[00:27.200 --> 00:28.160]  We're here.
[00:28.720 --> 00:29.900]  It's Sunday.
[00:29.900 --> 00:32.380]  It's getting towards the middle of the afternoon.
[00:32.760 --> 00:34.180]  I hope you're staying energized.
[00:34.180 --> 00:37.920]  There's a lot of good content, including this talk coming right up here.
[00:38.040 --> 00:40.300]  You're in the right place for the 3 o'clock session.
[00:40.480 --> 00:43.640]  I'll give you a couple of quick announcements about how the rest of the day is looking.
[00:43.820 --> 00:49.320]  There are still some other sessions going on after this one, including in the fourth track in the Coffee House.
[00:49.440 --> 00:52.180]  You can get that schedule on wiki.hope.net.
[00:52.180 --> 00:55.820]  See if there's something you want to do after your experience in this room.
[00:56.200 --> 01:00.960]  And also, we have closing ceremonies coming up at 6 p.m.
[01:01.260 --> 01:02.740]  Closing ceremonies are pretty cool.
[01:02.740 --> 01:10.140]  We just give sort of a nice thank you to everyone that's attended, talk a little bit about how the conference came together, a little bit of the, you know, behind-the-scenes information.
[01:10.300 --> 01:12.760]  So that's kind of a fun thing to attend.
[01:12.960 --> 01:24.660]  And at the end of closing ceremonies, there will be a band, a Brazilian salsa band, starting up right outside room 416 and processing through the building, ending up out front on the plaza.
[01:24.660 --> 01:29.840]  So we'll have a little bit of a street party atmosphere to wrap up the conference here.
[01:30.100 --> 01:31.300]  A New HOPE 2022.
[01:31.760 --> 01:32.740]  Pretty cool, right?
[01:33.220 --> 01:36.880]  So the invitation is to stick around.
[01:36.880 --> 01:40.620]  If you're thinking of leaving early to beat the traffic, forget about the traffic.
[01:40.840 --> 01:43.420]  The traffic gets better later at night, right?
[01:43.600 --> 01:45.280]  So stick around for closing ceremonies.
[01:45.280 --> 01:49.500]  And also, if you can, please stick around to volunteer with packing up.
[01:49.500 --> 01:52.660]  That usually takes us until 11 or even a little bit later at night.
[01:52.900 --> 01:56.120]  We got to remove all the rental equipment, pack up a bunch of trucks.
[01:56.540 --> 01:59.820]  To some extent, clean up the facility, take down our signs, stuff like that.
[01:59.900 --> 02:00.880]  Mostly unskilled labor.
[02:01.060 --> 02:02.240]  We really appreciate the help.
[02:02.660 --> 02:05.580]  And then tomorrow morning also, we have to load up a couple more rental trucks.
[02:05.720 --> 02:07.420]  So if you have the opportunity to stick around.
[02:07.660 --> 02:10.840]  If you're staying on campus or nearby, please do consider volunteering.
[02:11.100 --> 02:12.300]  It is another hot day.
[02:12.300 --> 02:13.780]  So don't skimp on hydration.
[02:13.780 --> 02:18.520]  Just because you made it this far into the weekend doesn't mean there's not still opportunity to become dehydrated.
[02:18.520 --> 02:20.920]  So please do drink.
[02:21.300 --> 02:23.140]  And that's all I got to say.
[02:23.360 --> 02:26.940]  Let's get rolling with our next fascinating talk.
[02:27.080 --> 02:27.780]  Please welcome our speaker.
[02:32.930 --> 02:33.690]  Thank you.
[02:33.990 --> 02:35.390]  What a lovely community.
[02:35.530 --> 02:36.850]  It's so great to be back at HOPE.
[02:37.030 --> 02:37.790]  Hello, everyone.
[02:38.130 --> 02:40.970]  I don't know if I should be speaking directly into this.
[02:41.610 --> 02:42.370]  Okay, great.
[02:42.490 --> 02:43.970]  Well, I hope you all are energized.
[02:44.030 --> 02:46.570]  I definitely have had a long night.
[02:46.570 --> 02:49.610]  So if I have to read my slides, it's just me trying to make sense.
[02:50.330 --> 02:51.650]  So let's get started.
[02:51.890 --> 02:53.610]  My name is Jamie Joyce.
[02:53.630 --> 02:56.310]  This is the second time I will be speaking at HOPE.
[02:57.390 --> 03:00.330]  I am the executive director of the Society Library.
[03:00.630 --> 03:05.030]  And what the Society Library is, is a nonprofit collective intelligence organization.
[03:05.310 --> 03:15.070]  And what we work on is developing tools to augment the way in which we perceive social and political issues, so that we can make more informed, inclusive, and less biased decisions.
[03:16.530 --> 03:19.610]  Today, I'll be talking a little bit more about what we do.
[03:19.790 --> 03:22.590]  But the name of this talk is Hacking Comprehension.
[03:22.870 --> 03:33.270]  And what it is about is essentially making sense of not only the big issues like climate change and reproductive rights and things like that, but also each other.
[03:33.270 --> 03:42.310]  And that means overcoming limitations of time, attention, perception, emotional empathy, all of the different things that get in the way of us understanding each other.
[03:42.510 --> 03:51.470]  So I'm going to be jumping into an exercise to kind of show how we don't see eye-to-eye, but there are some other things I'm going to be talking about, too.
[03:51.470 --> 03:53.650]  So, first, not seeing eye-to-eye.
[03:54.270 --> 04:04.950]  And then, secondly, I'm going to talk about the work that we do at the Society Library and how we try to improve the information environment by seeding it with different visualizations of knowledge and steel-manned knowledge.
[04:05.530 --> 04:18.470]  And then, thirdly, I'm going to inquire into what I think is going to be the next evolution of media and where we should start orienting our design in order to augment the way in which we see the world and each other.
[04:19.530 --> 04:22.570]  I'm actually going to scoot down so I have my notes closer.
[04:22.970 --> 04:23.090]  Okay.
[04:23.550 --> 04:25.750]  So, not seeing eye-to-eye.
[04:25.930 --> 04:27.010]  Are you all ready for this?
[04:27.190 --> 04:28.430]  I'm going to ask you to raise your hands.
[04:28.570 --> 04:29.170]  It's going to be great.
[04:29.670 --> 04:30.710]  Oh, you all did already.
[04:31.670 --> 04:33.830]  There's no reason yet, but just we're about to get started.
[04:34.390 --> 04:34.610]  Okay.
[04:35.010 --> 04:35.950]  Now, raise your hand.
[04:36.350 --> 04:37.770]  Do you recognize this dress?
[04:38.050 --> 04:38.950]  Have you seen this before?
[04:39.490 --> 04:41.010]  Oh, it's like everyone in the room.
[04:41.210 --> 04:41.610]  Wonderful.
[04:41.950 --> 04:42.170]  Okay.
[04:43.030 --> 04:44.790]  This is the dress that broke the Internet.
[04:44.850 --> 04:48.450]  If you remember, people thought about what color this dress is.
[04:49.130 --> 04:50.030]  So, question.
[04:50.210 --> 04:52.150]  Who sees this dress as black and blue?
[04:53.950 --> 04:54.430]  Okay.
[04:54.630 --> 04:56.290]  Who sees this dress as gold and white?
[04:57.330 --> 04:57.690]  A few.
[04:58.350 --> 04:59.430]  Does anyone know why?
[05:02.110 --> 05:02.890]  Go for it.
[05:02.970 --> 05:03.050]  Yeah.
[05:16.590 --> 05:17.630]  Well, basically.
[05:17.890 --> 05:19.910]  Basically, our brains are making an assumption.
[05:20.210 --> 05:29.210]  But the reason why is actually extremely interesting because not only did this dress break the Internet, but it also apparently broke our theory of vision.
[05:29.470 --> 05:32.610]  Because existing color theory could not explain this phenomenon.
[05:32.850 --> 05:36.050]  Why do people... Some people see this as black and blue.
[05:36.290 --> 05:37.690]  Other people see it as gold and white.
