56.870 --> 00:06:56.880 align:start position:0% simulation and so there's a number of 00:06:56.880 --> 00:06:58.460 align:start position:0% simulation and so there's a number of different<00:06:57.270> alternates<00:06:57.630> that<00:06:57.810> you<00:06:57.930> can<00:06:58.110> use<00:06:58.260> as 00:06:58.460 --> 00:06:58.470 align:start position:0% different alternates that you can use as 00:06:58.470 --> 00:07:00.170 align:start position:0% different alternates that you can use as I<00:06:58.590> mentioned<00:06:58.950> before<00:06:59.040> like<00:06:59.670> one<00:06:59.820> of<00:06:59.910> the<00:07:00.000> most 00:07:00.170 --> 00:07:00.180 align:start position:0% I mentioned before like one of the most 00:07:00.180 --> 00:07:03.620 align:start position:0% I mentioned before like one of the most famous<00:07:00.330> is<00:07:00.630> the<00:07:00.660> network's<00:07:01.500> many<00:07:02.460> fiddles<00:07:02.820> 103 00:07:03.620 --> 00:07:03.630 align:start position:0% famous is the network's many fiddles 103 00:07:03.630 --> 00:07:06.260 align:start position:0% famous is the network's many fiddles 103 Q<00:07:03.990> learning<00:07:04.140> algorithms<00:07:04.860> policy<00:07:05.280> 3<00:07:05.550> based 00:07:06.260 --> 00:07:06.270 align:start position:0% Q learning algorithms policy 3 based 00:07:06.270 --> 00:07:08.090 align:start position:0% Q learning algorithms policy 3 based algorithms<00:07:06.870> they<00:07:07.440> all<00:07:07.560> play<00:07:07.830> on<00:07:07.980> different 00:07:08.090 --> 00:07:08.100 align:start position:0% algorithms they all play on different 00:07:08.100 --> 00:07:11.360 align:start position:0% algorithms they all play on different sets<00:07:08.550> of<00:07:08.820> statistics<00:07:09.330> and<00:07:10.080> modeling<00:07:10.500> and<00:07:10.680> use 00:07:11.360 --> 00:07:11.370 align:start position:0% sets of statistics and modeling and use 00:07:11.370 --> 00:07:14.930 align:start position:0% sets of statistics and modeling and use of<00:07:11.670> light<00:07:11.910> chains<00:07:12.390> as<00:07:12.720> the<00:07:12.960> back<00:07:13.640> of<00:07:14.640> how<00:07:14.790> they 00:07:14.930 --> 00:07:14.940 align:start position:0% of light chains as the back of how they 00:07:14.940 --> 00:07:17.540 align:start position:0% of light chains as the back of how they calculate<00:07:15.390> different<00:07:16.230> states<00:07:16.470> and<00:07:16.710> wards<00:07:17.070> and 00:07:17.540 --> 00:07:17.550 align:start position:0% calculate different states and wards and 00:07:17.550 --> 00:07:19.460 align:start position:0% calculate different states and wards and so<00:07:18.360> I'm<00:07:18.480> excited<00:07:18.780> that<00:07:18.930> probably<00:07:19.350> the 00:07:19.460 --> 00:07:19.470 align:start position:0% so I'm excited that probably the 00:07:19.470 --> 00:07:22.600 align:start position:0% so I'm excited that probably the simplest<00:07:19.770> one<00:07:20.040> to<00:07:20.070> go<00:07:20.280> for<00:07:20.750> was<00:07:21.750> Q<00:07:22.080> learning 00:07:22.600 --> 00:07:22.610 align:start position:0% simplest one to go for was Q learning 00:07:22.610 --> 00:07:24.830 align:start position:0% simplest one to go for was Q learning and<00:07:23.610> so<00:07:23.760> illustrate<00:07:24.150> and<00:07:24.300> explain<00:07:24.390> a<00:07:24.660> little 00:07:24.830 --> 00:07:24.840 align:start position:0% and so illustrate and explain a little 00:07:24.840 --> 00:07:28.880 align:start position:0% and so illustrate and explain a little bit<00:07:25.020> what<00:07:25.950> Q<00:07:26.220> learning<00:07:26.550> is<00:07:26.750> essentially<00:07:27.750> so<00:07:28.560> we 00:07:28.880 --> 00:07:28.890 align:start position:0% bit what Q learning is essentially so we 00:07:28.890 --> 00:07:30.830 align:start position:0% bit what Q learning is essentially so we use<00:07:29.010> this<00:07:29.250> Python<00:07:30.210> framework<00:07:30.600> that's 00:07:30.830 --> 00:07:30.840 align:start position:0% use this Python framework that's 00:07:30.840 --> 00:07:32.420 align:start position:0% use this Python framework that's developed<00:07:31.230> by<00:07:31.320> an<00:07:31.470> opening<00:07:31.890> eye<00:07:31.980> and<00:07:32.310> it 00:07:32.420 --> 00:07:32.430 align:start position:0% developed by an opening eye and it 00:07:32.430 --> 00:07:33.290 align:start position:0% developed by an opening eye and it allows<00:07:32.640> you<00:07:32.790> to<00:07:32.940> create<00:07:33.090> your<00:07:33.270> own 00:07:33.290 --> 00:07:33.300 align:start position:0% allows you to create your own 00:07:33.300 --> 00:07:35.510 align:start position:0% allows you to create your own environment<00:07:34.080> and<00:07:34.230> interface<00:07:35.070> that<00:07:35.220> you<00:07:35.340> can 00:07:35.510 --> 00:07:35.520 align:start position:0% environment and interface that you can 00:07:35.520 --> 00:07:38.420 align:start position:0% environment and interface that you can use<00:07:35.670> and<00:07:35.970> so<00:07:36.720> we<00:07:36.900> developed<00:07:37.350> that<00:07:37.470> calling<00:07:38.190> the 00:07:38.420 --> 00:07:38.430 align:start position:0% use and so we developed that calling the 00:07:38.430 --> 00:07:41.240 align:start position:0% use and so we developed that calling the Java<00:07:38.640> code<00:07:38.910> and<00:07:39.120> then<00:07:39.960> we<00:07:40.080> tested<00:07:40.410> for<00:07:40.620> men 00:07:41.240 --> 00:07:41.250 align:start position:0% Java code and then we tested for men 00:07:41.250 --> 00:07:43.880 align:start position:0% Java code and then we tested for men so<00:07:41.970> we<00:07:42.600> wanted<00:07:42.810> to<00:07:42.900> see<00:07:43.170> what<00:07:43.470> was<00:07:43.620> the<00:07:43.740> best 00:07:43.880 --> 00:07:43.890 align:start position:0% so we wanted to see what was the best 00:07:43.890 --> 00:07:46.370 align:start position:0% so we wanted to see what was the best strategy<00:07:44.370> when<00:07:44.550> you<00:07:44.580> have<00:07:45.050> 10%<00:07:46.050> of<00:07:46.170> the<00:07:46.230> hash 00:07:46.370 --> 00:07:46.380 align:start position:0% strategy when you have 10% of the hash 00:07:46.380 --> 00:07:50.330 align:start position:0% strategy when you have 10% of the hash around<00:07:46.740> 40%<00:07:47.190> and<00:07:47.670> 60%<00:07:47.850> and<00:07:48.600> then<00:07:49.380> we<00:07:49.740> set<00:07:50.130> some 00:07:50.330 --> 00:07:50.340 align:start position:0% around 40% and 60% and then we set some 00:07:50.340 --> 00:07:51.950 align:start position:0% around 40% and 60% and then we set some hydro<00:07:50.670> parameters<00:07:51.180> that<00:07:51.330> are<00:07:51.420> related<00:07:51.810> to 00:07:51.950 --> 00:07:51.960 align:start position:0% hydro parameters that are related to 00:07:51.960 --> 00:07:54.560 align:start position:0% hydro parameters that are related to these<00:07:52.160> Q<00:07:53.160> tables<00:07:53.700> and<00:07:53.880> how<00:07:53.970> you<00:07:54.030> calculate<00:07:54.480> the 00:07:54.560 --> 00:07:54.570 align:start position:0% these Q tables and how you calculate the 00:07:54.570 --> 00:07:57.280 align:start position:0% these Q tables and how you calculate the reward<00:07:54.840> so<00:07:55.590> you<00:07:55.950> look<00:07:56.100> at<00:07:56.220> the<00:07:56.370> learning<00:07:56.670> rate 00:07:57.280 --> 00:07:57.290 align:start position:0% reward so you look at the learning rate 00:07:57.290 --> 00:07:59.510 align:start position:0% reward so you look at the learning rate which<00:07:58.290> the<00:07:58.470> session<00:07:58.800> is<00:07:58.890> just<00:07:58.920> telling<00:07:59.400> you 00:07:59.510 --> 00:07:59.520 align:start position:0% which the session is just telling you 00:07:59.520 --> 00:08:01.580 align:start position:0% which the session is just telling you how<00:08:00.060> much<00:08:00.180> you<00:08:00.240> love<00:08:00.570> you<00:08:01.020> and<00:08:01.320> new 00:08:01.580 --> 00:08:01.590 align:start position:0% how much you love you and new 00:08:01.590 --> 00:08:03.560 align:start position:0% how much you love you and new information<00:08:01.920> that<00:08:02.250> you<00:08:02.460> receive<00:08:02.820> versus 00:08:03.560 --> 00:08:03.570 align:start position:0% information that you receive versus 00:08:03.570 --> 00:08:05.030 align:start position:0% information that you receive versus information<00:08:03.870> that<00:08:04.230> you've<00:08:04.380> recorded<00:08:04.680> in<00:08:04.920> the 00:08:05.030 --> 00:08:05.040 align:start position:0% information that you've recorded in the 00:08:05.040 --> 00:08:07.760 align:start position:0% information that you've recorded in the past<00:08:05.280> and<00:08:05.720> we<00:08:06.720> also<00:08:06.840> have<00:08:07.080> a<00:08:07.110> discount<00:08:07.590> factor 00:08:07.760 --> 00:08:07.770 align:start position:0% past and we also have a discount factor 00:08:07.770 --> 00:08:10.760 align:start position:0% past and we also have a discount factor which<00:08:08.700> also<00:08:08.880> pertains<00:08:09.480> to<00:08:09.570> the<00:08:09.750> reward<00:08:10.110> you 00:08:10.760 --> 00:08:10.770 align:start position:0% which also pertains to the reward you 00:08:10.770 --> 00:08:13.460 align:start position:0% which also pertains to the reward you want<00:08:11.040> to<00:08:11.160> look<00:08:11.700> at<00:08:11.940> the<00:08:12.210> current<00:08:12.930> reward<00:08:13.350> that 00:08:13.460 --> 00:08:13.470 align:start position:0% want to look at the current reward that 00:08:13.470 --> 00:08:15.740 align:start position:0% want to look at the current reward that you<00:08:13.590> have<00:08:13.740> for<00:08:13.920> an<00:08:14.010> action<00:08:14.310> plus<00:08:14.970> there<00:08:15.600> were 00:08:15.740 --> 00:08:15.750 align:start position:0% you have for an action plus there