Latest / From First Principles / China’s AI Breakthrough, Time Crystals, Hidden Viruses, & Brightest Cosmic Signal (EP. 9)

Transcript

China’s AI Breakthrough, Time Crystals, Hidden Viruses, & Brightest Cosmic Signal (EP. 9) · From First Principles

• 0:00
  Hello Internet, this is your captain speaking Lester Nare, joined as always by my co-host and our resident PhD Krishna Chowdhury.
• 0:07
  This is From First Principles.
• 0:10
  We have a great episode for you this week.
• 0:11
  We're going to cover four stories in AI, space and immunology, starting with a new story out of China's DeepSeek, which says its AI model costs just $294.99.
• 0:24
  $1,000 to train.
• 0:26
  This is really important, and we're gonna go a deep dive into why.
• 0:29
  Followed up by our second story, sounds like we may have discovered the Ocarina of Time.
• 0:33
  We may need to get Zelda involved.
• 0:36
  As scientists just made the first time crystal that you can see.
• 0:40
  I'm really interested in talking about that one.
• 0:43
  The third story is about hidden viruses in our DNA that could be medicine's next big breakthrough.
• 0:49
  And we're going to end with space with the brightest fast radio burst ever detected and how it could help us solve an enduring cosmic mystery.
• 0:58
  This is From First Principles.
• 1:17
  My friend.
• 1:18
  How's it going?
• 1:19
  Episode 9.
• 1:20
  Episode 9.
• 1:21
  10k followers on Instagram.
• 1:22
  Woo!
• 1:24
  We still haven't been shut down by the FCC yet.
• 1:26
  Yeah, yeah, yeah.
• 1:27
  It's coming, though.
• 1:27
  It's coming.
• 1:28
  We're talking too much science.
• 1:29
  It's coming.
• 1:30
  It's coming.
• 1:30
  They're going to defund us and then ban us.
• 1:32
  But before they do, we're going to talk about some great stories.
• 1:36
  The first one, we always end up touching on AI.
• 1:39
  There's a lot of research going on in AI in a variety of ways.
• 1:44
  And our first story was a story in Reuters that also had two nature papers that were related to it.
• 1:52
  Yeah.
• 1:52
  The headline on this is China's DeepSeek says its AI model has hit a cost of just $294,000 to train.
• 2:02
  The sort of byline is in a paper that is likely to reignite the debate over Beijing's place in the race to develop AI.
• 2:09
  Yeah.
• 2:10
  This sort of cost, which is significantly lower than the figures that we see in the US.
• 2:16
  And I know we've talked about DeepSeek and all of this quite a bit.
• 2:20
  Yeah.
• 2:21
  Uh, it's a, it's a crazy model.
• 2:24
  And the fact that the team like put out a peer review paper, you know, no one else has done that in the field of AI.
• 2:30
  Uh,
• 2:31
  like gone to nature and put out, okay, this is how we trained.
• 2:34
  This is how we did it.
• 2:37
  It is, you know, it is, they're actually, it's, it's kind of crazy that China is setting the precedent for open science in the field of AI.
• 2:45
  Right.
• 2:46
  It is very interesting.
• 2:47
  And not only one, but two papers in, in nature, main brand, not, not a nature sub, you know, sub category.
• 2:54
  I triple E. This isn't even like, you know, Neurips.
• 2:57
  This is nature.
• 2:58
  And,
• 2:59
  And what I thought would be helpful, we're going to do sort of two aspects on this story.
• 3:04
  Yeah.
• 3:05
  Folks might have seen the Deep Seek news earlier in the year when there was sort of a big outcry about it.
• 3:11
  Yeah, they released in January, I think.
• 3:13
  Yes.
• 3:13
  And then all of our stocks failed.
• 3:15
  Yes, it was a big deal.
• 3:17
  Yeah.
• 3:17
  But January wasn't the start of this great power war over AI.
• 3:22
  Right.
• 3:23
  And I thought it would be good to just go for a couple of beats on the sort of geopolitical context of
• 3:29
  that this research paper kind of came out in.
• 3:31
  Yeah, yeah.
• 3:32
  And to sort of set the stage, right, there's an AI great power competition between the U.S. and China.
• 3:38
  AI, as we know, it has existed for quite some time, you know, before all this generative AI stuff.
• 3:44
  Yeah.
• 3:46
  But in the sort of recent context, we're going to start in 2017, where China launched its next generation AI plan, which said it was going to lead the world in AI by 2030.
• 3:58
  They're famous for their five-year plans.
• 3:59
  Right.
• 4:00
  Right.
• 4:00
  And the idea is U.S. has been the global hegemon forever.
• 4:03
  China's looking at a way in which it can sort of destabilize that control.
• 4:08
  And Xi Jinping, leader of the Chinese Communist Party, said that advanced tech is the sharp weapon of modern states.
• 4:16
  Yeah, that's true.
• 4:17
  And as we've discussed on this podcast previously, there is a very strong relationship between the military intelligence and national security apparatus within nation states and frontier technologies.
• 4:29
  Yeah.
• 4:30
  AI just being one of those.
• 4:31
  Yeah.
• 4:33
  And then in 2021, what was interesting is the US National Security Commission on AI, which was chaired by the former Google CEO Eric Schmidt, warned that American dominance in AI was not guaranteed.
• 4:46
  So we already have this setup of that friction.
• 4:49
  Wow.
• 4:49
  Now, if we go into the Biden era policies, because the US government, as we know, as it relates to technology innovation, is usually reactive as opposed to proactive.
• 4:59
  Yes.
• 4:59
  Yes.
• 5:23
  The Biden era was the first administration to take policy in earnest in terms of AI specifically.
• 5:29
  Okay.
• 5:29
  And this started in August 2022 when Biden signed the Chips and Science Act, which many people have probably heard the Chips Act.
• 5:36
  Yeah, the Chips Act.
• 5:37
  $250 billion bipartisan bill aimed at boosting U.S. semiconductor manufacturing and innovation.
• 5:42
  This dovetails with the re-onshoring of manufacturing conversation post-COVID.
• 5:47
  There was $52.7 billion in supplemental appropriations that was specific to semiconductor-related programs for 2023 through 2027.
• 5:56
  The idea being that it was also going to create guardrails.
• 6:00
  preventing recipients like Intel and TSMC from expanding their advanced fabs in China.
• 6:08
  So we can see the theory of the case here was, can we create a choke point at the hardware level?
• 6:14
  Because these chips were viewed at the time as being the key aspect of the stack that was gonna be the driver of progress and innovation.
• 6:23
  And TSMC is in Taiwan, which is like now iffy if China's going to do something, right?
• 6:29
  Correct.
• 6:30
  China's been doing a huge military buildup and training around this idea of we're going to take Taiwan.
• 6:35
  Yeah, yeah.
• 6:36
  That's basically their thesis.
• 6:37
  We're going to take Taiwan and you don't want to be here when that happens to us, the U.S. Yeah, yeah.
• 6:43
  And so there are a lot of layers to this.
• 6:46
  The Taiwan point's a really important one.
• 6:48
  Now, 2022, we're still in 22 August, CHIPS Act drops.
• 6:53
  And then October 7th, 2022, the October 7th controls were established, which are export controls.
• 7:00
  Now, they named it this prior to the unfortunate October 7th of 2023, which is the Israel-Gaza conflict.