[05:37.690 --> 05:45.110]  It turned out that Pascal Wallisch published a paper in the Journal of Vision.
[05:45.430 --> 05:48.210]  And he offered a reason as to why we see the dress differently.
[05:48.390 --> 05:52.950]  And it's not because of like rods or cones in our eyes or necessarily because of the contrast.
[05:53.090 --> 05:56.670]  Although the ambiguity of the lighting of the dress definitely does have an impact.
[05:56.670 --> 05:58.590]  But it actually has to do... Oops.
[05:58.970 --> 06:01.370]  It actually has to do with your prior experiences.
[06:01.730 --> 06:09.210]  So, the people who are likely to see this dress as black and blue are the ones who spend a lot of time under artificial light.
[06:09.550 --> 06:16.190]  When the people see it as gold and white actually have an existing prior of assuming that it is in natural light.
[06:16.190 --> 06:23.270]  So, it's literally your prior past experiences that is shaping the way that your brain is hallucinating color for you to experience.
[06:24.030 --> 06:32.190]  So, you know, early morning birds and night owls completely see this dress different colors based on prior experience, not physicality.
[06:33.010 --> 06:39.230]  Now, the important thing, of course, about scientific hypotheses is that you have to put it to the test and see if it has predictive power.
[06:39.230 --> 06:43.930]  And so, after Pascal Wallisch published this in the Journal of Vision...
[06:43.930 --> 06:44.930]  I just want to make sure I get that correct.
[06:45.330 --> 06:48.310]  They went ahead and proceeded with some other tests, including this one.
[06:48.850 --> 06:51.650]  This was an image that didn't just pop up on the Internet.
[06:51.650 --> 06:54.190]  This is something that they actually intentionally produced.
[06:54.410 --> 06:56.770]  They produced similar color conditions.
[06:56.910 --> 07:02.210]  So, I'm going to ask, who sees these Crocs as like green and grayish?
[07:03.830 --> 07:04.350]  Interesting.
[07:04.630 --> 07:04.790]  Okay.
[07:05.010 --> 07:07.470]  Who sees these Crocs as pink and white?
[07:08.210 --> 07:09.590]  This is so exciting.
[07:09.830 --> 07:09.970]  Okay.
[07:10.090 --> 07:10.450]  Amazing.
[07:10.670 --> 07:11.890]  I've never seen this result before.
[07:12.230 --> 07:12.410]  Okay.
[07:12.750 --> 07:18.970]  So, can anyone guess why you all are seeing these Crocs as different colors?
[07:26.620 --> 07:27.140]  Bingo!
[07:27.500 --> 07:27.800]  Okay.
[07:27.980 --> 07:28.500]  Perfect.
[07:28.740 --> 07:33.680]  So, the existing prior of the dress was your relationship to natural versus artificial light.
[07:33.820 --> 07:41.240]  That existing prior caused your brain to fill in the ambiguity of the image and hallucinate either a black and blue dress or a gold and white dress.
[07:41.240 --> 07:43.580]  You over there got that totally correct.
[07:44.000 --> 07:47.560]  For people, the existing prior actually is correlated with age.
[07:47.760 --> 07:52.860]  So, people who are older will assume that those socks are supposed to be white.
[07:52.980 --> 07:59.200]  When people who are a little bit younger don't necessarily have that existing prior of socks being white.
[07:59.380 --> 08:00.400]  Socks could be any color.
[08:00.600 --> 08:02.240]  Crocs don't really have a default color.
[08:02.240 --> 08:04.220]  And so, that's where they found the distribution.
[08:04.400 --> 08:05.580]  Which is extremely interesting.
[08:06.120 --> 08:15.940]  So, what they concluded is essentially our past experiences can exist as priors in the way in which we calculate probabilities under ambiguous conditions.
[08:16.240 --> 08:18.060]  Especially when it comes to visualizing.
[08:19.320 --> 08:21.100]  Has anyone seen this image before?
[08:22.200 --> 08:22.760]  Okay.
[08:23.200 --> 08:29.640]  So, what's really interesting about the dress and the Crocs and socks is that we can be looking at the exact same image.
[08:29.760 --> 08:34.280]  Human beings can be looking at the exact same image and conclude that they're seeing completely different things.
[08:34.560 --> 08:36.680]  We can hallucinate our differences.
[08:36.920 --> 08:39.760]  But we can also interestingly hallucinate consensus.
[08:40.240 --> 08:42.840]  Does everyone see these strawberries as red?
[08:45.520 --> 08:45.920]  Yes.
[08:46.120 --> 08:46.220]  Yeah.
[08:46.480 --> 08:47.400]  Depends on the monitor.
[08:47.520 --> 08:47.840]  Hilarious.
[08:48.000 --> 08:48.140]  Okay.
[08:48.640 --> 08:50.780]  This one right here that I'm pointing at.
[08:51.000 --> 08:51.760]  Do you see it as...
[08:51.760 --> 08:52.920]  Strawberries as red?
[08:53.380 --> 08:54.240]  I see nods.
[08:54.300 --> 08:54.940]  I see nods.
[08:55.080 --> 08:55.260]  Okay.
[08:55.580 --> 08:58.120]  Would you believe that there's no red in that image at all?
[08:58.960 --> 08:59.360]  Okay.
[08:59.460 --> 09:00.060]  Some more nods.
[09:00.100 --> 09:00.300]  Great.
[09:00.480 --> 09:00.660]  Okay.
[09:00.860 --> 09:03.820]  I plucked these pixels out because I couldn't believe it myself.
[09:04.140 --> 09:06.840]  I mean, to me, those strawberries definitely look red.
[09:07.000 --> 09:09.800]  But it's, again, just our brain filling in the details.
[09:09.920 --> 09:11.340]  Assuming what should be there.
[09:11.340 --> 09:14.540]  And generally, a lot of people tend to see this as red.
[09:14.540 --> 09:18.040]  It's not as divided as the Crocs and socks nor the dress.
[09:19.280 --> 09:19.720]  Okay.
[09:20.100 --> 09:20.480]  So.
[09:21.140 --> 09:22.260]  It gets worse though.
[09:22.660 --> 09:23.620]  Not only...
[09:23.620 --> 09:24.560]  Let me just conclude.
[09:25.500 --> 09:33.980]  So it seems as though one of the reasons why we don't see eye to eye is because even when we look at the exact same image, we can disagree because we are hallucinating different versions of reality.
[09:34.200 --> 09:37.580]  And sometimes when we look at the same image, we can all agree on what we see.
[09:37.580 --> 09:38.520]  But crazy enough.
[09:38.740 --> 09:43.540]  What we all are collectively seeing is a hallucinated fabrication.
[09:44.100 --> 09:49.800]  And it gets crazier because we are all not looking at the same information, of course, when it comes to social and political issues.
[09:50.020 --> 09:53.520]  We are not all zoomed into the same data, the same graphs, the same knowledge.
[09:53.880 --> 09:54.540]  Certainly not.
[09:54.540 --> 10:04.820]  And all of the ways in which we are interpreting that knowledge are, of course, informed by different experiences of life itself, different priors, traumas, lessons learned, tribes, and apparently sources of light.
[10:05.400 --> 10:13.820]  And the experience of life and therefore perception of reality can also warp moment to moment depending on whether we are rested, fed, sober, caffeinated, or stressed.
[10:14.000 --> 10:15.460]  Which I can definitely attest to.
[10:15.460 --> 10:20.160]  I crushed a low-carb monster energy drink right before this talk just for you.
[10:20.420 --> 10:23.140]  So I can tell my reality is definitely warped a little bit.
[10:23.860 --> 10:24.700]  Moving on.
[10:25.420 --> 10:38.500]  And even when we are really trying to be rational, we are really trying to survey the data landscape, put it in order, be logical, really inquire into understanding complex social and political issues like climate change and reproductive rights and all these things.
[10:38.500 --> 10:42.800]  Because our brains are primed to make over 300 kinds of logical reasoning errors.