were 00:08:15.750 --> 00:08:17.690 align:start position:0% you have for an action plus there were work<00:08:15.930> for<00:08:16.140> making<00:08:16.500> it<00:08:16.650> a<00:08:16.740> next<00:08:17.040> option<00:08:17.520> because 00:08:17.690 --> 00:08:17.700 align:start position:0% work for making it a next option because 00:08:17.700 --> 00:08:18.800 align:start position:0% work for making it a next option because you're<00:08:17.850> trying<00:08:18.000> to<00:08:18.090> discover<00:08:18.420> what<00:08:18.660> is<00:08:18.720> the 00:08:18.800 --> 00:08:18.810 align:start position:0% you're trying to discover what is the 00:08:18.810 --> 00:08:21.770 align:start position:0% you're trying to discover what is the most<00:08:19.020> optimal<00:08:19.470> path<00:08:19.710> and<00:08:20.520> so<00:08:20.880> it's<00:08:21.450> something 00:08:21.770 --> 00:08:21.780 align:start position:0% most optimal path and so it's something 00:08:21.780 --> 00:08:23.390 align:start position:0% most optimal path and so it's something that<00:08:21.900> you<00:08:22.050> have<00:08:22.170> to<00:08:22.350> -<00:08:22.530> and<00:08:22.560> in<00:08:22.800> terms<00:08:22.830> of<00:08:23.100> how 00:08:23.390 --> 00:08:23.400 align:start position:0% that you have to - and in terms of how 00:08:23.400 --> 00:08:25.910 align:start position:0% that you have to - and in terms of how do<00:08:23.460> you<00:08:23.610> value<00:08:24.290> the<00:08:25.290> future<00:08:25.620> of<00:08:25.680> the<00:08:25.740> Lord 00:08:25.910 --> 00:08:25.920 align:start position:0% do you value the future of the Lord 00:08:25.920 --> 00:08:28.760 align:start position:0% do you value the future of the Lord because<00:08:26.190> you<00:08:26.310> can<00:08:26.460> think<00:08:26.640> okay<00:08:27.420> I<00:08:27.960> kept<00:08:28.260> mine<00:08:28.440> a 00:08:28.760 --> 00:08:28.770 align:start position:0% because you can think okay I kept mine a 00:08:28.770 --> 00:08:30.410 align:start position:0% because you can think okay I kept mine a lot<00:08:28.950> right<00:08:29.190> now<00:08:29.340> just<00:08:29.880> have<00:08:30.030> it<00:08:30.150> as<00:08:30.240> soon<00:08:30.270> as 00:08:30.410 --> 00:08:30.420 align:start position:0% lot right now just have it as soon as 00:08:30.420 --> 00:08:30.950 align:start position:0% lot right now just have it as soon as possible 00:08:30.950 --> 00:08:30.960 align:start position:0% possible 00:08:30.960 --> 00:08:33.560 align:start position:0% possible and<00:08:31.110> again<00:08:31.350> or<00:08:31.740> if<00:08:31.950> -<00:08:32.099> but<00:08:32.400> if<00:08:32.909> I<00:08:33.060> decide<00:08:33.360> to 00:08:33.560 --> 00:08:33.570 align:start position:0% and again or if - but if I decide to 00:08:33.570 --> 00:08:35.750 align:start position:0% and again or if - but if I decide to wait<00:08:33.780> a<00:08:34.140> couple<00:08:34.979> of<00:08:35.250> what's<00:08:35.430> ahead 00:08:35.750 --> 00:08:35.760 align:start position:0% wait a couple of what's ahead 00:08:35.760 --> 00:08:38.300 align:start position:0% wait a couple of what's ahead publish<00:08:36.270> it<00:08:36.380> there's<00:08:37.380> traffic<00:08:37.710> was<00:08:38.099> buying 00:08:38.300 --> 00:08:38.310 align:start position:0% publish it there's traffic was buying 00:08:38.310 --> 00:08:41.529 align:start position:0% publish it there's traffic was buying some<00:08:38.640> money<00:08:39.050> might<00:08:40.050> cost<00:08:40.349> less<00:08:40.740> because<00:08:41.430> of 00:08:41.529 --> 00:08:41.539 align:start position:0% some money might cost less because of 00:08:41.539 --> 00:08:44.510 align:start position:0% some money might cost less because of inflation<00:08:42.539> or<00:08:42.719> you<00:08:42.870> can<00:08:43.020> just<00:08:43.200> think<00:08:43.349> well<00:08:44.099> I 00:08:44.510 --> 00:08:44.520 align:start position:0% inflation or you can just think well I 00:08:44.520 --> 00:08:47.270 align:start position:0% inflation or you can just think well I lined<00:08:45.330> the<00:08:45.750> four<00:08:46.470> blocks<00:08:46.770> in<00:08:46.980> one<00:08:47.130> minute 00:08:47.270 --> 00:08:47.280 align:start position:0% lined the four blocks in one minute 00:08:47.280 --> 00:08:48.890 align:start position:0% lined the four blocks in one minute that's<00:08:47.550> a<00:08:47.700> negligible<00:08:48.090> it's<00:08:48.360> worth<00:08:48.540> the<00:08:48.690> same 00:08:48.890 --> 00:08:48.900 align:start position:0% that's a negligible it's worth the same 00:08:48.900 --> 00:08:50.960 align:start position:0% that's a negligible it's worth the same so<00:08:49.440> based<00:08:49.680> on<00:08:49.800> what<00:08:49.860> you<00:08:49.980> value<00:08:50.070> most<00:08:50.370> you<00:08:50.850> will 00:08:50.960 --> 00:08:50.970 align:start position:0% so based on what you value most you will 00:08:50.970 --> 00:08:54.200 align:start position:0% so based on what you value most you will tune<00:08:51.210> on<00:08:51.750> this<00:08:51.930> parameter<00:08:52.290> and<00:08:52.710> then<00:08:53.490> finally 00:08:54.200 --> 00:08:54.210 align:start position:0% tune on this parameter and then finally 00:08:54.210 --> 00:08:56.120 align:start position:0% tune on this parameter and then finally you<00:08:54.270> have<00:08:54.330> epsilon<00:08:55.140> which<00:08:55.320> is<00:08:55.500> just<00:08:55.770> related 00:08:56.120 --> 00:08:56.130 align:start position:0% you have epsilon which is just related 00:08:56.130 --> 00:08:58.130 align:start position:0% you have epsilon which is just related to<00:08:56.190> randomness<00:08:56.670> how<00:08:57.090> many<00:08:57.300> choices<00:08:57.510> do<00:08:57.900> you<00:08:57.990> do 00:08:58.130 --> 00:08:58.140 align:start position:0% to randomness how many choices do you do 00:08:58.140 --> 00:09:00.830 align:start position:0% to randomness how many choices do you do that<00:08:58.860> our<00:08:58.980> random<00:08:59.490> person's<00:09:00.210> time<00:09:00.540> a<00:09:00.570> choice 00:09:00.830 --> 00:09:00.840 align:start position:0% that our random person's time a choice 00:09:00.840 --> 00:09:03.590 align:start position:0% that our random person's time a choice if<00:09:01.020> you<00:09:01.200> do<00:09:01.380> that<00:09:01.560> you're<00:09:01.710> maximizing<00:09:02.600> your 00:09:03.590 --> 00:09:03.600 align:start position:0% if you do that you're maximizing your 00:09:03.600 --> 00:09:06.269 align:start position:0% if you do that you're maximizing your strategy<00:09:04.110> by<00:09:04.350> thinking<00:09:04.620> that<00:09:04.740> they 00:09:06.269 --> 00:09:06.279 align:start position:0% strategy by thinking that they 00:09:06.279 --> 00:09:09.489 align:start position:0% strategy by thinking that they and<00:09:07.279> so<00:09:07.579> there's<00:09:07.999> two<00:09:08.059> states<00:09:08.509> as<00:09:08.989> I<00:09:09.169> mentioned 00:09:09.489 --> 00:09:09.499 align:start position:0% and so there's two states as I mentioned 00:09:09.499 --> 00:09:11.499 align:start position:0% and so there's two states as I mentioned before<00:09:09.589> the<00:09:10.459> first<00:09:10.489> one<00:09:10.909> is<00:09:11.029> driven<00:09:11.239> by 00:09:11.499 --> 00:09:11.509 align:start position:0% before the first one is driven by 00:09:11.509 --> 00:09:12.339 align:start position:0% before the first one is driven by Epsilon 00:09:12.339 --> 00:09:12.349 align:start position:0% Epsilon 00:09:12.349 --> 00:09:15.429 align:start position:0% Epsilon where<00:09:12.679> you're<00:09:12.919> just<00:09:13.159> trying<00:09:13.969> to<00:09:14.209> make<00:09:15.049> random 00:09:15.429 --> 00:09:15.439 align:start position:0% where you're just trying to make random 00:09:15.439 --> 00:09:16.929 align:start position:0% where you're just trying to make random choices<00:09:15.709> so<00:09:16.009> the<00:09:16.129> different<00:09:16.429> actions<00:09:16.729> we<00:09:16.789> can 00:09:16.929 --> 00:09:16.939 align:start position:0% choices so the different actions we can 00:09:16.939 --> 00:09:18.669 align:start position:0% choices so the different actions we can take<00:09:17.119> and<00:09:17.389> reporting<00:09:17.869> the<00:09:18.079> results<00:09:18.109> that<00:09:18.559> you 00:09:18.669 --> 00:09:18.679 align:start position:0% take and reporting the results that you 00:09:18.679 --> 00:09:21.849 align:start position:0% take and reporting the results that you get<00:09:18.859> in<00:09:19.159> this<00:09:19.339> table<00:09:19.789> and<00:09:20.049> then<00:09:21.049> this<00:09:21.559> will 00:09:21.849 --> 00:09:21.859 align:start position:0% get in this table and then this will 00:09:21.859 --> 00:09:24.129 align:start position:0% get in this table and then this will allow<00:09:22.189> to<00:09:22.519> generate<00:09:22.939> different<00:09:23.089> valleys<00:09:23.449> so 00:09:24.129 --> 00:09:24.139 align:start position:0% allow to generate different valleys so 00:09:24.139 --> 00:09:26.979 align:start position:0% allow to generate different valleys so that<00:09:24.289> you<00:09:24.469> can<00:09:25.189> reach<00:09:25.339> multiple<00:09:26.029> in<00:09:26.539> deciding 00:09:26.979 --> 00:09:26.989 align:start position:0% that you can reach multiple in deciding 00:09:26.989 --> 00:09:29.559 align:start position:0% that you can reach multiple in deciding what<00:09:27.259> is<00:09:27.289> the<00:09:27.409> best<00:09:27.529> action<00:09:27.919> at<00:09:28.249> each<00:09:28.569> state 00:09:29.559 --> 00:09:29.569 align:start position:0% what is the best action at each state 00:09:29.569 --> 00:09:33.489 align:start position:0% what is the best action at each state that<00:09:29.899> you<00:09:30.169> reach<00:09:30.409> and<00:09:31.059> so<00:09:32.229> then<00:09:33.229> there's<00:09:33.439> a 00:09:33.489 --> 00:09:33.499 align:start