• 7:08
  But at the time, this was a restriction on exports of specific NVIDIA and AMD chips.
• 7:15
  The H100s, A100s, and the AMD Mi250s, which basically meant this was the whole, we have a small yard, high fence.
• 7:25
  Yeah.
• 7:26
  Making it really, really difficult to get these sort of frontier chips.
• 7:29
  Yeah, and the thesis was that, like, these chips are the future.
• 7:32
  Yes.
• 7:32
  And without these chips, you can't do shit.
• 7:34
  Exactly.
• 7:34
  That was the theory of the case at the time.
• 7:36
  Yeah.
• 7:37
  As a small side note, it's interesting that a month, almost just over a month later, on November 22, 2022, ChatCBT launched.
• 7:48
  Ah, okay.
• 7:48
  So all this conversation was happening prior even to the meteoric rise of the first real production level LLM generative AI accessible to millions of people.
• 8:02
  And the growth curve, as we all know.
• 8:03
  I mean, ChatGPT was the moment where now everyone was like, oh, AI is here.
• 8:10
  Yes.
• 8:10
  It's not just this thing people are saying and da da da da.
• 8:13
  It's not theoretical.
• 8:15
  I can literally chat with it.
• 8:17
  Yes.
• 8:17
  And it's like giving me amazing answers.
• 8:19
  Yes.
• 8:19
  Yeah.
• 8:20
  Yes.
• 8:20
  Yeah.
• 8:20
  So now we're moving into 2023.
• 8:23
  Right.
• 8:24
  So throughout 2023, now the Biden administration is not only trying to create the choke points domestically, but now we're going out to our allies all across the world to sort of expand that tightening.
• 8:37
  So this included in the Netherlands with ASML.
• 8:41
  And in Japan with Nikon and Tokyo Electron blocking them from providing China access to their lithography tools, which are used to actually build and generate these more advanced chips.
• 8:53
  Yeah.
• 8:54
  This was followed up by in August of 23, Executive Order 14105, which now also restricted investment from U.S. investors into any China AI or quantum firm.
• 9:07
  Okay.
• 9:11
  like choking all pathways right to hardware innovation yeah for hardware and for hardware innovation okay right so while this was all happening in 23 uh interestingly enough uh there was the very very sort of public falling out of the founder or ceo at open ai sam altman so many
• 9:34
  folks may remember there was that huge board battle i forgot about yeah dude he was ousted he was ousted yes and this was around open ai was a non-profit they were doing a lot of for-profit moves they wanted to kind of rejigger their structure
• 9:50
  And it kind of signified the realization across the Silicon Valley consensus that this thing was going to be bigger, more powerful, more impactful, more money generating than even they had predicted.
• 10:04
  It was happening quicker than they expected.
• 10:07
  So there was that power struggle.
• 10:08
  Right.
• 10:09
  Ultimately, the king returned to the castle.
• 10:11
  Yeah.
• 10:12
  And Sam Reid got his position as CEO.
• 10:16
  But all this we've talked about so far has been the U.S. side.
• 10:18
  Yeah.
• 10:19
  So in 23, China's response, what has China been doing as the U.S. That's right.
• 10:23
  So, yeah, we've been battling internally, so to speak.
• 10:27
  And to sort of create this regulatory regime around access to chips.
• 10:31
  Right.
• 10:31
  So what's China's response?
• 10:33
  So China had firms like Huawei introduce their own, you know, what they call frontier chips with the Ascend AI chip, as well as shocking the global community when SMIC released their 7 nanometer process.
• 10:53
  for smartphone chips, which again, the idea was TSMC has really had a monopoly on the seven, five, and I believe three now, nanometer chip sizes.
• 11:02
  Their latest is three, but like sub 10 was them.
• 11:05
  Right, and no one else could really do it.
• 11:08
  And that was like a really big bottleneck.
• 11:10
  both for, and there's both operational reasons and proprietary intellectual property reasons why that's difficult.
• 11:18
  They also then had their other big conglomerate technology companies, Baidu, Alibaba, begin to put out their own chat GPT type LLMs.
• 11:27
  So we had Quinn, we had Ernie Bot,
• 11:30
  But nobody really... No one really thought anything.
• 11:32
  Nobody really cared.
• 11:33
  No one cared.
• 11:34
  They were all trained on the cut-down NVIDIA H800 GPUs, which were still allowed to be sold to China, but they were massively sort of handicapped.
• 11:43
  Yeah.
• 11:45
  Interestingly, now we're sort of getting into 2024, which is more of the same, but we talked about the intersection of the national security state, the intelligence community, military intelligence, and frontier technologies.
• 11:57
  Yeah.
• 11:58
  And Marc Andreessen made this interesting note in a podcast where he...
• 12:03
  Basically was trying to call out the folks who are called decels or the folks who are really heavy on AI safety.
• 12:10
  Deceleration.
• 12:11
  Yeah, the decelerationists who are saying let's slow down, let's slow our roll.
• 12:14
  Got it.
• 12:15
  Elon is a famous one.
• 12:16
  Elon is arguably the most famous decelerationist.
• 12:20
  And Marc Andreessen's point was, if you believe that this threat is as existential as nuclear weapons, you're still allowing these open AI employees to go to an unprotected, unsecured building in the same way that – not in the same way that, like, nuclear facilities are protected from an operational security perspective.
• 12:40
  So he's like, if you believe this to be true, the entire operating infrastructure of these companies needs to change.
• 12:45
  Yeah.
• 12:46
  That's fair, actually.
• 12:48
  I never thought about that, but yeah.
• 12:50
  It's a fair argument, right?
• 12:51
  Yeah.
• 12:51
  And in 2024, on June 13th,
• 12:56
  After all this board kerfuffle in 23, the former head of the National Security Agency and commander of US Cyber Command, Paul Nakasone, was added to the OpenAI board, which was the first salvo in what I like to call the nationalization of the frontier models in the United States.
• 13:14
  Got it.
• 13:15
  Right now, the national security state has a seat at the table.
• 13:19
  Shortly after, OpenAI announced several projects with the Pentagon, the DOD around using it.
• 13:25
  So this is all background leading up to the Sputnik moment for China.
• 13:30
  That was a big one.
• 13:31
  Which is in January of this year, 2025, when the Chinese startup DeepSeek released their R1 model, which was claiming at the time a Chachapiti 4-level reasoning.
• 13:44
  So this is like our best, most frontier.
• 13:46
  This is pre-ChatGPT5, which we now have today in September.
• 13:51
  Fully trained on NVIDIA H800s.
• 13:55
  Yes.
• 13:55
  The not good chips.
• 13:56
  The not good chips, yeah.
• 13:58
  Using a mixture of experts techniques for the training alongside with reinforcement learning at what we now understand to be a sub $300,000 training cost.
• 14:06
  Now it's important to note the amount of money...
• 14:09
  that Anthropic OpenAI Meta has raised to train, Grok, to train their foundation models is in the order of tens of millions, hundreds of millions of dollars.
• 14:22
  Yeah.
• 14:23
  So this is an orders of magnitude difference.
• 14:25
  Orders of magnitude.
• 14:27
  And so what we ended up creating by doing this choke point on the hardware side was necessity is the father invention or whatever, right?