[10:43.440 --> 10:47.900]  And there's over 190 different cognitive biases that we can be manipulated by as well.
[10:48.660 --> 10:55.080]  Plus, another way in which we don't see eye to eye is because each of us have different tools that we use to augment our perceptions of reality.
[10:55.480 --> 11:04.480]  Many of you in this room have a much different insight, vision, understanding of apps and software and websites because you know how to program.
[11:04.480 --> 11:07.000]  You can see and understand it in a different dimension.
[11:07.000 --> 11:08.640]  Not everybody has that.
[11:08.880 --> 11:17.800]  We also have asymmetric access to tools and understanding of things as small as like the physics of particles or as large as climate systems of Earth.
[11:18.020 --> 11:21.780]  If I take off my glasses right now, I believe I'm like technically legally blind.
[11:23.500 --> 11:27.280]  And luckily I'm aware of my terrible eyesight and I've gotten glasses.
[11:27.700 --> 11:29.860]  And sadly my prescription increases all the time.
[11:30.000 --> 11:32.580]  But many of us are not aware that we have blind spots.
[11:33.400 --> 11:38.060]  And we simply can't see what we can't see because we don't know what we don't know.
[11:38.960 --> 11:43.820]  Now, I'm going to extrapolate a little bit because this talk is about social and political issues.
[11:43.980 --> 11:55.920]  And I think that the way in which technology has amplified the amount of information that we're interacting with has potentially caused democracies to struggle a little bit.
[11:56.240 --> 11:59.460]  And it's no wonder people see the world differently.
[11:59.700 --> 12:04.660]  And many of us would probably agree that some of those differences in general are being agitated or exploited.
[12:04.680 --> 12:10.100]  And people are being weaponized and propagandized in battles over social control and political power.
[12:10.100 --> 12:23.800]  Frankly, I can understand why authoritarians may find it to be an efficient strategy to just make up people's minds for them by strictly controlling information access, confining their experiences, propagandizing people into tribes slash herd mentality,
[12:24.280 --> 12:31.900]  or coercing conformity and therefore restricting belief formation, and incentivizing decision making as a social control and political power mechanism.
[12:32.780 --> 12:38.140]  As opposed to just trying to get hundreds of millions of people to find consensus together.
[12:38.140 --> 12:47.160]  That seems like a very difficult thing to do, especially when people don't necessarily have the bandwidth of attention to all zoom into the same sets of information that are scattered across different forms of content.
[12:48.620 --> 13:08.920]  However, my hope is that our work at the Society Library will make a contribution towards improving that and causing change long term in part by building and hopefully democratizing access to new tools that can augment our perception of reality in order to increase comprehension of,
[13:09.060 --> 13:21.200]  at least as a starting point, the varying points of view that exist on issues that divide us, which can be rendered in new formats so we can make more informed, free, inclusive, and less biased decisions about political issues.
[13:21.380 --> 13:22.480]  If we choose to.
[13:22.640 --> 13:24.140]  If we want to opt in, of course.
[13:24.680 --> 13:34.260]  With such tools, I am hoping that we can more easily see our blind spots, see each other's views, and come to understand why some of us see black or gold, pink or green, or red or blue.
[13:35.720 --> 13:44.300]  So now, part two, I would love to talk about and show you some of the work of the Society Library a little bit, and then I'll talk about some open inquiries we have about enabling comprehension.
[13:44.860 --> 13:51.240]  Essentially, I am really curious about how far we could go augmenting our intelligence and comprehension with tech.
[13:52.180 --> 13:52.580]  Yay!
[13:52.740 --> 13:53.500]  So let's talk about work.
[13:54.960 --> 14:01.560]  Again, the Society Library is a non-profit organization, and we are dedicated to improving humanity's relationship to information.
[14:01.800 --> 14:03.000]  That's our big vision.
[14:03.220 --> 14:14.340]  We do a bunch of things from building political decision-making models to advising on the creation of legislation, providing educational curricula to various universities, and training professionals in our research methods.
[14:14.500 --> 14:20.260]  But today, I want to talk about how we build libraries of knowledge that are structured as societal-scale debates.
[14:20.260 --> 14:27.680]  In building these libraries, our goal is not to coerce someone into seeing a specific worldview or to arrive at a specific outcome.
[14:28.080 --> 14:33.580]  We are just looking to collect and compress knowledge to make wickedly complex debates comprehensible.
[14:33.880 --> 14:37.640]  We want everyone to be able to see from different points of view, make that actually possible.
[14:37.640 --> 14:54.580]  And for the first time, for some, make increased comprehension regarding the validity, depth, complexity, and variety of points of view on an issue accessible, as well as enabling the comprehension of new concepts and nuanced context of relevant knowledge.
[14:55.120 --> 14:57.900]  We will see some of the things that prevent comprehension.
[14:58.600 --> 15:05.200]  I would say that includes, as I mentioned before, limits in time, attention, empathy, patience, and access to information.
[15:05.200 --> 15:16.160]  So what we're specifically doing in the tech I'm about to show you is save users thousands of research hours and finding new ways to compress knowledge so that we can start seeing eye-to-eye a little bit more.
[15:16.820 --> 15:19.820]  But then I'm going to talk about limits to that approach because, of course, there always are.
[15:20.960 --> 15:21.400]  Okay.
[15:21.740 --> 15:22.700]  Oh, one more thing.
[15:23.120 --> 15:37.000]  Before I show you the library collection, a little background, how we build these libraries is we actually extract arguments, claims, and evidence from books, textbooks, scholarly works, government documents, television, podcasts, websites, documentaries,
[15:37.280 --> 15:43.940]  and social media, consolidate that content into a database in which we structure the data itself as a debate between all points of view.
[15:44.400 --> 15:47.280]  Last time I spoke at HOPE, I talked about our climate change collection.
[15:47.660 --> 15:52.720]  But today, I would like to show you a state-level issue having to do with nuclear energy.
[15:52.720 --> 15:53.920]  So let's dive in.
[15:55.540 --> 15:56.840]  It's going to autoplay.
[15:57.040 --> 15:57.360]  Okay.
[15:58.220 --> 16:01.040]  So here's the visualization of a database that we've built.
[16:01.220 --> 16:07.960]  The database shows the different points of view that people take on the last remaining nuclear power plant in California called Diablo Canyon.
[16:08.280 --> 16:13.000]  Diablo Canyon's two nuclear reactors are set to be decommissioned in 2024 and 2025.
[16:13.480 --> 16:18.340]  Some people believe that Diablo Canyon should remain open, and others believe it should be closed.
[16:18.340 --> 16:22.960]  At face value, it may look like it's a polarized issue.
[16:23.220 --> 16:24.760]  There's only two decisions, open, closed.
[16:25.700 --> 16:38.140]  However, what we found using our methods and extracting knowledge from over 4,000 references, including government documents and newspapers and blogs and petitions and yada, yada, we found that there's about nine positions that people take on this issue.
[16:38.300 --> 16:43.640]  And our work is to map out the arguments, claims, and evidence that support and refute those nine positions.
[16:44.200 --> 16:47.220]  I'm not confident in my ability to switch and give you a live demo.
[16:47.460 --> 16:50.560]  I don't, like, destroy the flow that we're in.
[16:50.880 --> 16:55.740]  But as you can see, what we've got here is conditions in which it should be opened or closed.
[16:55.920 --> 16:57.920]  So oftentimes we're forced to be binary.
[16:58.140 --> 17:01.860]  It's either open, it's closed, it's red, it's blue, it's yay or nay, pro and con.
[17:02.120 --> 17:10.060]  But in actuality, in reality, when we really dive into what people think, there are oftentimes only conditions in which someone will agree to something.
[17:10.060 --> 17:23.660]  So, for example, there is a position that people take where they want Diablo Canyon, the last remaining nuclear power plant in California, to remain open only if the current utility that operates it no longer operates it because they're deeply concerned about safety,
[17:24.020 --> 17:25.640]  oversight, all of these different things.