position:0% that you reach and so then there's a 00:09:33.499 --> 00:09:35.589 align:start position:0% that you reach and so then there's a theory<00:09:33.859> during<00:09:34.039> this<00:09:34.429> that<00:09:34.699> you<00:09:34.849> saw<00:09:35.269> some 00:09:35.589 --> 00:09:35.599 align:start position:0% theory during this that you saw some 00:09:35.599 --> 00:09:38.319 align:start position:0% theory during this that you saw some people<00:09:35.899> development<00:09:36.199> equations<00:09:36.739> and<00:09:36.999> it<00:09:37.999> just 00:09:38.319 --> 00:09:38.329 align:start position:0% people development equations and it just 00:09:38.329 --> 00:09:40.539 align:start position:0% people development equations and it just said<00:09:38.539> that<00:09:38.809> basically<00:09:39.289> by<00:09:39.619> maximizing<00:09:40.249> at 00:09:40.539 --> 00:09:40.549 align:start position:0% said that basically by maximizing at 00:09:40.549 --> 00:09:43.479 align:start position:0% said that basically by maximizing at each<00:09:40.809> state<00:09:41.809> that<00:09:42.079> you<00:09:42.289> find<00:09:42.529> yourself<00:09:42.919> what 00:09:43.479 --> 00:09:43.489 align:start position:0% each state that you find yourself what 00:09:43.489 --> 00:09:46.809 align:start position:0% each state that you find yourself what is<00:09:43.579> the<00:09:43.669> optimal<00:09:44.179> action<00:09:45.129> you<00:09:46.129> will<00:09:46.249> maximize 00:09:46.809 --> 00:09:46.819 align:start position:0% is the optimal action you will maximize 00:09:46.819 --> 00:09:48.819 align:start position:0% is the optimal action you will maximize for<00:09:47.029> the<00:09:47.119> entirety<00:09:47.479> of<00:09:48.199> your<00:09:48.259> decision 00:09:48.819 --> 00:09:48.829 align:start position:0% for the entirety of your decision 00:09:48.829 --> 00:09:51.429 align:start position:0% for the entirety of your decision process<00:09:48.979> and<00:09:49.669> so<00:09:50.089> you're<00:09:50.269> just<00:09:50.449> looking<00:09:51.139> at 00:09:51.429 --> 00:09:51.439 align:start position:0% process and so you're just looking at 00:09:51.439 --> 00:09:53.349 align:start position:0% process and so you're just looking at maybe<00:09:51.559> two<00:09:51.679> before<00:09:52.069> their<00:09:52.309> reward<00:09:52.609> and<00:09:52.879> then 00:09:53.349 --> 00:09:53.359 align:start position:0% maybe two before their reward and then 00:09:53.359 --> 00:09:55.899 align:start position:0% maybe two before their reward and then you're<00:09:53.689> also<00:09:53.929> adding<00:09:54.409> the<00:09:54.589> value<00:09:55.549> above<00:09:55.729> the 00:09:55.899 --> 00:09:55.909 align:start position:0% you're also adding the value above the 00:09:55.909 --> 00:09:58.239 align:start position:0% you're also adding the value above the future<00:09:56.269> award<00:09:56.629> of<00:09:56.779> the<00:09:56.899> next<00:09:57.289> state<00:09:57.619> in<00:09:57.859> action 00:09:58.239 --> 00:09:58.249 align:start position:0% future award of the next state in action 00:09:58.249 --> 00:10:00.609 align:start position:0% future award of the next state in action but<00:09:58.489> with<00:09:58.729> an<00:09:58.849> aesthetic<00:09:59.359> factor<00:09:59.749> so<00:10:00.379> that's 00:10:00.609 --> 00:10:00.619 align:start position:0% but with an aesthetic factor so that's 00:10:00.619 --> 00:10:01.899 align:start position:0% but with an aesthetic factor so that's what<00:10:00.769> I<00:10:00.799> was<00:10:00.919> talking<00:10:01.099> about<00:10:01.279> before<00:10:01.549> that 00:10:01.899 --> 00:10:01.909 align:start position:0% what I was talking about before that 00:10:01.909 --> 00:10:05.679 align:start position:0% what I was talking about before that zero<00:10:02.149> point<00:10:02.209> about<00:10:02.719> you<00:10:04.179> and<00:10:05.179> so<00:10:05.389> does<00:10:05.509> the 00:10:05.679 --> 00:10:05.689 align:start position:0% zero point about you and so does the 00:10:05.689 --> 00:10:06.789 align:start position:0% zero point about you and so does the reason<00:10:05.899> is<00:10:06.019> filling<00:10:06.319> with<00:10:06.349> more<00:10:06.619> concrete 00:10:06.789 --> 00:10:06.799 align:start position:0% reason is filling with more concrete 00:10:06.799 --> 00:10:09.369 align:start position:0% reason is filling with more concrete because<00:10:07.339> it<00:10:07.459> can<00:10:07.579> be<00:10:07.669> very<00:10:07.699> abstract<00:10:08.269> we<00:10:09.199> can 00:10:09.369 --> 00:10:09.379 align:start position:0% because it can be very abstract we can 00:10:09.379 --> 00:10:11.109 align:start position:0% because it can be very abstract we can think<00:10:09.619> about<00:10:09.709> this<00:10:10.009> Q<00:10:10.309> matrix<00:10:10.459> being 00:10:11.109 --> 00:10:11.119 align:start position:0% think about this Q matrix being 00:10:11.119 --> 00:10:13.689 align:start position:0% think about this Q matrix being something<00:10:11.389> like<00:10:11.629> this<00:10:11.809> where<00:10:12.199> you<00:10:12.379> have<00:10:13.369> four 00:10:13.689 --> 00:10:13.699 align:start position:0% something like this where you have four 00:10:13.699 --> 00:10:15.879 align:start position:0% something like this where you have four different<00:10:13.909> actions<00:10:14.899> so<00:10:15.139> you<00:10:15.199> can<00:10:15.439> hold<00:10:15.709> your 00:10:15.879 --> 00:10:15.889 align:start position:0% different actions so you can hold your 00:10:15.889 --> 00:10:17.949 align:start position:0% different actions so you can hold your block<00:10:16.159> and<00:10:16.459> not<00:10:16.729> publish<00:10:17.269> it<00:10:17.419> just<00:10:17.599> keep<00:10:17.749> it<00:10:17.869> in 00:10:17.949 --> 00:10:17.959 align:start position:0% block and not publish it just keep it in 00:10:17.959 --> 00:10:21.129 align:start position:0% block and not publish it just keep it in the<00:10:18.049> branching<00:10:18.879> you<00:10:19.879> can<00:10:20.119> polish<00:10:20.689> one<00:10:20.869> block 00:10:21.129 --> 00:10:21.139 align:start position:0% the branching you can polish one block 00:10:21.139 --> 00:10:23.079 align:start position:0% the branching you can polish one block two<00:10:21.439> or<00:10:21.619> three<00:10:21.859> as<00:10:22.099> I<00:10:22.279> mentioned<00:10:22.639> before<00:10:22.699> this 00:10:23.079 --> 00:10:23.089 align:start position:0% two or three as I mentioned before this 00:10:23.089 --> 00:10:23.790 align:start position:0% two or three as I mentioned before this is<00:10:23.269> the<00:10:23.329> difference 00:10:23.790 --> 00:10:23.800 align:start position:0% is the difference 00:10:23.800 --> 00:10:26.370 align:start position:0% is the difference some<00:10:23.980> actions<00:10:24.459> you<00:10:24.580> can<00:10:24.730> take<00:10:24.839> and<00:10:25.839> then<00:10:26.110> you 00:10:26.370 --> 00:10:26.380 align:start position:0% some actions you can take and then you 00:10:26.380 --> 00:10:28.769 align:start position:0% some actions you can take and then you can<00:10:26.589> hold<00:10:26.800> one<00:10:27.130> two<00:10:27.339> three<00:10:27.640> four<00:10:27.970> and<00:10:28.209> more<00:10:28.540> but 00:10:28.769 --> 00:10:28.779 align:start position:0% can hold one two three four and more but 00:10:28.779 --> 00:10:30.540 align:start position:0% can hold one two three four and more but for<00:10:28.990> simplicity<00:10:29.320> in<00:10:29.620> this<00:10:29.769> example<00:10:30.070> and 00:10:30.540 --> 00:10:30.550 align:start position:0% for simplicity in this example and 00:10:30.550 --> 00:10:33.150 align:start position:0% for simplicity in this example and that's<00:10:31.000> it<00:10:31.149> three<00:10:31.360> and<00:10:31.690> so<00:10:32.410> when<00:10:32.649> you<00:10:32.740> start<00:10:33.010> to 00:10:33.150 --> 00:10:33.160 align:start position:0% that's it three and so when you start to 00:10:33.160 --> 00:10:35.009 align:start position:0% that's it three and so when you start to initialize<00:10:33.640> it<00:10:33.670> and<00:10:34.000> all<00:10:34.450> the<00:10:34.570> values<00:10:34.839> aren't 00:10:35.009 --> 00:10:35.019 align:start position:0% initialize it and all the values aren't 00:10:35.019 --> 00:10:38.100 align:start position:0% initialize it and all the values aren't zero<00:10:35.440> and<00:10:35.940> as<00:10:36.940> you<00:10:37.149> keep<00:10:37.420> without<00:10:37.720> going 00:10:38.100 --> 00:10:38.110 align:start position:0% zero and as you keep without going 00:10:38.110 --> 00:10:39.420 align:start position:0% zero and as you keep without going around<00:10:38.320> them<00:10:38.470> the<00:10:38.560> connections<00:10:39.250> you'll 00:10:39.420 --> 00:10:39.430 align:start position:0% around them the connections you'll 00:10:39.430 --> 00:10:41.130 align:start position:0% around them the connections you'll change<00:10:39.670> the<00:10:39.760> values<00:10:40.149> so<00:10:40.570> let's<00:10:40.810> just<00:10:40.990> walk 00:10:41.130 --> 00:10:41.140 align:start position:0% change the values so let's just walk 00:10:41.140 --> 00:10:43.440 align:start position:0% change the values so let's just walk through<00:10:41.470> one<00:10:42.220> of<00:10:42.250> these<00:10:42.459> examples<00:10:42.850> so<00:10:43.269> let's 00:10:43.440 --> 00:10:43.450 align:start position:0% through one of these examples so let's 00:10:43.450 --> 00:10:45.300 align:start position:0% through one of these examples so let's say<00:10:43.600> that<00:10:43.630> you<00:10:43.839> start<00:10:44.079> and<00:10:44.529> you<00:10:44.589> start<00:10:44.890> in<00:10:45.040> that 00:10:45.300 --> 00:10:45.310 align:start position:0% say that you start and you start in that 00:10:45.310 --> 00:10:48.030 align:start position:0% say that you start and you start in that