• 14:38
  China then said, okay, let's innovate on the training model and infrastructure to try to get the same efficacy.
• 14:44
  On the algorithms and the software.
• 14:47
  Not on the hardware.
• 14:48
  So this created a huge...
• 14:53
  surprise from a lot of US observers because the consensus was if we choke off the hardware they got nothing.
• 15:00
  And this totally blew that out of the water.
• 15:02
  So at first there was fascination which was immediately followed by the national security narrative taking over.
• 15:08
  An interesting point about what China did with DeepSeek was it was also, if we exclude LAMA, Meta's LAMA models, which are open source, open weights, but arguably weren't as competitive as R1, China was the first frontier scale open weight model.
• 15:28
  Yeah, yeah.
• 15:28
  They just published their weights.
• 15:30
  Right.
• 15:30
  Which the point here is then you can take it and make it into whatever you want to.
• 15:35
  So the app for DeepSeek became the number one most downloaded app globally.
• 15:39
  Oh, I didn't know that.
• 15:40
  It was then banned in India, Italy, Australia, Taiwan, because the national security narrative came in.
• 15:46
  And it was like, this is a propaganda tool.
• 15:48
  It's a data siphoning tool.
• 15:51
  Don't trust DeepSeek.
• 15:53
  But it strategically left out the point about the fact that you could download the model.
• 15:59
  Yeah.
• 15:59
  Run it locally and change the weights however you wanted to.
• 16:02
  So with DeepSeek, if you tried to look up Tiananmen Square.
• 16:05
  Yeah.
• 16:06
  You weren't going to get an answer.
• 16:07
  Really?
• 16:08
  Right.
• 16:08
  So they were, you know.
• 16:10
  Classic.
• 16:11
  Doing a little bit on the.
• 16:11
  Yeah, yeah, yeah.
• 16:12
  Of course.
• 16:13
  On the information control.
• 16:13
  But it's open weights.
• 16:14
  Yeah.
• 16:15
  So if you want the Tiananmen Square answer.
• 16:17
  You can do it on your own.
• 16:18
  Oh, interesting.
• 16:19
  I see.
• 16:20
  Yeah.
• 16:20
  The narrative really tried to avoid talking about that point.
• 16:24
  At the same time, people started calling open AI closed AI.
• 16:28
  Yeah.
• 16:28
  Because they had never had an open source model.
• 16:32
  They never had a research paper, a peer review paper.
• 16:34
  Right.
• 16:34
  Yeah.
• 16:35
  Around their process.
• 16:36
  So now this brings us to where we are today with these two new papers in Nature, which go a little bit deeper into how DeepSeek went through their process.
• 16:46
  And we'll sort of talk about this.
• 16:48
  But I think that background is important context to understand, like, why does DeepSeek matter?
• 16:54
  Yeah.
• 16:55
  Why is it uniquely different than chat GPT, than Claude, than, you know, Gemini, Gemini, all these other models.
• 17:04
  And why is it important that they actually have a peer review paper?
• 17:08
  We always tell a lot of the U S conversation around science research is, Oh, China's papers.
• 17:14
  They fake the data.
• 17:15
  Yeah.
• 17:16
  They're not really good.
• 17:17
  They're not where we are.
• 17:18
  And I think this, this has been a, a unfortunate, uh,
• 17:24
  viewpoint in the U.S. about China generally, not only with cars saying they just steal our stuff, also with this, that is a little bit outdated.
• 17:33
  It is, definitely.
• 17:34
  Because we're seeing with cars, the best EV cars in the world are made by China.
• 17:38
  It's not arguable.
• 17:39
  They make them cheaper, faster, better, stronger.
• 17:42
  Yeah.
• 17:43
  which we've now put tariffs on in order to say, oh, don't look here, there's no problem, Ford is still the best.
• 17:48
  And AI is taking a similar, in terms of our posture, a similar political stance that China doesn't really know what they're doing, this is not really a big deal, blah, blah, blah.
• 17:59
  I think we need to be careful about ascribing a narrative that may have been true 15, 20 years ago
• 18:05
  Yeah, that's certainly not true now.
• 18:07
  That's certainly not true anymore.
• 18:08
  We've covered Chinese research.
• 18:10
  We've covered the hexagonal diamonds that were first created by China.
• 18:14
  Everyone's been going after them.
• 18:15
  Yes.
• 18:15
  But two labs in China were going after them, and actually one of them succeeded, right?
• 18:20
  Yes.
• 18:21
  And the proof is in the pudding.
• 18:22
  Yes.
• 18:23
  They made it.
• 18:23
  They have all the tests.
• 18:25
  Right.
• 18:25
  And there's peer-reviewed papers.
• 18:26
  And it's out in nature.
• 18:28
  It's out in nature.
• 18:29
  Right.
• 18:29
  And we always talk about, well, this and that.
• 18:31
  So now that we've given that context, let's talk about these two new papers.
• 18:36
  Yeah.
• 18:36
  These papers are pretty incredible in the fact that they're the first peer-reviewed paper from a sort of premier AI large language model.
• 18:48
  Right.
• 18:49
  You have, like, obviously, ChatGPT, Gemini.
• 18:51
  Yeah.
• 18:52
  Grok, all of these large language models, but they haven't actually published anything that shows sort of, OK, that that, you know, I'm sure that the team at DeepSeek has kept some secrets to themselves.
• 19:05
  But the fact that they've actually gone out and published this peer review paper is, I think, quite a watershed moment for the AI community.
• 19:12
  Right.
• 19:13
  Because they're trailblazing this idea that maybe.
• 19:16
  AI is a part of open science and maybe it should be out in the open just like normal science, right?
• 19:23
  Where it's like you as a scientist, you find something cool or you make something cool.
• 19:27
  You put it out in a peer reviewed paper, right?
• 19:29
  And you get let experts actually talk back to you and then like see how things actually work, right?
• 19:38
  You want to get into the nitty gritty of it.
• 19:40
  In the technical details, this paper is quite interesting because it tells us some of the tricks that they used in actually making it possible to train something like DeepSeq on perhaps only H-800s.
• 19:54
  Again, maybe not, right?
• 19:56
  There's always a possibility that they're not totally forthcoming.
• 20:00
  Right.
• 20:01
  But...
• 20:02
  They've sort of laid bare some of their strategies.
• 20:05
  And I think that's quite interesting.
• 20:07
  I want to actually just make two small context notes that I think are interesting here.
• 20:12
  One is that the founder of DeepSeek, his story is that the reason they ended up creating R1 was actually he was doing the trading, like financial trade, like finance.
• 20:27
  Right.
• 20:28
  It was related to a hedge fund.
• 20:29
  It was a hedge fund plan.
• 20:30
  And they were like, we want a more powerful model to utilize for.
• 20:35
  So it wasn't even their core.
• 20:37
  It was an outgrowth of another need.
• 20:41
  And then I want to talk about this open sourcing concept of why China did.
• 20:45
  Why are they doing this open source, open weights?
• 20:48
  Yeah.
• 20:48
  Move.
• 20:49
  There was an interesting there's a podcast, the BG2 podcast with two U.S. investors, Brad Gershner and Bill Gurley.
• 20:59
  And they've been talking about this a lot.
• 21:01
  If you look at China's strategy generally as it relates to if you're in a world where the U.S. already has dominance in a lot of these technologies that are closed source.