[17:26.440 --> 17:32.980]  Some people argue that it should remain open only if it becomes a polygeneration plan or only if it gets certain upgrades and things like that.
[17:33.080 --> 17:45.060]  So what's really important, I think, and what has been called mindset breaking by some people who've interacted with our work is that we start to break down the polarization because things are oftentimes not binary, even with that dress.
[17:45.240 --> 17:49.120]  It's not just people see it as black and blue or gold and white.
[17:49.280 --> 17:52.220]  There's actually people who see it in a lot of different colors, which is fascinating.
[17:52.460 --> 17:56.020]  And it's similar and we should keep that in mind with a lot of our different social problems.
[17:56.560 --> 18:01.740]  Moving on, we're going to dive into actually... this is just to kind of show like highest level view.
[18:01.740 --> 18:05.880]  We're going to dive down one little logical pathway and I'm going to show you both...
[18:06.340 --> 18:10.080]  We're going to zoom out and see the breadth of it as well as the depth of this argumentation.
[18:10.220 --> 18:14.760]  Because again, the whole purpose of these tools is to start augmenting the way we see social problems.
[18:14.920 --> 18:25.320]  If we're interacting with not databases of 4,000 different references, but instead just a news article here or there, a podcast here or there, it's very limited information that we're being given.
[18:25.320 --> 18:31.420]  And so the work that we're trying to do is greatly expand our comprehension of complexity of these problems.
[18:32.460 --> 18:36.860]  So let's take a walk down just one argument pathway in support of just one position.
[18:39.840 --> 18:40.660]  There we go.
[18:40.880 --> 18:40.980]  Okay.
[18:42.080 --> 18:47.720]  Position one states that Diablo Canyon should be decommissioned as scheduled due to a variety of different reasons.
[18:48.420 --> 18:55.600]  In part, because there's so many arguments to support and refute this position, we've clustered arguments by category.
[18:55.800 --> 18:59.220]  So we have economic, environmental, safety, energy-related, ethical, and political.
[18:59.900 --> 19:02.380]  These are extremely high-level categories.
[19:02.380 --> 19:09.220]  And because there's so many very high-level arguments underneath these, we've clustered them under these categories.
[19:09.220 --> 19:14.060]  But these claims that you're seeing here are also very high-level vague arguments.
[19:14.340 --> 19:22.880]  This one, for example, says market forces, financial incentives, and policy decisions have made the Diablo Canyon nuclear power plant redundant, noncompetitive, and undesirable.
[19:22.880 --> 19:25.100]  That's made of a number of different claims.
[19:25.280 --> 19:31.160]  Here's one specifically about it being noncompetitive and financially undesirable due to supply and demand.
[19:31.260 --> 19:32.940]  We can unpack that further.
[19:32.940 --> 19:42.580]  We get into economic assessments and argumentation about the specific impact of various different things happening in the economy that would impact the viability of the plant.
[19:42.740 --> 19:44.300]  That can break down further.
[19:44.560 --> 19:46.380]  There are papers that respond to that.
[19:46.500 --> 19:52.100]  And as you can see, there's pro and con argumentation arguing over the veracity and relevance of this data.
[19:52.100 --> 20:03.580]  And so the work that we do is just in collecting all of this different knowledge that's spread through different forms of media, through time, from different groups, and organizing that as a societal-scale communication.
[20:03.620 --> 20:05.680]  For us, this is just a state-level issue.
[20:06.400 --> 20:10.760]  Here I am on the video unpacking just the different sections.
[20:10.760 --> 20:15.200]  So that was a very high-level, vague economic claim that we just traveled down.
[20:15.300 --> 20:17.220]  But here I'm unpacking all the other categories.
[20:17.380 --> 20:26.080]  So we have economic, we've got environmental, we've got safety, we've got energy-related, we've got ethical and educational, and we've got political.
[20:26.380 --> 20:41.600]  And as soon as I'm done unpacking these, I'll zoom up and down, and you'll actually see even the most vague high-level arguments that people may be flippantly making on podcasts can be full of knowledge and argumentation about whether or not it's correct.
[20:42.200 --> 20:50.200]  And so the work that we've done in this particular database, and thank goodness this isn't how we render it for the public, although you can actually access this if you want.
[20:50.220 --> 20:51.780]  I personally love the debate map view.
[20:52.340 --> 20:55.040]  You can see it's just unbelievable amounts of knowledge.
[20:55.040 --> 21:02.400]  And so the purpose of visualizing it this way is to be able to really easily zoom out and see the complexity and volume.
[21:02.680 --> 21:05.960]  And hopefully that's a little bit mindset shifting.
[21:06.220 --> 21:20.080]  But we also render it in a different way where we deeply compress all of this knowledge to make it a little bit more comprehensible, and then also focus in on how do we contextualize these claims, provide really important indicators as to certain dimensions and characteristics of that.
[21:20.380 --> 21:23.520]  So we will move on and do that now.
[21:25.680 --> 21:29.140]  Okay, so this is a more readable narrative form.
[21:29.340 --> 21:31.760]  This is what we call society library papers.
[21:32.020 --> 21:36.420]  It is all of the knowledge of the database, but compressed into narrative form.
[21:36.600 --> 21:45.080]  So instead of the claims being in each little node, they could be partial clauses in sentences that read literally as a series of papers that you could unpack.
[21:45.320 --> 21:47.000]  So here's an example of that.
[21:48.040 --> 21:52.000]  So here, that's the first position, and all the categories are actually in line.
[21:52.000 --> 21:53.660]  And you can unpack.
[21:54.500 --> 21:57.560]  Instead of unpacking trees, you can open it up.
[21:58.040 --> 22:00.740]  Not only by unpacking it, as you'll see here.
[22:00.840 --> 22:02.260]  These are supporting argumentation.
[22:02.480 --> 22:12.120]  But you can also open each and every single one of these claims into like a Wikipedia-style page, full of knowledge about where did that claim come from.
[22:12.120 --> 22:16.580]  And there's information about like what type of media sources it came from.
[22:16.820 --> 22:18.580]  We do backups in the Internet Archive.
[22:18.680 --> 22:21.440]  So even if one of the links disappears, you can go and find it.
[22:21.580 --> 22:22.860]  There's some additional references.
[22:23.100 --> 22:25.060]  So essentially, you can check the integrity of our work.
[22:25.180 --> 22:28.620]  You can find the original context and quote in which that claim was expressed.
[22:28.620 --> 22:30.980]  And a bunch of other information on each one of these pages.
[22:32.680 --> 22:34.300]  Since I record this ahead of time, let's see.
[22:34.440 --> 22:34.620]  Oh, yeah.
[22:34.760 --> 22:38.840]  And then we also have this feature where you can opt into different expressions.
[22:38.840 --> 22:42.580]  So you can opt into technical or simple versions of each one of these paragraphs.
[22:42.700 --> 22:44.700]  We still have a little bit of work to do fleshing those out.
[22:44.860 --> 22:50.000]  But essentially, like that's another one of our features of enabling comprehension, which is what this is about.
[22:50.000 --> 23:02.100]  It's like if you're reading some technical claim about an economic assessment, and it's using terminology with which you're not familiar, you can just flip to the left and read a fourth grade reading level like version of that.
[23:02.280 --> 23:07.000]  Or perhaps you're an economist and you really want to know that they're using a specific model like the implant model.
[23:07.120 --> 23:13.100]  You can flip to the right and you can actually see a much more detailed explanation of that particular economic argumentation.
[23:13.280 --> 23:24.880]  And so this is just an example of how like we're trying to pull in and compress this knowledge to make comprehension of these really complex things actually possible in a format that's really familiar.
[23:26.720 --> 23:31.200]  But of course, I mean, haven't we moved on from text on paper?
[23:31.460 --> 23:37.020]  We're like in the land of TikToks and tweets and multimedia stuff and the metaverse is coming.
[23:37.280 --> 23:50.520]  So even though the Society Library is starting by just kind of reinventing the piece of paper, what's amazing about the databases that we're creating is I have no doubt that one, we will be able to render it in much more vivid, interactive and experiential knowledge.