chain<00:10:45.550> you<00:10:45.579> decide<00:10:45.970> to<00:10:46.089> hold<00:10:46.329> and<00:10:46.829> so<00:10:47.829> you're 00:10:48.030 --> 00:10:48.040 align:start position:0% chain you decide to hold and so you're 00:10:48.040 --> 00:10:49.410 align:start position:0% chain you decide to hold and so you're gonna<00:10:48.160> calculate<00:10:48.430> what<00:10:48.850> happens<00:10:48.880> when<00:10:49.360> you 00:10:49.410 --> 00:10:49.420 align:start position:0% gonna calculate what happens when you 00:10:49.420 --> 00:10:51.990 align:start position:0% gonna calculate what happens when you hold<00:10:49.600> and<00:10:49.959> then<00:10:50.380> maybe<00:10:50.860> the<00:10:51.010> next<00:10:51.160> step<00:10:51.430> you 00:10:51.990 --> 00:10:52.000 align:start position:0% hold and then maybe the next step you 00:10:52.000 --> 00:10:54.240 align:start position:0% hold and then maybe the next step you want<00:10:52.149> to<00:10:52.240> publish<00:10:52.390> it<00:10:52.720> so<00:10:52.959> that<00:10:53.800> will<00:10:53.920> generate 00:10:54.240 --> 00:10:54.250 align:start position:0% want to publish it so that will generate 00:10:54.250 --> 00:10:56.940 align:start position:0% want to publish it so that will generate a<00:10:54.370> reward<00:10:54.730> let's<00:10:55.089> say<00:10:55.240> to<00:10:55.300> me<00:10:55.779> and<00:10:56.680> so<00:10:56.920> you 00:10:56.940 --> 00:10:56.950 align:start position:0% a reward let's say to me and so you 00:10:56.950 --> 00:10:58.889 align:start position:0% a reward let's say to me and so you continue<00:10:57.310> doing<00:10:57.670> those<00:10:57.970> few<00:10:58.240> random<00:10:58.450> paths 00:10:58.889 --> 00:10:58.899 align:start position:0% continue doing those few random paths 00:10:58.899 --> 00:11:02.280 align:start position:0% continue doing those few random paths until<00:10:59.860> you<00:11:00.060> generated<00:11:01.060> enough<00:11:01.300> values<00:11:01.690> this 00:11:02.280 --> 00:11:02.290 align:start position:0% until you generated enough values this 00:11:02.290 --> 00:11:04.139 align:start position:0% until you generated enough values this is<00:11:02.440> like<00:11:02.620> random<00:11:03.279> values<00:11:03.550> not<00:11:03.820> the<00:11:03.940> actual 00:11:04.139 --> 00:11:04.149 align:start position:0% is like random values not the actual 00:11:04.149 --> 00:11:06.420 align:start position:0% is like random values not the actual ones<00:11:04.450> but<00:11:05.170> just<00:11:05.380> to<00:11:05.620> give<00:11:05.769> you<00:11:05.800> any<00:11:06.040> cable<00:11:06.339> what 00:11:06.420 --> 00:11:06.430 align:start position:0% ones but just to give you any cable what 00:11:06.430 --> 00:11:09.090 align:start position:0% ones but just to give you any cable what happens<00:11:06.820> and<00:11:07.089> then<00:11:07.410> after<00:11:08.410> a<00:11:08.440> certain<00:11:08.560> point 00:11:09.090 --> 00:11:09.100 align:start position:0% happens and then after a certain point 00:11:09.100 --> 00:11:11.400 align:start position:0% happens and then after a certain point what<00:11:09.640> you<00:11:09.760> start<00:11:10.029> doing<00:11:10.209> is<00:11:10.450> that<00:11:10.600> okay<00:11:10.959> I<00:11:11.140> find 00:11:11.400 --> 00:11:11.410 align:start position:0% what you start doing is that okay I find 00:11:11.410 --> 00:11:14.250 align:start position:0% what you start doing is that okay I find myself<00:11:11.560> that<00:11:11.980> I<00:11:12.010> speak<00:11:12.310> about<00:11:12.579> holy<00:11:13.000> -<00:11:13.420> so<00:11:14.230> I 00:11:14.250 --> 00:11:14.260 align:start position:0% myself that I speak about holy - so I 00:11:14.260 --> 00:11:16.590 align:start position:0% myself that I speak about holy - so I have<00:11:14.470> two<00:11:14.620> blocks<00:11:14.920> wise<00:11:15.250> my<00:11:15.459> next<00:11:15.790> next<00:11:16.120> action 00:11:16.590 --> 00:11:16.600 align:start position:0% have two blocks wise my next next action 00:11:16.600 --> 00:11:18.780 align:start position:0% have two blocks wise my next next action sure<00:11:17.260> polish<00:11:17.709> them<00:11:17.890> or<00:11:18.130> should<00:11:18.310> I<00:11:18.459> call<00:11:18.670> them 00:11:18.780 --> 00:11:18.790 align:start position:0% sure polish them or should I call them 00:11:18.790 --> 00:11:21.600 align:start position:0% sure polish them or should I call them and<00:11:19.390> so<00:11:19.570> you<00:11:19.630> think<00:11:19.839> the<00:11:19.990> maximum<00:11:20.399> value<00:11:21.399> for 00:11:21.600 --> 00:11:21.610 align:start position:0% and so you think the maximum value for 00:11:21.610 --> 00:11:23.730 align:start position:0% and so you think the maximum value for each<00:11:21.820> action<00:11:22.089> if<00:11:22.720> that's<00:11:22.959> how<00:11:23.140> you<00:11:23.200> can<00:11:23.380> find<00:11:23.589> a 00:11:23.730 --> 00:11:23.740 align:start position:0% each action if that's how you can find a 00:11:23.740 --> 00:11:27.889 align:start position:0% each action if that's how you can find a possible<00:11:24.279> path<00:11:24.540> and<00:11:25.860> so<00:11:26.860> I<00:11:27.190> don't<00:11:27.610> mention 00:11:27.889 --> 00:11:27.899 align:start position:0% possible path and so I don't mention 00:11:27.899 --> 00:11:31.110 align:start position:0% possible path and so I don't mention briefly<00:11:28.899> there's<00:11:29.410> a<00:11:29.500> the<00:11:30.399> marker<00:11:30.820> change 00:11:31.110 --> 00:11:31.120 align:start position:0% briefly there's a the marker change 00:11:31.120 --> 00:11:33.660 align:start position:0% briefly there's a the marker change crosses<00:11:31.570> bet<00:11:31.750> that's<00:11:32.500> all<00:11:32.800> this<00:11:33.010> information 00:11:33.660 --> 00:11:33.670 align:start position:0% crosses bet that's all this information 00:11:33.670 --> 00:11:35.519 align:start position:0% crosses bet that's all this information and<00:11:33.820> ensures<00:11:34.120> of<00:11:34.300> this<00:11:34.420> equation<00:11:34.959> holds<00:11:35.200> true 00:11:35.519 --> 00:11:35.529 align:start position:0% and ensures of this equation holds true 00:11:35.529 --> 00:11:38.010 align:start position:0% and ensures of this equation holds true and<00:11:35.829> it's<00:11:36.399> just<00:11:36.670> an<00:11:36.910> ATS<00:11:37.390> so<00:11:37.600> that<00:11:37.720> you<00:11:37.839> can 00:11:38.010 --> 00:11:38.020 align:start position:0% and it's just an ATS so that you can 00:11:38.020 --> 00:11:39.750 align:start position:0% and it's just an ATS so that you can also<00:11:38.350> visualize<00:11:38.890> this<00:11:39.339> in 00:11:39.750 --> 00:11:39.760 align:start position:0% also visualize this in 00:11:39.760 --> 00:11:41.879 align:start position:0% also visualize this in wait<00:11:40.029> you<00:11:40.660> can<00:11:40.870> have<00:11:41.079> a<00:11:41.110> chain<00:11:41.440> where<00:11:41.709> you're 00:11:41.879 --> 00:11:41.889 align:start position:0% wait you can have a chain where you're 00:11:41.889 --> 00:11:43.829 align:start position:0% wait you can have a chain where you're mining<00:11:42.699> and<00:11:42.970> then<00:11:43.089> you<00:11:43.209> can<00:11:43.329> serve<00:11:43.510> you<00:11:43.630> won't 00:11:43.829 --> 00:11:43.839 align:start position:0% mining and then you can serve you won't 00:11:43.839 --> 00:11:45.569 align:start position:0% mining and then you can serve you won't write<00:11:44.050> a<00:11:44.079> chain<00:11:44.380> and<00:11:44.649> keep<00:11:44.800> switching<00:11:45.250> between 00:11:45.569 --> 00:11:45.579 align:start position:0% write a chain and keep switching between 00:11:45.579 --> 00:11:48.300 align:start position:0% write a chain and keep switching between one<00:11:46.000> and<00:11:46.029> the<00:11:46.149> other<00:11:46.209> and<00:11:46.690> that<00:11:47.310> essentially 00:11:48.300 --> 00:11:48.310 align:start position:0% one and the other and that essentially 00:11:48.310 --> 00:11:49.949 align:start position:0% one and the other and that essentially because<00:11:48.610> of<00:11:48.639> the<00:11:48.790> properties<00:11:49.209> of<00:11:49.240> our<00:11:49.420> chain 00:11:49.949 --> 00:11:49.959 align:start position:0% because of the properties of our chain 00:11:49.959 --> 00:11:51.870 align:start position:0% because of the properties of our chain use<00:11:50.170> you<00:11:50.500> reach<00:11:50.709> the<00:11:50.980> steady<00:11:51.279> state<00:11:51.639> and 00:11:51.870 --> 00:11:51.880 align:start position:0% use you reach the steady state and 00:11:51.880 --> 00:11:53.879 align:start position:0% use you reach the steady state and that's<00:11:52.120> why<00:11:52.300> this<00:11:52.360> come<00:11:52.839> table<00:11:53.380> is<00:11:53.470> useful 00:11:53.879 --> 00:11:53.889 align:start position:0% that's why this come table is useful 00:11:53.889 --> 00:11:55.860 align:start position:0% that's why this come table is useful because<00:11:53.980> you<00:11:54.430> can<00:11:54.820> assume<00:11:55.209> that<00:11:55.240> to<00:11:55.600> exert<00:11:55.839> a 00:11:55.860 --> 00:11:55.870 align:start position:0% because you can assume that to exert a 00:11:55.870 --> 00:11:57.810 align:start position:0% because you can assume that to exert a number<00:11:56.170> of<00:11:56.260> steps<00:11:56.470> now<00:11:57.040> you've<00:11:57.279> reached<00:11:57.550> the 00:11:57.810 --> 00:11:57.820 align:start position:0% number of steps now you've reached the 00:11:57.820 --> 00:11:59.910 align:start position:0% number of steps now you've reached the same<00:11:58.029> and<00:11:58.389> you<00:11:58.570> were<00:11:58.690> off<00:11:58.870> to<00:11:58.990> infinity<00:11:59.110> and<00:11:59.620> so 00:11:59.910 --> 00:11:59.920 align:start position:0% same and you were off to infinity and so 00:11:59.920 --> 00:12:03.290 align:start position:0% same