• 21:09
  From a how do we remain competitive angle, this is actually what Facebook did by trying to have Lama be open, is the best thing you can do to battle closed source is to create an open source ecosystem.
• 21:22
  Right.
• 21:22
  That now becomes competitive with that closed source because you're going to then now get input from a wider array of experts that invest in your version of the future.
• 21:32
  And so for China, there was an incentive there.
• 21:34
  geopolitically to go open source being a second mover as the means by which to combat the closed source dominance of the us so oh this deep seek model is not an isolated incident of china using open source as a way to backstab u.s's closed source dominance
• 21:54
  Yeah, there's a strategy behind this.
• 21:58
  That is a larger part of how China is trying to spread its influence across these areas in which the current West's control and buy-in is waning.
• 22:11
  Yeah.
• 22:12
  I mean, I think they figure that like the world is a really big place.
• 22:15
  OK. And these developing economies are exactly that developing.
• 22:21
  Right.
• 22:21
  They're getting bigger.
• 22:23
  People are getting more affluent.
• 22:25
  Technologies there are going to start getting better and better.
• 22:28
  And they're going to require AI to keep up with the rest of the world.
• 22:33
  Right.
• 22:33
  And maybe they don't want to pay money.
• 22:36
  whatever royalties, right, to these closed systems.
• 22:41
  I think it's really cool.
• 22:43
  And I think the idea that DeepSea came out, right, with their model in January, I think it was also a watershed moment for the rest of the AI community in terms of...
• 22:54
  finding out the efficacy of something like a pure reinforcement learning approach.
• 22:59
  Okay.
• 23:00
  Rather than some of the techniques that ChatGPT and these other firms had used, which was a supervised fine tuning or this sort of human feedback reinforcement learning, which always requires a human in the loop.
• 23:15
  Yes.
• 23:16
  Right.
• 23:16
  What these guys did was...
• 23:19
  kind of a pure reinforcement learning approach.
• 23:21
  And that's something that they've laid out in this paper in nature.
• 23:24
  Okay.
• 23:24
  The idea is that sort of traditional LLM reasoning models, what they do is they require human input in order to steer them towards the correct answer.
• 23:34
  Right.
• 23:34
  In order to be like, okay, this is, this was not quite what I wanted.
• 23:38
  This is, this is what I wanted.
• 23:39
  Right.
• 23:39
  There were a bunch of like Kenyans that were hired by open AI.
• 23:44
  Yes.
• 23:44
  To basically do this.
• 23:46
  Yes.
• 23:46
  And I don't know how many, but like that was a whole thing back in the day.
• 23:50
  And, you know, that's fine for that.
• 23:52
  That's fine.
• 23:52
  But it's quite expensive in terms of like getting human annotated chain of thought data and so on and so forth.
• 24:01
  But on the other hand, there's actually...
• 24:04
  a disadvantage to using humans, because then you are limited to human reasoning.
• 24:10
  Okay?
• 24:11
  There could be other ways to reason about a problem that wouldn't quite occur to a human being, that perhaps if we gave the entire control to the machine itself...
• 24:22
  the machine could figure out on its own.
• 24:24
  Right.
• 24:25
  So there's an inherent disadvantage that I think these guys at DeepSeek actually like latched on to.
• 24:31
  Right.
• 24:32
  And they figured that what we could do is we could use a pure reinforcement learning strategy where there's a reward signal.
• 24:38
  So the way the reinforcement learning works is you've got your model.
• 24:41
  Right.
• 24:42
  And then it tries to find an answer.
• 24:45
  This works for well-defined tasks where there's a right and wrong answer.
• 24:48
  And there's a really nice way to compare your answer to the right answer and figure out what a reward should be.
• 24:55
  A reward being if my answer is closely aligned to the right answer, then I'm on the right track.
• 25:02
  And the strategy that I've been using is sort of on the right track.
• 25:06
  And otherwise, maybe I need to try a different strategy.
• 25:08
  It's actually very close to how neural systems work, right?
• 25:12
  We've studied reinforcement learning, like, at the biological level when it comes to, like, you know, a rat trying to find cheese in a maze, right?
• 25:23
  We've literally gone into the brain and found that these... How...
• 25:28
  at the nitty-gritty level, at the cellular level, how this stuff works.
• 25:31
  And now we're kind of translating that into the AI neural network space, the artificial neural network space.
• 25:39
  So if we use pure reinforcement learning, what that does is it bypasses human-labeled reasoning steps.
• 25:45
  And then we can actually...
• 25:48
  And explore just computationally the entire paradigm of how to solve the problem, which is, I think, very cool.
• 25:57
  So that's one of the things that's revealed by this paper, the fact that they used pure reinforcement learning.
• 26:04
  It was this autonomous exploration and a development of problem solving strategies.
• 26:09
  And what you could see and what they report in the paper is you could see the network sort of use that RL technique, reinforcement learning, to get better at finding the right answer.
• 26:20
  So what they noticed is...
• 26:23
  For example, the length of the answers would get progressively longer because the neural network wanted to get better and better at finding that right answer in this large language model type of thing.
• 26:37
  They also had this thing called the aha moment, which is when I was like, oh, this is like working way better than we thought, which is when, you know, in the chain of thought, you can actually output the thought that this thing is doing.
• 26:48
  Right.
• 26:49
  And the model would start using the word wait.
• 26:57
  That was never explicitly put into the training regime.
• 27:01
  But what the model would start doing is as it was thinking, if it was trying to get to an answer and it was like sensing that it was rushing things and just trying to get it done quickly, it would use this thing wait.
• 27:14
  And then it would backtrack and try to figure out if there's a better way to rethink what it had just thought and then keep going.
• 27:21
  That's like some meta level weird stuff, dude.
• 27:27
  Like that it's just using this word wait in the context that the word wait means wait and let's try to figure out what we just said in a nice way.
• 27:37
  You know, I think that's kind of crazy to backtrack.
• 27:41
  And then...
• 27:42
  So it developed these autonomous, sophisticated reasoning patterns to try to, you know, solve the problem.
• 27:50
  And it's no longer limited by the human guided approach.
• 27:54
  Right.
• 27:55
  Right.
• 27:55
  Which also has bias issues.
• 27:57
  Yeah, exactly.
• 27:58
  Related to human reinforcement learning.
• 28:00
  Because now, if you have a bunch of Kenyans doing it, there's cultural differences.
• 28:05
  Exactly.
• 28:06
  There's educational differences.
• 28:08
  And so this is where people sort of have concerns about...
• 28:12
  the Silicon Valley elite being the ones controlling the parameters and mind space that these models sort of have.
• 28:22
  But this is like a very pure play version that kind of abstracts away the need for...
• 28:29
  at least in some cases, for the human input to be such a big driver in how it determines answers.
• 28:35
  That's right.
• 28:35
  Yeah.
• 28:36
  So that was the reinforcement learning part, right?
• 28:38
  Which is like, okay, that's a training strategy.
• 28:41
  They also had this philosophy sort of efficiency over everything, right?
• 28:45
  Because of what you talked about where, you know, the powers that be, let's say, have crippled their access to hardware.
• 28:54
  So what can we do?
• 28:55
  We can innovate hardware,
• 28:56
  with algorithms and we can innovate with strategy.