[23:50.740 --> 23:52.360]  But I think we'll also have to.
[23:54.040 --> 23:58.660]  And I mean, yes, more and more diverse ways to visualize knowledge are possible.
[23:59.420 --> 24:00.960]  So I'm not concerned about that.
[24:01.100 --> 24:05.640]  I think it's very easy for us to just give away the data and designers can render it however they wish.
[24:05.800 --> 24:12.360]  And I hope there's lots of experimentation and iteration to make sure that it's meeting those goals of it being comprehensible.
[24:12.360 --> 24:18.380]  But what I am concerned about, which is a part of what this talk is about, is the ethics of inducing comprehension.
[24:19.280 --> 24:24.080]  How do we decide what information should be expressed to a level of vividness?
[24:24.300 --> 24:25.600]  You know, what if it's too impactful?
[24:25.780 --> 24:26.840]  What if it's too persuasive?
[24:27.100 --> 24:28.480]  There's just so many pitfalls.
[24:29.320 --> 24:36.580]  I imagine many people here who work in security are aware of the nefarious ways to force someone to comprehend something through manipulation.
[24:36.580 --> 24:42.300]  Ways to coerce someone to change their mind or come to a conclusion without necessarily having their consent.
[24:42.500 --> 24:44.220]  This is very popular in politics.
[24:44.400 --> 24:49.420]  There are people who hit the streets who have mastered techniques to make people doubt their own beliefs.
[24:49.560 --> 24:52.820]  And, you know, I'm not saying that's wrong or change their mind or things like that.
[24:53.060 --> 24:54.940]  Politics is a power game.
[24:55.100 --> 24:57.560]  And so the society library is not interested in that.
[24:57.700 --> 25:00.380]  We're interested in serving information needs without being coercive.
[25:00.380 --> 25:09.700]  So part of this talk is me bringing to you all this ethical conundrum of how do we figure out the right way of visualizing this knowledge without accidentally prioritizing something and being persuasive.
[25:09.880 --> 25:10.820]  We've thought a lot about it.
[25:10.920 --> 25:12.000]  There's policies on our website.
[25:12.180 --> 25:13.260]  They're rendered as Google Docs.
[25:13.360 --> 25:14.160]  Please give us feedback.
[25:14.360 --> 25:16.660]  But we're going to talk about something in particular.
[25:17.360 --> 25:33.080]  Some of the tactics before I move on, though, include, like, compelling disinformation, agenda setting, manipulative language, appeals to personality and psychographic makeup, outright repetition, negging, bots to drive perception of consensus by playing on tribal or in-group desirability, deepfakes.
[25:34.660 --> 25:41.980]  And then, like, have any of you seen that magic trick where a magician will come up to you and then they'll engage in conversation?
[25:41.980 --> 25:56.140]  But what they're really doing is seeding everything that they're saying with movie titles and famous movie quotes where they're actually, like, building a semantic web in your brain so that you're thinking of the lead actor of all of those movies without them ever explicitly stating it.
[25:56.480 --> 25:58.220]  And then they ask you, like, to think of an actor.
[25:58.360 --> 26:01.600]  And, of course, you think of that person because your brain, that's how your brain generates information.
[26:01.600 --> 26:03.560]  And they pretend to read your mind and it's like...
[26:04.260 --> 26:05.480]  Yeah, that's manipulation, too.
[26:06.540 --> 26:13.640]  And, obviously, the easiest way to go about avoiding these pitfalls in the work that we do is just, like, not executing any of these strategies intentionally.
[26:13.640 --> 26:19.840]  But there are also tough calls that we have to make, like, who decides what language is a dog whistle or not?
[26:20.000 --> 26:26.400]  How do you go about contextualizing content so that it is represented in a deliberation but not given extreme undue weight?
[26:26.580 --> 26:31.820]  How do you give people enough information to induce comprehension without coercing people to a conclusion?
[26:32.120 --> 26:38.800]  There's an argument to be made that technology is not neutral, and that means it's important to be clear about the intentions of the tools that we make.
[26:38.800 --> 26:48.720]  At the Society Library, our intentions for producing these libraries are to enable comprehension and appreciation of the legitimate nuance and complexity of our social and political issues.
[26:48.960 --> 26:57.240]  We want our tools to give people a new site, to overcome their blind spots that are inherent simply because there's too much information spread all over the web and world.
[26:57.400 --> 27:01.560]  No one has the time to, like, map all the knowledge that we have.
[27:01.660 --> 27:05.520]  We've saved you 6,000 hours of research just by creating that database.
[27:05.520 --> 27:07.680]  It certainly won't take you that long to read it all.
[27:09.340 --> 27:14.580]  And in giving people this new site, it's both the Google Earth and Street View of the issue.
[27:14.900 --> 27:29.580]  And we're hoping that what this does is inspire increased intellectual humility, depolarized and deradicalized attitudes, increased curiosity, increased resilience to disinforming content, and we also hope our products and services will increase subject matter knowledge.
[27:29.780 --> 27:32.380]  These are the things that we're aiming for and, of course, can test for.
[27:32.380 --> 27:39.960]  But this is where the talk turns from a kind of, like, show and tell about what we're doing and more of an open inquiry.
[27:40.200 --> 27:43.960]  Beyond comprehension, my question is, can we induce wisdom?
[27:44.980 --> 27:52.220]  I imagine many of you in this room have read sci-fi or fictional stories in which wise sages are depicted.
[27:52.860 --> 28:00.180]  So does anyone want to tell me what are some characteristics of, like, the wise oracle, the wise sage, the wise guide?
[28:01.000 --> 28:01.440]  Yes?
[28:02.860 --> 28:03.300]  Experience.
[28:03.300 --> 28:03.520]  Experience.
[28:03.800 --> 28:03.960]  Okay.
[28:22.480 --> 28:22.920]  Fantastic.
[28:23.280 --> 28:28.020]  So understanding all the different points of view at the table and the arguments that support them as to why.
[28:29.520 --> 28:31.200]  Not having a dog in the fight.
[28:31.380 --> 28:32.920]  Not having a dog in the fight.
[28:33.100 --> 28:34.200]  So an objectivity.
[28:34.540 --> 28:35.060]  Interesting.
[28:35.320 --> 28:35.600]  Yes?
[28:39.800 --> 28:41.180]  No interest in fighting.
[28:41.360 --> 28:42.080]  Very interesting.
[28:42.820 --> 28:43.620]  Anyone else?
[28:43.740 --> 28:43.960]  Yes?
[28:54.170 --> 28:55.170]  Good question.
[28:55.390 --> 28:57.210]  Does anyone not have a dog in the fight?
[29:04.860 --> 29:05.840]  Mm-hmm.
[29:06.640 --> 29:06.940]  Fair.
[29:08.220 --> 29:09.280]  Well, everyone.
[29:10.180 --> 29:11.060]  Great answers.
[29:11.380 --> 29:14.440]  I'm going to add a few myself just to weave it into the talk.
[29:14.440 --> 29:17.980]  I'm going to say they can perceive of multiple points of view.
[29:18.280 --> 29:20.420]  That's kind of what this gentleman over here was talking about.
[29:21.040 --> 29:22.440]  Understanding the big picture.
[29:23.040 --> 29:25.200]  Are able to make accurate predictions.
[29:25.740 --> 29:33.400]  And then I also want to add, and I think this is actually very important, empathize and comprehend emotions, values, and culture.
[29:33.760 --> 29:39.680]  So I actually think that feelings are a very, like, important distinguishing feature to wisdom.
[29:39.680 --> 29:50.080]  I think a wise one, for example, may understand the frustration of a young Padawan, predict the pettiness of an enemy, or empathize with the devastation of the loss of another.
[29:51.140 --> 29:56.840]  And I am wondering if inducing wisdom must include inducing a comprehension of feeling.
[29:57.160 --> 29:59.440]  And that's where we get to that kind of, like, tricky part.
[29:59.440 --> 30:03.160]  Because removing kind of feelings is a way of removing bias to knowledge.