and you were off to infinity and so that's<00:12:00.130> how<00:12:00.310> you<00:12:00.399> can<00:12:00.579> maximize<00:12:01.420> the<00:12:01.839> values 00:12:03.290 --> 00:12:03.300 align:start position:0% that's how you can maximize the values 00:12:03.300 --> 00:12:05.400 align:start position:0% that's how you can maximize the values and<00:12:04.300> so<00:12:04.480> there's<00:12:04.660> a<00:12:04.750> lot<00:12:04.870> of<00:12:04.959> attention 00:12:05.400 --> 00:12:05.410 align:start position:0% and so there's a lot of attention 00:12:05.410 --> 00:12:07.350 align:start position:0% and so there's a lot of attention challenges<00:12:06.010> we<00:12:06.610> can<00:12:06.790> get<00:12:06.910> this<00:12:07.000> approach 00:12:07.350 --> 00:12:07.360 align:start position:0% challenges we can get this approach 00:12:07.360 --> 00:12:09.090 align:start position:0% challenges we can get this approach because<00:12:08.110> you<00:12:08.199> need<00:12:08.350> to<00:12:08.440> understand<00:12:08.889> what's 00:12:09.090 --> 00:12:09.100 align:start position:0% because you need to understand what's 00:12:09.100 --> 00:12:11.069 align:start position:0% because you need to understand what's the<00:12:09.370> best<00:12:09.579> data<00:12:09.790> that<00:12:09.880> we<00:12:10.089> eat<00:12:10.600> you<00:12:10.810> need<00:12:10.930> to 00:12:11.069 --> 00:12:11.079 align:start position:0% the best data that we eat you need to 00:12:11.079 --> 00:12:12.840 align:start position:0% the best data that we eat you need to test<00:12:11.350> for<00:12:11.410> all<00:12:11.620> these<00:12:11.740> hyperparameters<00:12:12.370> and 00:12:12.840 --> 00:12:12.850 align:start position:0% test for all these hyperparameters and 00:12:12.850 --> 00:12:14.970 align:start position:0% test for all these hyperparameters and keep<00:12:12.970> changing<00:12:13.420> them<00:12:13.600> until<00:12:14.410> you<00:12:14.589> can<00:12:14.649> find 00:12:14.970 --> 00:12:14.980 align:start position:0% keep changing them until you can find 00:12:14.980 --> 00:12:17.550 align:start position:0% keep changing them until you can find something<00:12:15.459> optimal<00:12:16.029> and<00:12:16.300> it's<00:12:17.290> also<00:12:17.440> a 00:12:17.550 --> 00:12:17.560 align:start position:0% something optimal and it's also a 00:12:17.560 --> 00:12:19.110 align:start position:0% something optimal and it's also a consideration<00:12:17.920> of<00:12:18.190> how<00:12:18.459> do<00:12:18.519> you<00:12:18.699> build<00:12:18.880> this 00:12:19.110 --> 00:12:19.120 align:start position:0% consideration of how do you build this 00:12:19.120 --> 00:12:20.790 align:start position:0% consideration of how do you build this what<00:12:20.019> are<00:12:20.110> the<00:12:20.199> behaviors<00:12:20.620> that<00:12:20.649> you 00:12:20.790 --> 00:12:20.800 align:start position:0% what are the behaviors that you 00:12:20.800 --> 00:12:22.769 align:start position:0% what are the behaviors that you incentivize<00:12:21.579> how<00:12:21.910> do<00:12:21.970> you<00:12:22.089> to<00:12:22.180> find<00:12:22.420> something 00:12:22.769 --> 00:12:22.779 align:start position:0% incentivize how do you to find something 00:12:22.779 --> 00:12:25.470 align:start position:0% incentivize how do you to find something that's<00:12:22.870> visiting<00:12:23.380> and<00:12:23.680> so<00:12:24.610> on<00:12:24.790> there's<00:12:25.209> a<00:12:25.269> lot 00:12:25.470 --> 00:12:25.480 align:start position:0% that's visiting and so on there's a lot 00:12:25.480 --> 00:12:28.019 align:start position:0% that's visiting and so on there's a lot of<00:12:25.630> things<00:12:25.839> that<00:12:25.959> you<00:12:26.079> can<00:12:26.430> buy<00:12:27.430> challenges 00:12:28.019 --> 00:12:28.029 align:start position:0% of things that you can buy challenges 00:12:28.029 --> 00:12:31.500 align:start position:0% of things that you can buy challenges with<00:12:28.240> and<00:12:28.980> so<00:12:29.980> it<00:12:30.160> just<00:12:30.190> to<00:12:30.519> go<00:12:31.029> back<00:12:31.060> to<00:12:31.389> where 00:12:31.500 --> 00:12:31.510 align:start position:0% with and so it just to go back to where 00:12:31.510 --> 00:12:34.079 align:start position:0% with and so it just to go back to where you<00:12:31.630> can<00:12:31.779> look<00:12:32.019> through<00:12:32.199> some<00:12:32.440> work<00:12:32.699> we<00:12:33.699> took<00:12:33.880> a 00:12:34.079 --> 00:12:34.089 align:start position:0% you can look through some work we took a 00:12:34.089 --> 00:12:35.670 align:start position:0% you can look through some work we took a survey<00:12:34.269> shion's<00:12:34.660> we<00:12:34.839> were<00:12:34.930> distracted<00:12:35.290> what 00:12:35.670 --> 00:12:35.680 align:start position:0% survey shion's we were distracted what 00:12:35.680 --> 00:12:38.220 align:start position:0% survey shion's we were distracted what is<00:12:35.800> the<00:12:36.300> number<00:12:37.300> of<00:12:37.329> blocks<00:12:37.690> that<00:12:37.959> you've 00:12:38.220 --> 00:12:38.230 align:start position:0% is the number of blocks that you've 00:12:38.230 --> 00:12:40.680 align:start position:0% is the number of blocks that you've published<00:12:38.860> in<00:12:39.279> the<00:12:39.370> public<00:12:39.699> chain<00:12:39.940> private 00:12:40.680 --> 00:12:40.690 align:start position:0% published in the public chain private 00:12:40.690 --> 00:12:42.840 align:start position:0% published in the public chain private chain<00:12:40.870> I'm<00:12:41.589> keeping<00:12:41.889> track<00:12:41.920> of<00:12:42.160> all<00:12:42.370> calls<00:12:42.639> and 00:12:42.840 --> 00:12:42.850 align:start position:0% chain I'm keeping track of all calls and 00:12:42.850 --> 00:12:45.780 align:start position:0% chain I'm keeping track of all calls and forts<00:12:43.240> and<00:12:43.420> then<00:12:43.980> letting<00:12:44.980> it<00:12:45.130> know<00:12:45.190> so<00:12:45.490> we<00:12:45.639> can 00:12:45.780 --> 00:12:45.790 align:start position:0% forts and then letting it know so we can 00:12:45.790 --> 00:12:51.019 align:start position:0% forts and then letting it know so we can make<00:12:45.940> the<00:12:46.060> best<00:12:46.240> actions<00:12:47.250> and<00:12:48.510> so<00:12:49.510> we<00:12:49.839> ran<00:12:50.050> this 00:12:51.019 --> 00:12:51.029 align:start position:0% make the best actions and so we ran this 00:12:51.029 --> 00:12:52.889 align:start position:0% make the best actions and so we ran this similar<00:12:52.029> to<00:12:52.180> what<00:12:52.300> we<00:12:52.420> were<00:12:52.480> in<00:12:52.600> in<00:12:52.750> the 00:12:52.889 --> 00:12:52.899 align:start position:0% similar to what we were in in the 00:12:52.899 --> 00:12:55.350 align:start position:0% similar to what we were in in the self-contained<00:12:53.380> simulation<00:12:54.310> lying<00:12:55.000> on<00:12:55.180> its 00:12:55.350 --> 00:12:55.360 align:start position:0% self-contained simulation lying on its 00:12:55.360 --> 00:12:56.620 align:start position:0% self-contained simulation lying on its miners<00:12:55.690> and<00:12:55.839> their<00:12:55.930> one<00:12:56.139> is 00:12:56.620 --> 00:12:56.630 align:start position:0% miners and their one is 00:12:56.630 --> 00:12:59.100 align:start position:0% miners and their one is my<00:12:56.750> nerdiness<00:12:57.110> top<00:12:57.500> date<00:12:57.770> through<00:12:58.670> the 00:12:59.100 --> 00:12:59.110 align:start position:0% my nerdiness top date through the 00:12:59.110 --> 00:13:01.660 align:start position:0% my nerdiness top date through the reinforcement<00:13:00.110> learning<00:13:00.260> agent<00:13:00.860> so<00:13:01.340> we 00:13:01.660 --> 00:13:01.670 align:start position:0% reinforcement learning agent so we 00:13:01.670 --> 00:13:04.450 align:start position:0% reinforcement learning agent so we measured<00:13:02.000> it<00:13:02.120> at<00:13:02.270> 10<00:13:02.990> 40<00:13:03.620> X<00:13:03.770> 60<00:13:03.830> percent 00:13:04.450 --> 00:13:04.460 align:start position:0% measured it at 10 40 X 60 percent 00:13:04.460 --> 00:13:06.760 align:start position:0% measured it at 10 40 X 60 percent hashing<00:13:05.000> power<00:13:05.030> and<00:13:05.540> we<00:13:06.260> collected<00:13:06.650> the 00:13:06.760 --> 00:13:06.770 align:start position:0% hashing power and we collected the 00:13:06.770 --> 00:13:11.710 align:start position:0% hashing power and we collected the results<00:13:09.310> and<00:13:10.310> so<00:13:10.520> as<00:13:10.700> I<00:13:10.730> mentioned<00:13:11.150> before<00:13:11.210> we 00:13:11.710 --> 00:13:11.720 align:start position:0% results and so as I mentioned before we 00:13:11.720 --> 00:13:13.720 align:start position:0% results and so as I mentioned before we use<00:13:11.870> the<00:13:12.050> simulator<00:13:12.500> and<00:13:12.590> John<00:13:12.950> we<00:13:13.340> use<00:13:13.550> the 00:13:13.720 --> 00:13:13.730 align:start position:0% use the simulator and John we use the 00:13:13.730 --> 00:13:16.870 align:start position:0% use the simulator and John we use the spray<00:13:13.940> wearing<00:13:14.270> company<00:13:14.690> I<00:13:14.720> -<00:13:15.460> so<00:13:16.460> that's<00:13:16.730> a 00:13:16.870 --> 00:13:16.880 align:start position:0% spray wearing company I - so that's a 00:13:16.880 --> 00:13:18.610 align:start position:0% spray wearing company I - so that's a big<00:13:17.060> publishing<00:13:17.360> memory<00:13:17.690> libraries<00:13:18.110> and<00:13:18.440> then 00:13:18.610 --> 00:13:18.620 align:start position:0% big publishing memory libraries and then 00:13:18.620 --> 00:13:22.650 align:start position:0% big publishing memory libraries and then I<00:13:18.830> ran<00:13:19.100> all<00:13:19.400> of<00:13:19.430> these<00:13:19.700> tests<00:13:20.420> in<00:13:20.690> a<00:13:20.810> double<00:13:21.170> us 