• 29:00
  So there's a two-pronged approach that they at least revealed in this nature paper.
• 29:05
  The first one is this idea of mixture of experts, right?
• 29:08
  This architecture where what you can do is have a giant model.
• 29:12
  The DeepSync model itself is 671 billion total parameters, which is on par with some of the big ChatGPTs and things like that.
• 29:21
  I think ChatGPT5 has like on the order of trillion or more.
• 29:25
  But some of the earlier models have hundreds of billions.
• 29:27
  So it's in the same regime.
• 29:29
  But...
• 29:30
  when it comes to actually inference, the number of active parameters used is only 37 billion.
• 29:36
  Okay, which is 20 times less than the total size of this network.
• 29:41
  Because what they're doing is, they've got this sort of network on top that comes in in the very beginning, and it's kind of a router.
• 29:49
  Okay, what it does is, you've got the 600 billion parameter network, but if you ask it a certain question,
• 29:58
  There's only a certain subset of that network that would be very good at answering that question.
• 30:03
  And that becomes that 37 billion part.
• 30:06
  So they've implemented this router mechanism that takes in the prompt of whatever you put in, and then it sort of routes...
• 30:15
  that prompt to a subset of the network and now inference becomes faster it becomes less expensive in terms of maybe you don't need an h100 gpu right maybe you can do with that h 800 or you know that that previous model and you can actually like get a quick answer yes and it'll be just as good right
• 30:36
  Right.
• 30:36
  So that's one of the strategies that they use.
• 30:38
  The other strategy that they used was during training.
• 30:41
  They had this idea of using a reward system.
• 30:47
  Right.
• 30:47
  In reinforcement learning, you have to have a way of sort of quantifying how much the reward is and then steering the network towards getting more of that kind of reward.
• 31:00
  Right.
• 31:00
  So the way that traditional LLMs have been trained is through this process called proximal policy optimization.
• 31:09
  Okay.
• 31:09
  PPO.
• 31:11
  It's this idea that you have your policy maker, which is your LLM.
• 31:15
  That's like actually outputting the language that gets output whenever you do the chat.
• 31:21
  But then there's this thing on top, which is kind of like the critic model.
• 31:25
  And what that critic model is doing is it's during the training process.
• 31:29
  It is taking the output of the policy, which is the output of the model, and then it's trying to guess at what the reward is going to be.
• 31:40
  Okay?
• 31:41
  And then what you can do is you can use this critic model and the advantage that you get, the advantage being what is my guest advantage at this particular strategy?
• 31:51
  That's what this critic model calculates.
• 31:53
  And what this critic model does is feedback that advantage to the big LLM to tell it, okay, you're on the right track.
• 32:00
  This strategy is working or no, it's not working.
• 32:03
  Crucially, this is very expensive because now you've got two models that you're training, right?
• 32:08
  You've got two models.
• 32:09
  And as the LLM gets bigger and bigger, the critic model has to get bigger and bigger to try to figure out.
• 32:15
  how to calculate that advantage in whatever scenario, whatever game you're playing, whether it's text generation, whether it's whatever, right?
• 32:24
  So if you've got a bunch of H-100s, that's fine, right?
• 32:28
  If you've got billions of dollars of funding from random VC firms, that's fine.
• 32:33
  But if you're DeepSeek,
• 32:35
  That's no longer tenable.
• 32:36
  So what these guys did was they used a different strategy called group relative policy optimization.
• 32:41
  Okay.
• 32:42
  Okay.
• 32:42
  They don't have this second critic model that they're doubly training.
• 32:47
  Right.
• 32:47
  They don't have this 2x parameter blow up.
• 32:51
  What they have is their original LLM.
• 32:54
  And what it's going to do is it's going to output a bunch of different answers.
• 32:59
  Okay, let's say for the sake of argument, let's say 10 different answers.
• 33:03
  And then what it's going to do is it's going to calculate the reward for each which we can get because this is a reinforcement learning strategy.
• 33:09
  And...
• 33:11
  There's going to be a mean and standard deviation.
• 33:13
  There's going to be a distribution that you can calculate about that 10 different strategies.
• 33:19
  There's going to be some that are above the mean, which means that that is better than those others.
• 33:24
  The whole idea is you're trying to get this idea of advantage.
• 33:28
  Which are the strategies that give me advantage and which are the strategies that don't?
• 33:33
  In the ones before, in the traditional LLMs, the ones that we're doing in the States and in the West,
• 33:38
  We've got this separate neural network that is calculating that advantage and telling the original one where to go.
• 33:44
  Here, the original one is coming up with different answers and then creating a distribution out of that and then figuring out, okay, I should go here to this part.
• 33:55
  Whatever I did here, that's what's good.
• 33:56
  Whatever I did here, that's what's bad.
• 33:58
  Because I came up with eight, ten different answers.
• 34:01
  These guys were bad.
• 34:01
  These guys were good.
• 34:02
  Yes.
• 34:03
  Right?
• 34:03
  So you don't need this second giant neural network to train.
• 34:07
  Right?
• 34:08
  So it's a really interesting...
• 34:11
  strategy where you're choked by the hardware.
• 34:15
  Yes.
• 34:15
  But you're saying, okay, well, I've got a few tricks up my sleeve.
• 34:19
  Yes.
• 34:19
  Right?
• 34:19
  Yes.
• 34:20
  I think it's rather cool.
• 34:22
  I think that this is an important, you know, there is always, because of this nexus between nation states trying to accrue and maintain power and the importance of frontier technologies to that goal, we're always going to see policy
• 34:41
  impact how these technologies grow and progress.
• 34:46
  The Biden administration made a bet that hardware strangulation to China was the best path at creating advantage for the U.S. in the race.
• 35:01
  And it was a well-informed bet, right?
• 35:02
  Because like everyone who had been doing
• 35:06
  AI thought that that was the masterpiece, right?
• 35:09
  The hardware was it, right?
• 35:11
  Ever since 2012 when AlexNet came out.
• 35:14
  And what they could do is actually use GPUs to train in parallel.
• 35:18
  And then just all of the mathematics of matrix multiplication and backpropagation could be done on GPUs.
• 35:25
  That's when it was, you know, ever since then we've sort of taken for granted that, oh, GPUs are really...
• 35:30
  What matter?
• 35:31
  But we've created good enough GPUs.
• 35:34
  Yes.
• 35:34
  That we've taken for granted that like the best GPU is always the best.
• 35:38
  But actually the H800 is like a good enough GPU now.
• 35:41
  Yes.
• 35:42
  That it can do enough of the math.
• 35:46
  Yes.
• 35:47
  That now these guys can actually go and...
• 35:51
  You know, start tweaking the knobs.
• 35:54
  Yeah, start tweaking the knobs.
• 35:55
  It's a really, you know, the cost difference is crazy.
• 36:00
  The equivalence of efficacy of the models, given the cost difference, is even crazier.
• 36:06
  And it is going to continue to be...
• 36:08
  This hardware-software tension, where do you innovate more?
• 36:14
  Where has the most aggregate value in terms of continuing to increase the power and efficacy of these models?
• 36:22
  It's going to continue to be a rolling story.
• 36:25
  I think, again, to your point at the top of when we started this story, it is very interesting.
• 36:31
  That for the amount of time, I remember using GPT-2 in my company trying to see, you know, trying to gauge when is this stuff going to be ready for production years ago.