[30:03.360 --> 30:11.520]  And yet, if comprehension of feeling is important to that greater understanding, that wisdom, how do we go about doing that ethically and carefully?
[30:13.220 --> 30:20.640]  Especially when, sometimes, the feelings that are necessary in order to understand a situation can be, like, really harmful and traumatizing.
[30:20.640 --> 30:22.680]  Or actually just, like, impossible to understand.
[30:23.460 --> 30:28.960]  So perhaps, to a certain extent, fiction can actually be a safe, wisdom-inducing psychotechnology.
[30:29.320 --> 30:34.200]  By reading or watching a story, we can gain understanding, compassion, empathy, and insight.
[30:34.400 --> 30:35.940]  And it may improve our judgment.
[30:35.940 --> 30:41.480]  I can at least go far as to say that media can be a feeling-inducing thing.
[30:41.720 --> 30:46.520]  We can even be forced to feel things based on completely imaginary things.
[30:46.820 --> 30:52.620]  I simply, like, cannot not cry when I, like, watch one of those wholesome Pixar movies.
[30:52.760 --> 30:54.740]  Is anyone bawl like a baby at Pixar?
[30:55.040 --> 30:56.240]  There's a few, yes.
[30:56.860 --> 30:57.740]  I'm on a plane.
[30:57.740 --> 30:58.500]  I don't want to cry.
[30:58.500 --> 30:59.640]  And it forces me.
[30:59.780 --> 31:00.620]  I can't help it.
[31:00.680 --> 31:01.840]  It just strikes me with it.
[31:01.840 --> 31:09.000]  And then also, I will also forcibly acquire tons of nightmare fuel if I watch something like Texas Chainsaw Massacre.
[31:09.220 --> 31:16.340]  So, like, the vividness of media, which it seems like Hollywood has really mastered, can induce a comprehension of something.
[31:16.360 --> 31:20.160]  But what's kind of odd about that is that it can be completely irrational.
[31:20.160 --> 31:22.140]  It can be based on completely imaginary things.
[31:22.160 --> 31:24.360]  And maybe that's not the thing we want to be angling toward.
[31:24.360 --> 31:30.540]  But maybe there are things that we can borrow from film and vivid visual media to induce comprehension of other things.
[31:32.380 --> 31:32.900]  Yeah.
[31:33.140 --> 31:47.560]  And the thing that I'm concerned about, especially when we're dissecting documentaries and the videos that people make when they're warring over political issues, that slowly what that's doing is building up strong emotional priors that may be really difficult to overcome.
[31:47.560 --> 31:54.520]  We saw earlier in the beginning, priors can be, like, very impactful in terms of how we perceive things.
[31:54.680 --> 31:59.920]  And so, like, should we be more careful about the type of media that we're absorbing into our psyche?
[32:00.620 --> 32:03.820]  I think art, to a certain extent, is provocative and irresponsible.
[32:04.380 --> 32:09.760]  So, thank goodness a lot of movies, at least in Hollywood, have trigger warnings so we can avoid the nightmare fuel.
[32:11.180 --> 32:21.020]  But I also think that there's an opportunity to start translating complex civic knowledge into vivid, rich media interactions that could induce emotional comprehension and potentially wisdom.
[32:21.240 --> 32:26.380]  Through visual stimulation, story, and media, perhaps we could hack our way to opt in wisdom.
[32:26.580 --> 32:27.840]  What would be the costs?
[32:28.060 --> 32:28.980]  Would there be trauma?
[32:29.260 --> 32:31.780]  Could we make it an informed, consensual process?
[32:31.780 --> 32:36.900]  Does some designer have to know what wisdom is meant to be derived or isn't that coercive?
[32:37.040 --> 32:41.160]  Or do we just need to make knowledge vivid, visceral, experiential?
[32:41.340 --> 32:42.380]  What are the pitfalls?
[32:42.660 --> 32:45.400]  Will it be so powerful that we'll need emotional trigger warnings?
[32:45.620 --> 32:47.040]  Where could this go wrong?
[32:47.200 --> 32:48.920]  Who decides what gets amplified?
[32:49.200 --> 32:51.220]  Or should anyone do anything?
[32:53.100 --> 32:58.780]  In our database about Diablo Canyon, for example, there are arguments about the degradation of pipes.
[32:58.780 --> 33:04.740]  And so would it be ethical for us to highlight and suggest that readers definitely look at these photos?
[33:05.020 --> 33:08.060]  How do we know that this wouldn't be disproportionately persuasive?
[33:08.280 --> 33:16.540]  Are the visuals important to comprehending the nature of the perceived and argued threat of the state of the nuclear power plant?
[33:16.540 --> 33:23.640]  And then in the database also, we have a long timeline of incidents when PG&E's negligence has led to death and destruction.
[33:23.980 --> 33:35.620]  Is it important to those who are advocating that Diablo Canyon should remain open but not be managed by PG&E that people look at these videos of the fires and see the faces of the people who have died in explosions and disasters?
[33:35.620 --> 33:37.280]  Or is that too traumatizing?
[33:37.480 --> 33:39.180]  Is that extraneous and unnecessary?
[33:39.600 --> 33:44.440]  Is it too biasing because of how impactful feelings are in our sense-making and decision-making?
[33:44.700 --> 33:50.420]  There are so many pitfalls, so many ways to err into bias perception, which is totally antithetical to our mission.
[33:50.760 --> 33:54.220]  Even a single word can accidentally add undue weight.
[33:54.380 --> 34:01.120]  And then there is the impossibility of accounting for how people's priors can shape how people perceive the things that they are seeing.
[34:01.120 --> 34:03.680]  I can't imagine it all can be accounted for.
[34:04.120 --> 34:10.020]  And yet, if we even marginally increase the wisdom of people in general, what kind of impact would that have?
[34:11.420 --> 34:14.820]  Well, I'm afraid I probably left you with way more questions than answers.
[34:15.050 --> 34:15.610]  Sorry about that.
[34:15.740 --> 34:17.050]  It was, again, a very late night.
[34:17.940 --> 34:21.570]  But I think that we do have a series of important questions to consider.
[34:21.780 --> 34:33.480]  Like, is it possible to augment human intelligence to help people overcome biasing priors in order to see problems and solutions in new light, but also in a manner that isn't meant to just overtly coerce or persuade?
[34:34.090 --> 34:40.590]  If we were a design technology to help us hack our way to wisdom, what kinds of emotional costs and demands would that require?
[34:40.590 --> 34:52.920]  I'm not sure we at the Society Library would ever dare to assume that we would know the way to wisdom, but beyond enabling comprehension of the knowledge we're working with today, it seems like the highest ideal of knowledge management work.
[34:53.070 --> 34:57.550]  So we would love to hear any ideas you may have, hackers, and thanks for listening.
[35:05.260 --> 35:08.120]  I think we have time for questions in case any of y'all have any.
[35:12.830 --> 35:13.310]  Yes?
[35:18.910 --> 35:23.890]  Well, my job is pretty much not to have opinions on the things that we map.
[35:24.490 --> 35:32.430]  And I will say Diablo Canyon is the first collection we will actually be releasing, because it is the only one that has reached what we call meaningful completion.
[35:32.830 --> 35:40.310]  Climate change, COVID-19, election integrity, these other topics that we've worked on, are unbelievably enormous.
[35:40.310 --> 35:44.750]  And what the Society Library is really trying to do is set a new kind of standard for rigor.
[35:45.090 --> 35:50.550]  I mean, like, there's like a threshold of how rigorous people are when they're forming their arguments.
[35:50.750 --> 35:53.470]  Scholarly papers tend to perform the best that I've seen.
[35:53.650 --> 35:57.290]  Government documentation, news and journalism kind of falls underneath.
[35:57.290 --> 36:03.210]  I'm constantly shocked by, like, substandards of people not sourcing their references and things like that.
[36:03.210 --> 36:16.190]  So the Society Library is really trying to just, like, introduce a new level of rigor, a new almost kind of language for how to talk about our problems that is so ridiculously comprehensive that we're really trying to raise the bar.