00:13:22.650 --> 00:13:22.660 align:start position:0% I ran all of these tests in a double us 00:13:22.660 --> 00:13:25.750 align:start position:0% I ran all of these tests in a double us so<00:13:23.660> that<00:13:23.870> we<00:13:24.050> could<00:13:24.230> test<00:13:24.590> it<00:13:24.770> collect<00:13:25.160> all<00:13:25.580> the 00:13:25.750 --> 00:13:25.760 align:start position:0% so that we could test it collect all the 00:13:25.760 --> 00:13:29.290 align:start position:0% so that we could test it collect all the data<00:13:26.110> and<00:13:27.110> so<00:13:27.470> this<00:13:28.100> is<00:13:28.250> just<00:13:28.400> an<00:13:28.850> overview<00:13:29.120> of 00:13:29.290 --> 00:13:29.300 align:start position:0% data and so this is just an overview of 00:13:29.300 --> 00:13:30.580 align:start position:0% data and so this is just an overview of what<00:13:29.420> we<00:13:29.510> were<00:13:29.630> doing<00:13:29.810> and<00:13:30.170> the<00:13:30.260> different 00:13:30.580 --> 00:13:30.590 align:start position:0% what we were doing and the different 00:13:30.590 --> 00:13:32.410 align:start position:0% what we were doing and the different hashing<00:13:30.920> rates<00:13:31.160> we<00:13:31.760> collected<00:13:32.240> the<00:13:32.360> reward 00:13:32.410 --> 00:13:32.420 align:start position:0% hashing rates we collected the reward 00:13:32.420 --> 00:13:35.050 align:start position:0% hashing rates we collected the reward only<00:13:33.140> a<00:13:33.260> total<00:13:33.710> of<00:13:33.830> one<00:13:34.100> hour<00:13:34.220> remaining<00:13:34.700> we 00:13:35.050 --> 00:13:35.060 align:start position:0% only a total of one hour remaining we 00:13:35.060 --> 00:13:38.790 align:start position:0% only a total of one hour remaining we also<00:13:35.240> collected<00:13:35.720> over<00:13:36.320> the<00:13:36.500> total<00:13:36.830> of 00:13:38.790 --> 00:13:38.800 align:start position:0% also collected over the total of 00:13:38.800 --> 00:13:42.100 align:start position:0% also collected over the total of episodes<00:13:39.800> or<00:13:40.190> simulation<00:13:40.970> so<00:13:41.180> essentially 00:13:42.100 --> 00:13:42.110 align:start position:0% episodes or simulation so essentially 00:13:42.110 --> 00:13:44.500 align:start position:0% episodes or simulation so essentially all<00:13:42.200> raw<00:13:42.560> element<00:13:43.040> my<00:13:43.190> agent<00:13:43.910> run<00:13:44.060> for<00:13:44.300> an<00:13:44.360> hour 00:13:44.500 --> 00:13:44.510 align:start position:0% all raw element my agent run for an hour 00:13:44.510 --> 00:13:48.100 align:start position:0% all raw element my agent run for an hour or<00:13:45.200> until<00:13:45.230> it<00:13:45.710> has<00:13:45.890> mine<00:13:46.220> a<00:13:46.850> thousand<00:13:47.840> watts 00:13:48.100 --> 00:13:48.110 align:start position:0% or until it has mine a thousand watts 00:13:48.110 --> 00:13:50.250 align:start position:0% or until it has mine a thousand watts and<00:13:48.470> record<00:13:48.860> that<00:13:49.040> behavior<00:13:49.460> and<00:13:49.700> I<00:13:49.820> run<00:13:50.060> it 00:13:50.250 --> 00:13:50.260 align:start position:0% and record that behavior and I run it 00:13:50.260 --> 00:13:53.890 align:start position:0% and record that behavior and I run it for<00:13:51.260> ten<00:13:51.740> thousand<00:13:52.160> episodes<00:13:52.820> while<00:13:53.690> you're 00:13:53.890 --> 00:13:53.900 align:start position:0% for ten thousand episodes while you're 00:13:53.900 --> 00:13:56.500 align:start position:0% for ten thousand episodes while you're changing<00:13:54.140> the<00:13:54.920> actions<00:13:55.700> that<00:13:55.760> you<00:13:55.940> select<00:13:56.270> and 00:13:56.500 --> 00:13:56.510 align:start position:0% changing the actions that you select and 00:13:56.510 --> 00:13:57.970 align:start position:0% changing the actions that you select and we<00:13:57.020> also<00:13:57.140> measure<00:13:57.440> for<00:13:57.740> three<00:13:57.890> different 00:13:57.970 --> 00:13:57.980 align:start position:0% we also measure for three different 00:13:57.980 --> 00:13:59.320 align:start position:0% we also measure for three different strategies<00:13:58.580> we<00:13:58.730> just<00:13:58.910> wanted<00:13:59.060> to<00:13:59.150> make<00:13:59.270> sure 00:13:59.320 --> 00:13:59.330 align:start position:0% strategies we just wanted to make sure 00:13:59.330 --> 00:14:01.960 align:start position:0% strategies we just wanted to make sure that<00:13:59.800> we<00:14:00.800> weren't<00:14:01.220> performing<00:14:01.520> better<00:14:01.820> than 00:14:01.960 --> 00:14:01.970 align:start position:0% that we weren't performing better than 00:14:01.970 --> 00:14:04.750 align:start position:0% that we weren't performing better than if<00:14:02.270> you<00:14:02.390> were<00:14:02.720> just<00:14:02.750> mine<00:14:03.230> honestly<00:14:03.860> or<00:14:04.400> if 00:14:04.750 --> 00:14:04.760 align:start position:0% if you were just mine honestly or if 00:14:04.760 --> 00:14:07.420 align:start position:0% if you were just mine honestly or if that<00:14:05.000> your<00:14:05.150> mind<00:14:05.510> just<00:14:05.750> randomly<00:14:06.200> and<00:14:06.650> we 00:14:07.420 --> 00:14:07.430 align:start position:0% that your mind just randomly and we 00:14:07.430 --> 00:14:11.170 align:start position:0% that your mind just randomly and we wanted<00:14:07.730> to<00:14:07.850> prepare<00:14:08.090> the<00:14:08.240> results<00:14:10.000> so<00:14:11.000> whether 00:14:11.170 --> 00:14:11.180 align:start position:0% wanted to prepare the results so whether 00:14:11.180 --> 00:14:14.410 align:start position:0% wanted to prepare the results so whether there<00:14:11.390> was<00:14:11.480> also<00:14:11.630> play<00:14:12.400> we<00:14:13.400> found<00:14:13.610> it<00:14:13.760> when<00:14:14.270> you 00:14:14.410 --> 00:14:14.420 align:start position:0% there was also play we found it when you 00:14:14.420 --> 00:14:18.190 align:start position:0% there was also play we found it when you ran<00:14:14.660> for<00:14:15.820> even<00:14:16.820> just<00:14:17.000> like<00:14:17.180> a<00:14:17.420> thousand 00:14:18.190 --> 00:14:18.200 align:start position:0% ran for even just like a thousand 00:14:18.200 --> 00:14:21.850 align:start position:0% ran for even just like a thousand episodes<00:14:18.320> the<00:14:19.010> agent<00:14:19.730> that<00:14:20.110> was<00:14:21.110> incentivized 00:14:21.850 --> 00:14:21.860 align:start position:0% episodes the agent that was incentivized 00:14:21.860 --> 00:14:23.490 align:start position:0% episodes the agent that was incentivized to<00:14:21.920> hold<00:14:22.220> his<00:14:22.460> chain<00:14:22.730> a<00:14:22.940> ten<00:14:23.150> percent 00:14:23.490 --> 00:14:23.500 align:start position:0% to hold his chain a ten percent 00:14:23.500 --> 00:14:25.680 align:start position:0% to hold his chain a ten percent for<00:14:23.980> plumbing<00:14:24.160> work<00:14:24.430> because<00:14:24.940> they're<00:14:25.510> all 00:14:25.680 --> 00:14:25.690 align:start position:0% for plumbing work because they're all 00:14:25.690 --> 00:14:27.270 align:start position:0% for plumbing work because they're all ideas<00:14:26.020> that<00:14:26.050> I've<00:14:26.350> done<00:14:26.470> percent<00:14:26.920> you're<00:14:27.100> not 00:14:27.270 --> 00:14:27.280 align:start position:0% ideas that I've done percent you're not 00:14:27.280 --> 00:14:29.100 align:start position:0% ideas that I've done percent you're not getting<00:14:27.670> are<00:14:27.730> gonna<00:14:27.940> be<00:14:28.090> ahead<00:14:28.210> of<00:14:28.660> all<00:14:29.080> the 00:14:29.100 --> 00:14:29.110 align:start position:0% getting are gonna be ahead of all the 00:14:29.110 --> 00:14:30.510 align:start position:0% getting are gonna be ahead of all the other<00:14:29.410> miners<00:14:29.590> who<00:14:29.830> don't<00:14:29.980> have<00:14:30.100> enough<00:14:30.280> power 00:14:30.510 --> 00:14:30.520 align:start position:0% other miners who don't have enough power 00:14:30.520 --> 00:14:34.980 align:start position:0% other miners who don't have enough power so<00:14:30.910> you're<00:14:31.570> just<00:14:31.780> in<00:14:31.960> the<00:14:32.110> losing<00:14:32.470> one<00:14:33.990> however 00:14:34.980 --> 00:14:34.990 align:start position:0% so you're just in the losing one however 00:14:34.990 --> 00:14:36.990 align:start position:0% so you're just in the losing one however for<00:14:35.800> any<00:14:35.860> person<00:14:36.220> they<00:14:36.460> already<00:14:36.580> start 00:14:36.990 --> 00:14:37.000 align:start position:0% for any person they already start 00:14:37.000 --> 00:14:39.800 align:start position:0% for any person they already start gaining<00:14:37.210> an<00:14:37.780> increase<00:14:38.140> in<00:14:38.290> your<00:14:38.440> reward<00:14:38.860> after 00:14:39.800 --> 00:14:39.810 align:start position:0% gaining an increase in your reward after 00:14:39.810 --> 00:14:42.090 align:start position:0% gaining an increase in your reward after maybe<00:14:40.810> like<00:14:40.990> five<00:14:41.230> hundred<00:14:41.380> sites<00:14:41.830> you<00:14:42.040> can 00:14:42.090 --> 00:14:42.100 align:start position:0% maybe like five hundred sites you can 00:14:42.100 --> 00:14:43.860 align:start position:0% maybe like five hundred sites you can already<00:14:42.460> see<00:14:42.700> a<00:14:43.000> little<00:14:43.480> bit<00:14:43.660> of<00:14:43.720> an 00:14:43.860 --> 00:14:43.870 align:start position:0% already see a little bit of an 00:14:43.870 --> 00:14:45.330 align:start position:0% already see a little bit of an improvement<00:14:44.200> in<00:14:44.620> terms<00:14:44.800> of<00:14:45.010> what<00:14:45.160> you're 00:14:45.330 --> 00:14:45.340 align:start position:0% improvement in terms of what you're 00:14:45.340 --> 