• 36:42
  And it is, the jury is not out yet on...
• 36:51
  on this journey to AGI and artificial super intelligence, which of these two levers that we're pulling are going to create those incremental?
• 37:00
  Because we've sort of seen, the models are kind of starting to aggregate at a similar level of performance.
• 37:07
  Yeah.
• 37:08
  And the gains from your walled garden are becoming less and less.
• 37:14
  And this is a classic battle in every technology revolution that's ever existed
• 37:20
  in terms of the tension between closed source
• 37:23
  and open source.
• 37:25
  Open source has always been, all servers are on Linux.
• 37:28
  Linux is open source.
• 37:30
  Why is that?
• 37:32
  Everyone uses Python.
• 37:33
  Everyone uses Python, open source.
• 37:34
  And the reason, the theory of the case is when you open source, there's a more of a distributed incentive to create security.
• 37:40
  Exactly.
• 37:41
  Because everyone wants it to work.
• 37:43
  You're having more viewpoints inputting into the system itself.
• 37:47
  And you're limited by your resources if you go closed source by your talent.
• 37:51
  The way in which in the U.S. we deal with that is Facebook gives billion-dollar offers to the best AI researchers.
• 37:58
  But at some point, this is a forever tension in technology revolutions.
• 38:02
  And it's interesting to see it play out in this geopolitical sphere where the –
• 38:09
  the winner or the import of the battle is so high.
• 38:15
  Yeah, yeah, the stakes are really high here.
• 38:17
  It's really high.
• 38:18
  Yeah, yeah, yeah.
• 38:19
  And so great first story.
• 38:21
  I love the AI story is always great because a lot of times people just talk about the end product.
• 38:26
  We haven't had a lot of research papers to talk about.
• 38:28
  No, but it's kind of cool that we got a sort of look under the box to see how it all worked.
• 38:35
  And they were also forthcoming about some of the challenges that they had.
• 38:41
  When they first started out with DeepSeek R1-0, which is the one that they didn't release, they were having this trouble where the output would be half Chinese, half English.
• 38:51
  Right?
• 38:51
  So then what they did was they had cold start data.
• 38:54
  So it wasn't completely all blind reinforcement learning.
• 38:57
  What they did was they primed the model with human data, okay, to actually sort of give it a running start.
• 39:05
  Yes.
• 39:06
  And not have it, like, from the very beginning start doing random stuff.
• 39:10
  So what you do is you can have a small subset of human-labeled training data.
• 39:14
  Yes.
• 39:14
  That, like, creates a sort of...
• 39:17
  neighborhood of parameters okay where where okay like the the model sort of settles in on this neighborhood of where the numbers should be for all of these weights and then you can fine tune it and fine tuning is a coarse-grained word because obviously like a lot of training was done afterwards but like
• 39:35
  After that initial sort of like, okay, don't do like the stupid stuff of like half Chinese, half English.
• 39:41
  But then now get into the, you know, how do you problem solve?
• 39:45
  How do you actually make a really nice LLM?
• 39:49
  So they were forthcoming about some of their challenges, which I think is kind of interesting.
• 39:53
  And again, it goes back to the sort of the narrative violation that the current CCP has put ourselves in where...
• 40:02
  in the US we're still viewing their operating system as what it was 20 years ago, but they are being nimble and understanding the changing environment and trying to reposition.
• 40:11
  Yeah.
• 40:12
  Again, you can argue that it's contrived or not.
• 40:16
  Yeah.
• 40:16
  I mean.
• 40:18
  Knowing China, of course, there's stuff that we don't know about, right?
• 40:22
  Of course, probably DeepSeek is entrenched with the CCP.
• 40:26
  I'm just saying.
• 40:28
  I'm just saying.
• 40:29
  It's probably true.
• 40:30
  But they're doing the work.
• 40:32
  They're publishing peer-reviewed papers.
• 40:34
  They're coming out with cutting-edge AI, and they're doing it without the H-100s, or at least not with the kind of access...
• 40:43
  that companies have in America.
• 40:45
  Perhaps they do have H-100s from shell companies in Bhutan or whatever.
• 40:50
  Which is one of the arguments is that they've been doing this illicit market trading in order to get access.
• 40:55
  But the point is they don't even need it.
• 40:57
  Yeah, yeah.
• 40:57
  It turns out like...
• 40:59
  You know, people are, now that this is out, they're going to try using these strategies now.
• 41:03
  And we're going to see if the rubber meets the road.
• 41:07
  And, you know, since DeepSeek came out, Meta was already releasing open weight models.
• 41:14
  I think they've actually incorporated some of the mixture of experts architecture into the llama models.
• 41:19
  I think people are taking this sort of reinforcement learning strategy very seriously, too, where they're saying, OK, maybe I don't need a bunch of Kenyans to just keep annotating my outputs.
• 41:29
  Right.
• 41:29
  Correct.
• 41:30
  Yeah.
• 41:30
  And we now just in the last month and a half have an open source, open AI model.
• 41:36
  And it is very easy to argue that we would not have an open source.
• 41:43
  Weights open AI model without Deep Seagull.
• 41:47
  Yeah.
• 41:48
  Yeah.
• 41:48
  There was not an incentive for.
• 41:49
  No, why would they do it?
• 41:50
  Open AI to do so.
• 41:50
  Yeah.
• 41:51
  But they had to answer investors and the competition, et cetera.
• 41:55
  Great, great update here.
• 41:56
  We'll definitely be touching on AI more.
• 41:59
  We're going to go to our next story, which I'm very excited about because I don't know what any of it means.
• 42:02
  Yeah.
• 42:03
  The headline on our second story, scientists just made the first time crystal that you can see.
• 42:11
  We have an ocarina of time.
• 42:12
  Yes.
• 42:13
  Physicists at the University of Colorado Boulder have created the first time crystal that humans can actually see using liquid crystals that swirl into a never ending pattern when illuminated by light.
• 42:24
  They put out their paper in nature materials.
• 42:27
  There's been a lot of coverage of this one.
• 42:29
  I always see all these TikToks about time crystals and all this.
• 42:33
  Yeah, it's very Marvel, very Doctor Who.
• 42:36
  Yeah, yeah, yeah.
• 42:36
  You know, like, this is how we beat Thanos.
• 42:39
  Mr.
• 42:39
  Strange, Doctor Strange.
• 42:41
  Yeah, yeah, all sorts of, yeah.
• 42:43
  So help me understand, you know, what exactly is going on with this time crystals paper?
• 42:46
  Yeah, so, well, first, let's talk about what are time crystals.
• 42:51
  I would like to know that.
• 42:52
  Okay.
• 42:52
  So before we even do that, normal crystals.
• 42:56
  Okay.
• 42:57
  We talked about diamond the other day.
• 42:59
  Yes.
• 43:00
  The simplest crystal I can think of is table salt.
• 43:03
  Okay.
• 43:03
  N-A-C-L, sodium chloride.
• 43:05
  You can imagine there's a sodium atom and a chlorine atom.
• 43:08
  Sodium atom, chlorine atom.
• 43:09
  Sodium atom, chlorine atom.
• 43:11
  In this regular sort of chessboard-like structure, but a 3D sort of chessboard.
• 43:16
  Mm-hmm.