[36:16.290 --> 36:23.990]  So the Diablo Canyon issue is a state-level issue and is something that we've been able to fundraise enough to reach meaningful completion and are releasing.
[36:24.210 --> 36:27.730]  Everything else still needs a lot more work because they're international issues.
[36:27.730 --> 36:29.190]  They are enormous.
[36:29.190 --> 36:30.530]  Yes?
[36:35.820 --> 36:39.920]  Societal impact, those tend to be somewhat ephemeral.
[36:40.540 --> 36:46.000]  So, like, maybe it lasts for a decade, maybe 50 years, but probably not 100.
[36:48.180 --> 36:52.940]  You're also tackling a very, very small portion of society currently.
[36:53.540 --> 37:05.740]  So, given that you're currently able to nibble a small amount and it's going to expire, how would you scale this up to something that's generally useful?
[37:05.960 --> 37:08.400]  Yeah, that's a fantastic question, and you're so perceptive.
[37:09.240 --> 37:11.980]  I don't know how much information I revealed, but you hit the nail on the head.
[37:12.080 --> 37:17.340]  Right now, we are working on issues of contemporary importance that are oftentimes fleeting.
[37:17.580 --> 37:25.440]  Right now, we're focused on high-impact, persistent, polarizing issues in the English-speaking United States, having to do with human rights, resource management, et cetera.
[37:25.440 --> 37:28.820]  But really, the Society Library is a very long-term institution.
[37:29.080 --> 37:36.320]  I'm concerned with... I mean, we've named ourselves the Society Library because I'm concerned with the knowledge infrastructure that future generations are going to be growing up with.
[37:36.480 --> 37:48.620]  We have old institutions that have not yet been fully digitized, and yet, if we just directly digitize them, which is a lot of what we've seen, we've directly digitized the newspaper, we've directly digitized the book, we've maintained a lot of the structure and format,
[37:48.780 --> 37:50.060]  I think that's inherently limiting.
[37:50.060 --> 37:53.780]  So what the Society Library is looking to do is, what are our information needs?
[37:53.800 --> 38:02.120]  And what are the technologies that we have at our disposal to just greatly amplify the ability to induce comprehension, account for comprehensiveness, yada, yada.
[38:02.280 --> 38:06.180]  So really, the work that we're doing now is because it falls within our mission.
[38:06.420 --> 38:10.280]  People are willing to fund it because they're interested in near-term, present issues.
[38:10.320 --> 38:18.000]  But really, a lot of it is just R&D to developing the pipeline for ingesting knowledge, processing knowledge, the ontological development.
[38:18.000 --> 38:34.480]  And long-term, I think, potentially even in my lifetime, because of how unbelievable AI is accelerating with chain-of-thought reasoning and things like that, I think we'll be able to see live libraries being generated in my lifetime, which would be an absolute dream come true.
[38:34.640 --> 38:36.060]  But there is a number of different issues.
[38:36.160 --> 38:39.800]  If you go to our website, you can see the full scope of the different topics that we would cover.
[38:40.140 --> 38:46.800]  And I think also that some issues are probably going to be persistently problematic, especially around values and culture.
[38:46.800 --> 38:49.240]  So those ones are things that we also cover.
[38:49.620 --> 38:51.620]  And most likely, we'll be mapping those things for a while.
[38:52.420 --> 38:53.180]  Yes, yes, yes, yes.
[38:55.800 --> 38:56.320]  Go ahead.
[38:56.440 --> 38:56.560]  Yes.
[39:07.520 --> 39:07.880]  Biases.
[39:08.040 --> 39:09.680]  Oh, so we're actually radically inclusive.
[39:09.920 --> 39:12.980]  So we have 22 different methods for overcoming our own research biases.
[39:13.320 --> 39:20.320]  And essentially, the service that the Society Library is rendering, or one of the services, is that we undertake the work of doing a lot of steel manning.
[39:20.320 --> 39:31.740]  So like, even if something seems like really obscure or extraneous, we, and like low quality content, we treat those like hints, and we spend a lot of time steel manning that and trying to find evidence for it.
[39:31.800 --> 39:49.780]  Like, for example, there was this one claim in the Diablo Canyon database, people persistently on Facebook and in blogs and even in news articles would just persist this idea that 1.5 billion fish would get, I believe, was entrained and like boiled and killed every year in the Diablo Canyon cooling system.
[39:49.780 --> 39:52.220]  And we just couldn't find evidence for it.
[39:52.280 --> 39:56.020]  And I was beginning to think like, perhaps it was just a persistent myth, we couldn't track down the providence.
[39:56.640 --> 39:58.900]  Eventually, we found where the data came from.
[39:59.180 --> 40:05.980]  And what's really wonderful is that now there's like this centralized source where there's so many different people who have referenced that number.
[40:06.120 --> 40:08.020]  And now there's a citation that they can just link to.
[40:08.480 --> 40:11.160]  So we actually spend a lot of time like steel manning content.
[40:11.160 --> 40:18.240]  We've got virtues and values on our page that guide our orientation towards the information that we're interacting with.
[40:18.620 --> 40:22.260]  Essentially, we consider our duty to leave no intellectual stone unturned.
[40:22.400 --> 40:25.360]  And in general, the culture of libraries is to be radically inclusive.
[40:25.820 --> 40:30.000]  I encourage you all to check out the American Library Association's Library Bill of Rights.
[40:30.380 --> 40:31.680]  Libraries are super radical.
[40:32.020 --> 40:33.880]  They're like ridiculously anti-censorship.
[40:34.140 --> 40:35.140]  It's like pretty awesome.
[40:35.580 --> 40:39.460]  So we borrow a lot of that culture and have imbued it in our own research methods.
[40:39.820 --> 40:54.980]  And we have found consistently time and time again, that even if someone immediately discounts an idea or like, we hear someone bring up an idea that sounds ridiculous, if you just spend a little bit of time digging, maybe the way in which that claim was expressed wasn't the best representation of the sentiment,
[40:55.240 --> 40:56.460]  and is technically wrong.
[40:56.460 --> 41:03.300]  But if you do a little bit of digging, like, it may just actually be a representation, a malformed representation of something that is substantial.
[41:03.380 --> 41:05.920]  And it probably is important that that could get surfaced.
[41:06.080 --> 41:12.860]  And we're grateful to the people who, whether it's on a conspiratorial website or an unsighted rant online, for bringing that to our attention.
[41:15.960 --> 41:16.420]  Yes?
[41:35.680 --> 41:41.500]  Thank you.
[41:43.220 --> 41:44.560]  Yes?
[42:11.610 --> 42:11.990]  Great.
[42:12.230 --> 42:15.530]  Okay, so let me try and restate the gist of what you asked.
[42:16.030 --> 42:29.710]  So correct me if I'm wrong, but you are asking essentially, you know, how do we balance like arguments about moral conscience and values and things like that in comparison to collecting facts and data and whatnot.
[42:29.990 --> 42:30.870]  Is that about right?
[42:38.180 --> 42:39.000]  Yeah, great.
[42:39.460 --> 42:40.200]  So perfect.
[42:40.500 --> 42:45.500]  So there's a very specific reason why I use the words arguments, claims and evidence.
[42:45.760 --> 42:48.380]  And that's because claims can be facts or opinions.
[42:48.620 --> 42:52.320]  They could be moral arguments or they could be, you know, evidence driven arguments.
[42:53.300 --> 42:56.400]  Claims can be fact opinion, right or wrong, doesn't matter.
[42:56.800 --> 42:57.660]  They're just claims.
[42:57.840 --> 42:59.420]  They're assertions that something is true.
[42:59.640 --> 43:04.860]  And someone can assert that it is morally right or morally true that something is done or not done.
[43:04.860 --> 43:16.120]  And in fact, for the Diablo Canyon content in that position one, almost every single position has an entire ethics section, where people argue about the ethics of something, which can be moral, it can be religious.