00:14:47.910 align:start position:0% improvement in terms of what you're doing<00:14:45.490> and<00:14:45.880> then<00:14:46.780> at<00:14:46.990> sixty<00:14:47.290> percent<00:14:47.440> it's 00:14:47.910 --> 00:14:47.920 align:start position:0% doing and then at sixty percent it's 00:14:47.920 --> 00:14:50.220 align:start position:0% doing and then at sixty percent it's just<00:14:48.160> like<00:14:48.340> taking<00:14:48.670> over<00:14:48.880> the<00:14:49.120> network<00:14:49.420> so 00:14:50.220 --> 00:14:50.230 align:start position:0% just like taking over the network so 00:14:50.230 --> 00:14:53.250 align:start position:0% just like taking over the network so their<00:14:50.380> results<00:14:50.860> I<00:14:51.670> never<00:14:52.540> had<00:14:52.720> that<00:14:52.750> here 00:14:53.250 --> 00:14:53.260 align:start position:0% their results I never had that here 00:14:53.260 --> 00:14:55.200 align:start position:0% their results I never had that here anywhere<00:14:54.010> else<00:14:54.220> for<00:14:54.340> looking<00:14:54.700> at<00:14:55.060> this 00:14:55.200 --> 00:14:55.210 align:start position:0% anywhere else for looking at this 00:14:55.210 --> 00:14:57.540 align:start position:0% anywhere else for looking at this nitrate<00:14:55.750> which<00:14:55.960> is<00:14:55.990> the<00:14:56.350> ratio<00:14:56.770> of<00:14:57.010> lost<00:14:57.370> of 00:14:57.540 --> 00:14:57.550 align:start position:0% nitrate which is the ratio of lost of 00:14:57.550 --> 00:15:00.090 align:start position:0% nitrate which is the ratio of lost of mine<00:14:57.760> so<00:14:58.090> this<00:14:58.300> is<00:14:58.510> how<00:14:59.260> many<00:14:59.470> blots<00:14:59.710> that<00:14:59.890> you 00:15:00.090 --> 00:15:00.100 align:start position:0% mine so this is how many blots that you 00:15:00.100 --> 00:15:02.700 align:start position:0% mine so this is how many blots that you mind<00:15:00.340> versus<00:15:00.940> the<00:15:01.060> number<00:15:01.570> total<00:15:02.080> watts<00:15:02.440> mine 00:15:02.700 --> 00:15:02.710 align:start position:0% mind versus the number total watts mine 00:15:02.710 --> 00:15:06.540 align:start position:0% mind versus the number total watts mine so<00:15:03.550> what<00:15:04.180> you<00:15:04.330> see<00:15:04.510> here<00:15:04.780> is<00:15:05.050> that<00:15:05.400> for<00:15:06.400> ten 00:15:06.540 --> 00:15:06.550 align:start position:0% so what you see here is that for ten 00:15:06.550 --> 00:15:09.690 align:start position:0% so what you see here is that for ten percent<00:15:06.670> you're<00:15:07.240> doing<00:15:07.660> worse<00:15:07.920> than<00:15:08.920> if<00:15:09.580> you 00:15:09.690 --> 00:15:09.700 align:start position:0% percent you're doing worse than if you 00:15:09.700 --> 00:15:12.720 align:start position:0% percent you're doing worse than if you were<00:15:09.880> just<00:15:10.300> being<00:15:10.780> honest<00:15:11.200> but<00:15:12.190> forty<00:15:12.580> percent 00:15:12.720 --> 00:15:12.730 align:start position:0% were just being honest but forty percent 00:15:12.730 --> 00:15:16.560 align:start position:0% were just being honest but forty percent though<00:15:13.150> you<00:15:13.210> you<00:15:13.960> started<00:15:14.320> moving<00:15:14.680> we<00:15:15.610> set<00:15:16.330> two 00:15:16.560 --> 00:15:16.570 align:start position:0% though you you started moving we set two 00:15:16.570 --> 00:15:19.620 align:start position:0% though you you started moving we set two different<00:15:17.020> reward<00:15:17.530> systems<00:15:18.010> just<00:15:18.630> testing 00:15:19.620 --> 00:15:19.630 align:start position:0% different reward systems just testing 00:15:19.630 --> 00:15:23.700 align:start position:0% different reward systems just testing and<00:15:20.130> for<00:15:21.130> one<00:15:21.310> of<00:15:21.340> those<00:15:22.380> starting<00:15:23.380> to<00:15:23.560> be 00:15:23.700 --> 00:15:23.710 align:start position:0% and for one of those starting to be 00:15:23.710 --> 00:15:26.460 align:start position:0% and for one of those starting to be better<00:15:24.220> at<00:15:24.460> twenty<00:15:24.700> percent<00:15:25.050> while<00:15:26.050> acting 00:15:26.460 --> 00:15:26.470 align:start position:0% better at twenty percent while acting 00:15:26.470 --> 00:15:28.740 align:start position:0% better at twenty percent while acting selfishly<00:15:26.680> so<00:15:27.130> he's<00:15:27.430> getting<00:15:27.550> like<00:15:27.910> almost 00:15:28.740 --> 00:15:28.750 align:start position:0% selfishly so he's getting like almost 00:15:28.750 --> 00:15:31.079 align:start position:0% selfishly so he's getting like almost 50%<00:15:29.200> of<00:15:29.740> the<00:15:29.860> watts<00:15:30.100> which<00:15:30.520> is<00:15:30.670> gonna<00:15:30.790> slay 00:15:31.079 --> 00:15:31.089 align:start position:0% 50% of the watts which is gonna slay 00:15:31.089 --> 00:15:34.230 align:start position:0% 50% of the watts which is gonna slay increase<00:15:32.100> and<00:15:33.100> at<00:15:33.220> sixty<00:15:33.550> percent<00:15:33.700> is<00:15:34.000> just 00:15:34.230 --> 00:15:34.240 align:start position:0% increase and at sixty percent is just 00:15:34.240 --> 00:15:36.240 align:start position:0% increase and at sixty percent is just mining<00:15:34.960> all<00:15:35.140> of<00:15:35.290> the<00:15:35.410> boss<00:15:35.589> because<00:15:35.800> you<00:15:36.040> get 00:15:36.240 --> 00:15:36.250 align:start position:0% mining all of the boss because you get 00:15:36.250 --> 00:15:39.000 align:start position:0% mining all of the boss because you get such<00:15:36.520> a<00:15:36.550> huge<00:15:37.000> advantage<00:15:37.440> from<00:15:38.440> selfish<00:15:38.770> money 00:15:39.000 --> 00:15:39.010 align:start position:0% such a huge advantage from selfish money 00:15:39.010 --> 00:15:41.310 align:start position:0% such a huge advantage from selfish money that<00:15:39.700> essentially<00:15:40.390> you're<00:15:40.480> publishing<00:15:40.870> most 00:15:41.310 --> 00:15:41.320 align:start position:0% that essentially you're publishing most 00:15:41.320 --> 00:15:43.380 align:start position:0% that essentially you're publishing most of<00:15:41.500> the<00:15:41.589> bots<00:15:41.800> because<00:15:42.100> nobody<00:15:42.400> can<00:15:42.550> catch<00:15:43.150> up 00:15:43.380 --> 00:15:43.390 align:start position:0% of the bots because nobody can catch up 00:15:43.390 --> 00:15:45.000 align:start position:0% of the bots because nobody can catch up to<00:15:43.540> you<00:15:43.720> and<00:15:43.990> what<00:15:44.260> happens<00:15:44.589> is<00:15:44.710> that<00:15:44.830> you're 00:15:45.000 --> 00:15:45.010 align:start position:0% to you and what happens is that you're 00:15:45.010 --> 00:15:48.300 align:start position:0% to you and what happens is that you're also<00:15:45.220> decreasing<00:15:45.730> the<00:15:46.030> difficulty<00:15:46.600> of<00:15:47.160> lot<00:15:48.160> so 00:15:48.300 --> 00:15:48.310 align:start position:0% also decreasing the difficulty of lot so 00:15:48.310 --> 00:15:50.540 align:start position:0% also decreasing the difficulty of lot so you<00:15:48.460> can<00:15:48.610> my<00:15:48.760> investor<00:15:49.330> with<00:15:49.750> the<00:15:49.870> same<00:15:50.080> I 00:15:50.540 --> 00:15:50.550 align:start position:0% you can my investor with the same I 00:15:50.550 --> 00:15:55.140 align:start position:0% you can my investor with the same I should<00:15:51.550> rate<00:15:53.010> and<00:15:54.010> so<00:15:54.280> what<00:15:54.670> are<00:15:54.790> some<00:15:54.850> of<00:15:55.089> the 00:15:55.140 --> 00:15:55.150 align:start position:0% should rate and so what are some of the 00:15:55.150 --> 00:15:58.500 align:start position:0% should rate and so what are some of the key<00:15:55.330> takeaways<00:15:55.450> of<00:15:55.870> all<00:15:56.020> of<00:15:56.140> this<00:15:56.290> is<00:15:56.770> that<00:15:57.510> at 00:15:58.500 --> 00:15:58.510 align:start position:0% key takeaways of all of this is that at 00:15:58.510 --> 00:16:00.990 align:start position:0% key takeaways of all of this is that at sixty<00:15:58.839> percent<00:15:58.990> you're<00:15:59.770> doing<00:16:00.220> a<00:16:00.490> lot<00:16:00.880> of 00:16:00.990 --> 00:16:01.000 align:start position:0% sixty percent you're doing a lot of 00:16:01.000 --> 00:16:03.030 align:start position:0% sixty percent you're doing a lot of damage<00:16:01.270> but<00:16:01.420> we<00:16:01.750> all<00:16:01.839> know<00:16:01.990> that<00:16:02.170> even<00:16:02.530> like<00:16:02.950> I 00:16:03.030 --> 00:16:03.040 align:start position:0% damage but we all know that even like I 00:16:03.040 --> 00:16:04.230 align:start position:0% damage but we all know that even like I thought<00:16:03.160> this<00:16:03.310> interview<00:16:03.670> should<00:16:03.790> be<00:16:03.880> able<00:16:04.150> to 00:16:04.230 --> 00:16:04.240 align:start position:0% thought this interview should be able to 00:16:04.240 --> 00:16:07.980 align:start position:0% thought this interview should be able to move<00:16:04.360> a<00:16:04.630> lot<00:16:05.200> of<00:16:05.290> damage<00:16:05.620> a<00:16:05.910> forty<00:16:06.910> percent<00:16:07.060> you 00:16:07.980 --> 00:16:07.990 align:start position:0% move a lot of damage a forty percent you 00:16:07.990 --> 00:16:09.720 align:start position:0% move a lot of damage a forty percent you start<00:16:08.320> achieving<00:16:08.410> some<00:16:08.920> interesting<00:16:09.550> results 00:16:09.720 --> 00:16:09.730 align:start position:0% start achieving some interesting results 00:16:09.730 --> 00:16:12.410 align:start position:0% start achieving some interesting results but<00:16:10.480> one<00:16:11.290> of<00:16:11.380> our<00:16:11.470> hosts<00:16:11.710> was<00:16:11.950> about 00:16:12.410 --> 00:16:12.420 align:start position:0% but