• 43:17
  of sodium chlorine, sodium chlorine.
• 43:19
  The reason why this is a crystal, right?
• 43:21
  What characterizes it as a crystal is you've got a discrete space symmetry breaking, okay?
• 43:28
  This is a technical term in physics that just means if I take the crystal and I move it over by one...
• 43:36
  spacing, I will get the same thing back.
• 43:39
  Okay.
• 43:39
  Okay?
• 43:40
  You can imagine I've got the NaCl and then there's another NaCl here, right?
• 43:45
  I can take that entire crystal and move it by a whole molecule and I'll have the same thing repeating, right?
• 43:52
  Just like in a chessboard, right?
• 43:54
  Imagine an infinite chessboard
• 43:56
  that thing is a crystal in some sense, right?
• 43:58
  It's a 2D crystal because there's black, white, black, white.
• 44:01
  If I take up the chessboard and I move it over two sort of squares, I'll get the same exact chessboard, right?
• 44:09
  So that's what's called symmetry breaking in space, okay?
• 44:13
  Where it's no longer continuous because you can't just arbitrarily move this thing around.
• 44:18
  You have to move it by the lattice spacing, right?
• 44:20
  The lattice constant of the crystal.
• 44:22
  Yes.
• 44:23
  So that's what makes it a crystal.
• 44:25
  Okay.
• 44:25
  And it's also resilient to forces and stress, right?
• 44:29
  Like a diamond, for example, great way.
• 44:32
  I put stress on it.
• 44:33
  The crystalline structure of the carbon atoms isn't really going to change because the way that these carbon atoms are bonded to each other, they're resilient to like push and pull.
• 44:43
  You know, I heat it up.
• 44:45
  That...
• 44:46
  that structure is going to be resilient.
• 44:48
  Right?
• 44:48
  So there was a guy, Frank Wilczek, he was a Nobel laureate.
• 44:52
  He won the Nobel Prize in 2004 for discovering something called asymptotic freedom, which has to do with the ways that quarks behave inside the atomic nucleus.
• 45:03
  We can get into that some other day.
• 45:05
  I think we need some asymptotic freedom in the US right now, but that's a whole nother story That's a whole nother story as well for another day But he won the Nobel Prize in 2004 and he's been very active in physics ever since he's discovered things like called axions and like he's done stuff with cosmology And one of the things that he proposed back in 2012 was this idea of a time crystal Okay
• 45:28
  So he figured it's a very interesting, quite simple argument.
• 45:32
  Okay.
• 45:33
  He said that the laws of physics are symmetric to space translations.
• 45:39
  Right.
• 45:40
  Meaning if I do a, if I do an experiment here and then I, you know, take up the experiment and I move it somewhere else, the results of the experiment should be the same.
• 45:49
  Right.
• 45:49
  If, if all the environments are the same and that's because there's no special coordinate
• 45:54
  in the universe right there's no like like there's no like oh it's just like here is where where things happen right there's no um there's no special spatial coordinate well similarly the laws of the universe are also symmetric in time right if i do an experiment now and then i do it the next day should be the same yeah so if we have things that break spatial symmetry
• 46:18
  By creating crystals in space, right, where you have clearly like carbon making the diamond or NACL making table salt.
• 46:27
  So you clearly have spatial crystals.
• 46:30
  What if I have a breaking of the time symmetry to create a time crystal?
• 46:35
  And what would that look like?
• 46:36
  What that would look like is you've got some kind of periodic system that returns to its space.
• 46:42
  original form periodically.
• 46:45
  So just like how with a space crystal, like the chessboard, I can take it up and I can move it in space and then put it back and it'll be the same thing in time.
• 46:54
  For example, a sine wave, right?
• 46:56
  Mathematically, a sine wave is symmetric in time to that period, right?
• 47:02
  I can pick up the sine wave, I can move it one entire period and put it back and it'll be the same sine wave.
• 47:08
  I should be able to have that
• 47:11
  in like the physical world.
• 47:13
  Okay.
• 47:14
  Okay.
• 47:14
  Because, you know, it's kind of same, same.
• 47:17
  Yeah.
• 47:17
  Right.
• 47:17
  Theoretically.
• 47:18
  Theoretically, it should be possible.
• 47:20
  And it kind of makes sense.
• 47:22
  If I have the same sort of symmetry breaking in space, I should have the same sort of symmetry breaking in time.
• 47:27
  Well, it wasn't quite that simple.
• 47:28
  Okay.
• 47:29
  So right after he published that, it was, you know, Frank Wilczek is doing this Nobel Prize winner.
• 47:35
  He's a very famous physicist.
• 47:36
  So everyone's like, okay, is this possible?
• 47:38
  And actually it turns out the way that he had envisioned it, it is not.
• 47:43
  So people came out and they published something called the no-go theorem for time crystals, which was that.
• 47:50
  Frank Wilczek's original idea was that this time crystal is going to be in the ground state, which means it's in its lowest energy state.
• 47:58
  There's nothing pumping on it.
• 47:59
  It's just, you know, crystals can be in their lower energy state, right?
• 48:02
  I can cool a crystal down to like near zero Kelvin and it's going to remain in its crystal form.
• 48:08
  And that's what really like that means that it's an inherent property of the material itself, right?
• 48:14
  And it's not something that coming from outside creating this order.
• 48:17
  Mm-hmm.
• 48:18
  Frank Wilczek's idea was that you can have a system that is in its ground state, right?
• 48:24
  In its lowest energy state.
• 48:25
  There's nothing coming in.
• 48:26
  And it's still going to behave in this sort of periodic manner.
• 48:29
  Yes.
• 48:30
  Okay?
• 48:30
  That turns out can't be the case.
• 48:32
  Okay.
• 48:32
  Okay?
• 48:33
  So the no-go theorem said that a true equilibrium ground state cannot have persistent periodic motion.
• 48:39
  Meaning, like, if you've cooled something down to its ground state, it's going to be stationary.
• 48:44
  Right.
• 48:44
  And that kind of makes sense, right?
• 48:46
  Because if it's moving around in its ground state, that means there's some kinetic energy that I can extract.
• 48:51
  And it's not really the true ground state, right?
• 48:54
  Okay, so that's fine.
• 48:56
  But then people started wondering, all right, fine.
• 48:58
  I don't need something to be in equilibrium at its ground state.
• 49:01
  What if I pump it with stuff?
• 49:04
  Can I create a crystal that way?
• 49:05
  What if I pump it with energy?
• 49:07
  So it's now a non-equilibrium state.
• 49:09
  Like there's energy coming in and then there's energy going out.
• 49:13
  But the way that this energy is transformed inside the system creates this periodic motion within the system.
• 49:19
  Right?
• 49:19
  Right.
• 49:20
  Turns out that was possible.
• 49:22
  And for the longest time, people were studying these things called discrete time crystals.
• 49:28
  They were systems that were kept out of equilibrium by this outside force.
• 49:32
  And it's not as simple as just like you just pump it rhythmically and then this thing goes by that.
• 49:37
  Then it's like, oh, like me on a swing, like doing my legs is a time crystal.
• 49:43
  Like me at the playground is a time.
• 49:45
  No.
• 49:45
  No, that's not it's not quite as simple.
• 49:47
  What you need is you need you need the pumping to be done at a certain frequency.
• 49:51
  And then the crystal itself is doing it at a different frequency.