[43:16.260 --> 43:20.480]  I don't know if I know of any particularly like religious arguments in the nuclear PowerPoint one.
[43:20.480 --> 43:31.840]  But it's, again, we're very interested in being radically inclusive because what we're trying to do is create a one centralized repository library where someone can say, listen, I really want to know all dimensions of this issue.
[43:32.140 --> 43:38.440]  And all dimensions definitely includes how people feel about it, the opinions they have about it, their morals, their values.
[43:38.580 --> 43:40.300]  It has a specific section, right?
[43:40.300 --> 43:46.020]  And it's identified as such, unless there happens to be evidence to support it, and then bam, it's an evidence driven claim.
[43:46.800 --> 43:51.880]  But those are important dimensions to understanding ourselves as a collective and functioning as a democracy.
[43:52.220 --> 43:54.180]  People cannot read those things if they choose.
[43:54.360 --> 43:56.540]  They don't have to read everything in the library if they don't want to.
[43:56.760 --> 43:58.920]  But I think it definitely deserves representation.
[43:59.340 --> 44:03.640]  And so how the second part of your question, how do we promote that?
[44:03.640 --> 44:13.860]  I mean, it's our human right, actually, as defined by the United Nations declarations of human rights to have opinions, and also have religious beliefs.
[44:13.920 --> 44:15.720]  I believe it's Article 18 and 19.
[44:15.960 --> 44:26.240]  So it is inherently your human right to have an opinion, have a belief, and like, it doesn't need to be informed, and no one has the right to really, like, challenge or try to coerce or change that.
[44:26.240 --> 44:36.580]  Which is an ideal that I don't think we have fully realized, because I think in order to realize the freedom of, like, having opinions and religious beliefs, I think people should have the freedom of choice to begin with.
[44:36.760 --> 44:38.920]  I think a lot of people inherit their religious beliefs.
[44:39.200 --> 44:40.520]  Nothing necessarily wrong with that.
[44:40.720 --> 44:54.820]  But I think that even a library of religious perspectives, as it relates to, like, fundamental philosophical questions about the nature of reality life in the universe, could be something that people could have access to so they can more freely choose what they want to believe.
[44:54.820 --> 45:00.000]  And then at least that's, like, a free choice, as opposed to, I've inherited this, no one tried to take it away from me.
[45:00.140 --> 45:03.480]  But ultimately, like, it's all of our human rights to believe whatever the heck we want to believe.
[45:06.560 --> 45:06.800]  Five minutes.
[45:06.940 --> 45:07.100]  Okay.
[45:07.320 --> 45:07.980]  One more question.
[45:08.080 --> 45:08.260]  Yes.
[45:08.260 --> 45:09.600]  Five minutes.
[46:08.190 --> 46:10.110]  Your question is about undue weight, right?
[46:10.250 --> 46:12.590]  And like platforming things and giving false equivalents.
[46:12.690 --> 46:21.730]  Does something look like it has weight, it has validity, just because it's right next to something that, you know, has a bunch of evidence and things like that.
[46:21.730 --> 46:25.930]  So I don't know if you saw in the example, but we actually do label claims.
[46:26.050 --> 46:33.290]  So if something has no evidence, there's a whole taxonomy of different things where we describe something that we haven't been able to find evidence for.
[46:33.510 --> 46:38.070]  It could be that no evidence was provided, no evidence was found yet, we need to look for evidence.
[46:38.190 --> 46:45.310]  So there's like a whole like set of ways in which we label and contextualize data because we are also concerned with that.
[46:45.310 --> 46:50.370]  If people are just like reading at face value all of these claims and absorbing them, right?
[46:50.470 --> 46:56.850]  What is that doing to their priors, especially if they happen to already like have an inclination to believe a specific thing?
[46:56.850 --> 46:59.450]  So we actually put notes in front of things.
[46:59.450 --> 47:09.950]  So people see like, this is logical fallacy, this is cherry pick data, this is outdated, we've not been able to find any evidence for this, this is so vague, it's actually not fact checkable, like we let people know.
[47:10.170 --> 47:16.150]  And then as you unpack things, you can also see, you know, counter argumentation and what the counter argumentation is made out of.
[47:16.150 --> 47:30.690]  So I mean, this visualization, the paper visualization is one of our first iterations of how to express this knowledge, I have no doubt, and it is my hope that we will partner with designers that will make these things more obvious, because this is all very important,
[47:30.850 --> 47:34.930]  right, and this is kind of like getting to the gist of the talk is like, how do we go about enabling comprehension?
[47:34.930 --> 47:47.690]  There's so many facets of absorbing knowledge that are really important that we really need to be careful about in order to get as much knowledge in without like causing someone to end up at a certain conclusion or manipulating their thoughts.
[47:47.870 --> 47:56.590]  So it's like, just I think we're at the dawning of like, a new era, an enlightenment era of knowledge management and consolidation and presentation.
[47:57.270 --> 47:59.910]  And yeah, I'm hoping there's just so much more to come.
[48:00.070 --> 48:01.370]  So thank you so much for your concern.
[48:01.370 --> 48:02.350]  We also have it.
[48:04.930 --> 48:05.830]  Yes, quickly.
[48:16.080 --> 48:18.080]  Sorry, can you say that one more time?
[48:36.480 --> 48:41.600]  So our, we don't necessarily drive anyone to a conclusion.
[48:41.940 --> 48:46.040]  However, the Society Library does build political decision-making models.
[48:46.080 --> 48:47.640]  So we work at the city council level.
[48:47.640 --> 48:52.540]  We essentially execute our method by extracting arguments, claims and evidence from various forms of media.
[48:52.740 --> 49:02.060]  And instead of rendering it as debate maps, what we do is we break down every single dimension of a decision, which can be along categorical lines like economic or social or whatever.
[49:02.060 --> 49:09.180]  Sometimes it can be much more precise, like those high level, like vague orientations underneath economic and safety and those things that you saw.
[49:09.640 --> 49:14.840]  And what we do is we show them, here's the packet of pro-con argumentation in each one of those dimensions.
[49:15.080 --> 49:17.940]  And then we visualize that so they essentially micro-vote.
[49:18.060 --> 49:20.920]  I've got a talk on our website about how we do this.
[49:20.920 --> 49:24.700]  And we actually give away a toolkit for making your own decision-making models.
[49:24.700 --> 49:32.040]  So one of the things that we do want to do is take our decision-making model process, digitize that, and then make that an interface through which our data can be rendered.
[49:32.280 --> 49:37.920]  So we're not going to tell anyone the decision that they should make, right, because some of the stuff is pretty subjective and value-driven.
[49:38.080 --> 49:39.160]  Like, do you value safety?
[49:39.160 --> 49:41.380]  Do you not care about the reputation of the utility?
[49:41.680 --> 49:43.020]  Like, that's kind of subjective.
[49:43.180 --> 49:49.140]  But if someone actually wanted to go through the process of methodically going through all of this knowledge, we happen to have a process for that.
[49:49.140 --> 49:50.800]  And it just takes them through step-by-step.
[49:50.800 --> 49:53.340]  So we're hoping to raise some funds.
[49:53.500 --> 49:58.680]  Actually, we've got proposals out for that to be the next iteration of an interface for people to opt into.
[49:58.840 --> 50:06.840]  And what's really cool about that is that once people go through the process, they don't have to remember the decision that they made in the black box of their minds going through potentially thousands of arguments.
[50:07.060 --> 50:08.460]  Instead, it mirrors back to them.
[50:08.700 --> 50:14.740]  Once you have gone through the process of looking at every dimension, this is what you decide along economic, environmental safety lines.
[50:14.940 --> 50:15.800]  And that's my time.
[50:15.980 --> 50:16.740]  Thank you all so much.
[50:21.460 --> 50:27.220]  People probably want to keep talking with you and just take your stuff out in the hallway as long as you want.
[50:27.460 --> 50:27.760]  Great.
[50:27.980 --> 50:28.200]  Thank you.
[50:28.400 --> 50:28.600]  Thank you.
[50:28.900 --> 50:29.280]  Thank you.