one of our hosts was about 00:16:12.420 --> 00:16:14.210 align:start position:0% but one of our hosts was about this<00:16:12.870> automated<00:16:13.260> wait<00:16:13.590> we<00:16:13.860> would<00:16:13.980> find 00:16:14.210 --> 00:16:14.220 align:start position:0% this automated wait we would find 00:16:14.220 --> 00:16:17.150 align:start position:0% this automated wait we would find perhaps<00:16:14.700> other<00:16:14.970> lost<00:16:15.690> something<00:16:16.680> different 00:16:17.150 --> 00:16:17.160 align:start position:0% perhaps other lost something different 00:16:17.160 --> 00:16:18.710 align:start position:0% perhaps other lost something different than<00:16:17.340> selfish<00:16:17.790> fighting<00:16:18.270> a<00:16:18.420> different 00:16:18.710 --> 00:16:18.720 align:start position:0% than selfish fighting a different 00:16:18.720 --> 00:16:21.290 align:start position:0% than selfish fighting a different strategy<00:16:18.930> and<00:16:19.410> then<00:16:20.280> a<00:16:20.310> type<00:16:20.610> person<00:16:20.940> is<00:16:21.090> just 00:16:21.290 --> 00:16:21.300 align:start position:0% strategy and then a type person is just 00:16:21.300 --> 00:16:22.910 align:start position:0% strategy and then a type person is just doing<00:16:21.600> very<00:16:21.900> poor<00:16:22.140> thing 00:16:22.910 --> 00:16:22.920 align:start position:0% doing very poor thing 00:16:22.920 --> 00:16:25.610 align:start position:0% doing very poor thing and<00:16:23.370> so<00:16:24.000> I<00:16:24.270> never<00:16:24.390> mentioned<00:16:24.660> before<00:16:25.080> that<00:16:25.110> 60% 00:16:25.610 --> 00:16:25.620 align:start position:0% and so I never mentioned before that 60% 00:16:25.620 --> 00:16:29.030 align:start position:0% and so I never mentioned before that 60% your<00:16:26.100> producing<00:16:26.700> over<00:16:27.000> 90%<00:16:27.660> of<00:16:27.690> the<00:16:27.840> boss<00:16:28.050> but 00:16:29.030 --> 00:16:29.040 align:start position:0% your producing over 90% of the boss but 00:16:29.040 --> 00:16:30.320 align:start position:0% your producing over 90% of the boss but what's<00:16:29.250> interesting<00:16:29.430> is<00:16:29.820> that<00:16:29.850> you're<00:16:30.180> not 00:16:30.320 --> 00:16:30.330 align:start position:0% what's interesting is that you're not 00:16:30.330 --> 00:16:32.360 align:start position:0% what's interesting is that you're not actually<00:16:30.540> making<00:16:31.230> any<00:16:31.470> more<00:16:31.740> money<00:16:31.890> if 00:16:32.360 --> 00:16:32.370 align:start position:0% actually making any more money if 00:16:32.370 --> 00:16:34.130 align:start position:0% actually making any more money if anything<00:16:32.940> you'll<00:16:33.360> be<00:16:33.540> making<00:16:33.720> around<00:16:34.080> the 00:16:34.130 --> 00:16:34.140 align:start position:0% anything you'll be making around the 00:16:34.140 --> 00:16:38.300 align:start position:0% anything you'll be making around the same<00:16:34.350> amount<00:16:34.380> of<00:16:34.740> etherium<00:16:36.170> and<00:16:37.170> so<00:16:37.500> I<00:16:37.800> mean 00:16:38.300 --> 00:16:38.310 align:start position:0% same amount of etherium and so I mean 00:16:38.310 --> 00:16:40.100 align:start position:0% same amount of etherium and so I mean you're<00:16:38.520> just<00:16:38.730> attacking<00:16:38.970> the<00:16:39.240> difficulty<00:16:39.750> but 00:16:40.100 --> 00:16:40.110 align:start position:0% you're just attacking the difficulty but 00:16:40.110 --> 00:16:44.420 align:start position:0% you're just attacking the difficulty but it's<00:16:40.440> not<00:16:40.740> necessarily<00:16:41.280> beneficial<00:16:41.640> and<00:16:43.430> yeah 00:16:44.420 --> 00:16:44.430 align:start position:0% it's not necessarily beneficial and yeah 00:16:44.430 --> 00:16:45.740 align:start position:0% it's not necessarily beneficial and yeah we<00:16:44.580> couldn't<00:16:44.820> find<00:16:44.940> a<00:16:45.030> better<00:16:45.210> strategy<00:16:45.420> that 00:16:45.740 --> 00:16:45.750 align:start position:0% we couldn't find a better strategy that 00:16:45.750 --> 00:16:47.870 align:start position:0% we couldn't find a better strategy that selfish<00:16:46.080> mining<00:16:46.440> but<00:16:47.340> it's<00:16:47.520> interesting 00:16:47.870 --> 00:16:47.880 align:start position:0% selfish mining but it's interesting 00:16:47.880 --> 00:16:50.330 align:start position:0% selfish mining but it's interesting because<00:16:48.180> then<00:16:48.840> what<00:16:49.380> you<00:16:49.500> can<00:16:49.680> do<00:16:49.950> is<00:16:50.160> use 00:16:50.330 --> 00:16:50.340 align:start position:0% because then what you can do is use 00:16:50.340 --> 00:16:52.700 align:start position:0% because then what you can do is use these<00:16:50.580> types<00:16:50.820> of<00:16:50.970> systems<00:16:51.420> to<00:16:51.530> test<00:16:52.530> different 00:16:52.700 --> 00:16:52.710 align:start position:0% these types of systems to test different 00:16:52.710 --> 00:16:55.160 align:start position:0% these types of systems to test different particles<00:16:53.340> and<00:16:53.610> understand<00:16:54.180> if<00:16:54.570> there's 00:16:55.160 --> 00:16:55.170 align:start position:0% particles and understand if there's 00:16:55.170 --> 00:16:57.500 align:start position:0% particles and understand if there's something<00:16:55.830> going<00:16:55.950> on<00:16:56.190> so<00:16:56.520> perhaps<00:16:57.030> you<00:16:57.330> can 00:16:57.500 --> 00:16:57.510 align:start position:0% something going on so perhaps you can 00:16:57.510 --> 00:17:01.310 align:start position:0% something going on so perhaps you can see<00:16:58.230> CNN<00:16:59.010> signal<00:16:59.910> detection<00:17:00.270> system<00:17:01.140> that 00:17:01.310 --> 00:17:01.320 align:start position:0% see CNN signal detection system that 00:17:01.320 --> 00:17:02.150 align:start position:0% see CNN signal detection system that there's<00:17:01.500> something<00:17:01.800> wrong<00:17:01.920> with<00:17:02.070> the 00:17:02.150 --> 00:17:02.160 align:start position:0% there's something wrong with the 00:17:02.160 --> 00:17:04.010 align:start position:0% there's something wrong with the protocol<00:17:02.550> because<00:17:02.850> if<00:17:03.360> you<00:17:03.480> don't<00:17:03.660> follow<00:17:03.810> the 00:17:04.010 --> 00:17:04.020 align:start position:0% protocol because if you don't follow the 00:17:04.020 --> 00:17:06.079 align:start position:0% protocol because if you don't follow the normal<00:17:04.589> strategy<00:17:05.070> you're<00:17:05.190> doing<00:17:05.459> better<00:17:05.610> and 00:17:06.079 --> 00:17:06.089 align:start position:0% normal strategy you're doing better and 00:17:06.089 --> 00:17:08.840 align:start position:0% normal strategy you're doing better and so<00:17:06.420> this<00:17:07.050> can<00:17:07.290> help<00:17:07.500> you<00:17:07.680> understand<00:17:08.250> evaluate 00:17:08.840 --> 00:17:08.850 align:start position:0% so this can help you understand evaluate 00:17:08.850 --> 00:17:11.630 align:start position:0% so this can help you understand evaluate in<00:17:09.089> different<00:17:09.390> ways<00:17:10.160> what's<00:17:11.160> happening<00:17:11.550> and 00:17:11.630 --> 00:17:11.640 align:start position:0% in different ways what's happening and 00:17:11.640 --> 00:17:14.840 align:start position:0% in different ways what's happening and you<00:17:11.880> shouldn't<00:17:12.180> just<00:17:13.400> think<00:17:14.400> that<00:17:14.579> this<00:17:14.699> is 00:17:14.840 --> 00:17:14.850 align:start position:0% you shouldn't just think that this is 00:17:14.850 --> 00:17:16.550 align:start position:0% you shouldn't just think that this is not<00:17:15.000> a<00:17:15.030> wholesome<00:17:15.360> solution<00:17:15.839> but<00:17:15.930> rather<00:17:15.990> a 00:17:16.550 --> 00:17:16.560 align:start position:0% not a wholesome solution but rather a 00:17:16.560 --> 00:17:20.960 align:start position:0% not a wholesome solution but rather a good<00:17:16.860> indicator<00:17:17.250> of<00:17:17.339> what<00:17:17.610> can<00:17:17.760> happen<00:17:18.680> and<00:17:19.970> so 00:17:20.960 --> 00:17:20.970 align:start position:0% good indicator of what can happen and so 00:17:20.970 --> 00:17:22.610 align:start position:0% good indicator of what can happen and so this<00:17:21.120> is<00:17:21.240> an<00:17:21.329> open<00:17:21.570> source<00:17:21.780> project<00:17:22.020> it's<00:17:22.470> all 00:17:22.610 --> 00:17:22.620 align:start position:0% this is an open source project it's all 00:17:22.620 --> 00:17:25.160 align:start position:0% this is an open source project it's all I<00:17:22.650> can<00:17:22.829> get<00:17:22.890> hope<00:17:23.160> so<00:17:23.400> if<00:17:23.670> anyone<00:17:24.510> feels<00:17:24.839> like 00:17:25.160 --> 00:17:25.170 align:start position:0% I can get hope so if anyone feels like 00:17:25.170 --> 00:17:26.569 align:start position:0% I can get hope so if anyone feels like they<00:17:25.350> want<00:17:25.500> to<00:17:25.620> test<00:17:25.800> their<00:17:26.010> own<00:17:26.040> protocols 00:17:26.569 --> 00:17:26.579 align:start position:0% they want to test their own protocols 00:17:26.579 --> 00:17:29.900 align:start position:0% they want to test their own protocols you're<00:17:26.790> willing<00:17:27.180> to<00:17:27.209> do<00:17:27.449> it<00:17:27.600> and<00:17:28.460> yeah<00:17:29.460> that's 00:17:29.900 --> 00:17:29.910 align:start position:0% you're willing to do it and yeah that's 00:17:29.910 --> 00:17:33.540 align:start position:0% you're willing to do it and yeah that's it 00:17:33.540 --> 00:17:33.550 align:start position:0% 00:17:33.550 --> 00:17:42.610 align:start position:0% [Applause] 00:17:42.610 --> 00:17:42.620 align:start position:0% 00:17:42.620 --> 00:17:44.680 align:start position:0% you