• 49:55
  Maybe it's at a harmonic.
• 49:56
  Right.
• 49:57
  And there were specialized quantum systems that could do this.
• 50:01
  Right.
• 50:02
  There were nuclear spins, trapped ions, cold, cold atoms.
• 50:06
  at extremely low temperatures, quantum weirdness.
• 50:10
  You know, it's like cool, but it's cool.
• 50:13
  But that's about it.
• 50:15
  Like maybe in 100 years, it can lead to like, you know, quantum devices and things like that.
• 50:21
  This paper is very cool because it's doing it in a classical system, which means no quantum weirdness.
• 50:27
  And it's doing it at room temperature.
• 50:29
  Wow.
• 50:30
  And it's doing it at a big enough scale that we can see with our naked eye.
• 50:34
  This is giving me LK99 vibes.
• 50:37
  Yeah, but this is out of UC Boulder.
• 50:39
  Yeah, so it's legit.
• 50:40
  And so it's pretty legit.
• 50:41
  UC Boulder has probably the best AMO department out of anywhere in the world.
• 50:46
  Okay.
• 50:46
  Okay?
• 50:47
  Atomic Molecular Optical.
• 50:50
  Those guys do it.
• 50:51
  Let's invest in our university research institutions.
• 50:54
  Yeah.
• 50:54
  I mean, you know, Boulder has like the most PhDs out of any city.
• 50:58
  I heard this recently.
• 50:59
  Yeah, and a lot of it has to do with UC Boulder.
• 51:01
  Out of UC Boulder came NIST, which is the National Institute for Standards and Technology.
• 51:06
  There's a bunch of other science labs just in Boulder.
• 51:08
  UC Boulder itself has like four Nobel Prizes in physics just for AMO.
• 51:13
  Okay, like not like, okay, there's one for cosmology, one for...
• 51:16
  It's just all like in the lab making like materials, right?
• 51:20
  They made the first Bose-Einstein condensate.
• 51:22
  They're the ones who figured out how to trap ions using lasers and cool them down using lasers...
• 51:28
  So they're very good.
• 51:29
  So when a paper comes out of UC Boulder and it's an AMO, it's like, okay, these guys probably talked to their department.
• 51:35
  And if it didn't pass, like, their own.
• 51:38
  Yeah, they have the best in the world there.
• 51:39
  Yeah, yeah, yeah.
• 51:40
  It's like, so it's a really cool little thing that they've made out of liquid crystals, right?
• 51:47
  These are the LCDs.
• 51:50
  And they made a macroscopic time crystal.
• 51:53
  Okay.
• 51:55
  And what you can do is, what's really cool is, first of all, you don't need a laser or oscillating magnetic field like in those old time crystals.
• 52:04
  In the old ones, you needed a quantum system and you needed a laser, which means coherent light, single frequency.
• 52:11
  Or like a magnetic field, it's oscillating at a single frequency.
• 52:13
  And then that sort of...
• 52:16
  Drives this quantum system to do this periodic motion.
• 52:20
  This is just like a light.
• 52:21
  You just shine a light on it, and then it uses the energy from that light to create this periodic motion.
• 52:26
  Okay?
• 52:27
  And it just uses LCD, like liquid crystal technology, pneumatic liquid crystals.
• 52:33
  So the way that liquid crystals work, like, you know, you've used liquid crystals before.
• 52:39
  It's a really simple technology.
• 52:40
  It actually won the Nobel Prize.
• 52:42
  The idea is...
• 52:44
  So what does it mean by a liquid crystal?
• 52:46
  Okay.
• 52:47
  A crystal you usually think of as a solid.
• 52:49
  Yes.
• 52:50
  Right?
• 52:51
  A solid meaning that you've got atoms in a regularly spaced pattern and they don't really move.
• 52:56
  In the liquid crystal, you've got molecules that are all sort of... You can think of them as like rods, many tiny rods.
• 53:04
  They're molecules that are elongated.
• 53:07
  And you can manipulate their direction...
• 53:11
  using an electric field.
• 53:12
  Okay.
• 53:13
  Okay?
• 53:14
  So if you put an electric field in this direction, all of the molecules align in that direction.
• 53:19
  And you can have an electric field in this direction, which means all the molecules align in this direction.
• 53:23
  And what these molecules do is because they're small enough, they can actually influence the propagation of light.
• 53:29
  Okay?
• 53:30
  And so that's exactly how liquid crystal displays work.
• 53:33
  What you do is you have a backlight.
• 53:35
  Okay?
• 53:35
  OK, that like lets and then you have a polarizer that polarizes all the light.
• 53:39
  So all of the electric fields are propagating in one direction.
• 53:43
  And what you can do is you can have a liquid crystal interface in between another polarizer.
• 53:49
  So now if the light gets to the other polarizer and it's polarized in the same direction, it's going to be let through.
• 53:55
  Yeah.
• 53:55
  But if the light is polarized in a different direction, in the perpendicular direction, then it's going to be blocked.
• 54:00
  And what you can do is use an electric field to manipulate the liquid crystals in the middle to influence the propagation of light.
• 54:09
  That's the fundamental technology.
• 54:12
  Okay.
• 54:12
  So you can have an electric field that sort of rotates the light.
• 54:16
  And then if it rotates in exactly the right way, then the light is going to go through and I'm going to see something.
• 54:20
  But if there's no electric field or if I make the electric field this way, then all the rods are going to be, you know, sort of in the perpendicular direction.
• 54:29
  The light is not going to rotate and I'm going to get blocked.
• 54:31
  Yes.
• 54:31
  And that's fundamentally how liquid crystals work.
• 54:34
  They're liquid in the sense that, you know, just how in a liquid the molecules can move around.
• 54:39
  In this case, these liquid molecules, these rod-like molecules can move around and they can be controlled by an electric field.
• 54:46
  And they can be controlled by light itself, right?
• 54:50
  So they figured what I can do is I can create a liquid crystal.
• 54:56
  And this liquid crystal has exactly the right properties where it can...
• 55:02
  sort of act like a giant sort of, I guess, okay.
• 55:09
  So this liquid crystal can be oriented in such a way that when it's shined on by light,
• 55:18
  What you're going to get is deformations in the crystal.
• 55:22
  Yes.
• 55:22
  Okay?
• 55:22
  Tiny little deformations.
• 55:24
  So what you can imagine is, let's say all of the liquid crystals, all of these rods are oriented in one direction.
• 55:31
  Okay?
• 55:32
  And now what I do is I introduce a tiny kink.
• 55:36
  in this lattice.
• 55:37
  So they're all oriented in one direction, but I've taken like a tiny little neighborhood and I've like bent them in a certain way.
• 55:44
  Okay.
• 55:44
  What you can do is create a material such that that bend, which becomes this sort of topological defect.
• 55:50
  Yes.
• 55:50
  This like idea of like, okay, this neighborhood has a tiny little bend or a twist.
• 55:54
  That twist can then influence the stuff around it, which can influence the stuff around it, which can influence the stuff around it.
• 56:00
  And what you get is this sort of traveling wave of disturbance.
• 56:04
  Yes.
• 56:04
  Okay.
• 56:05
  There are what are called topological solitons in physics.
• 56:09
  This idea that like I've got this like disturbance and that disturbance kind of acts like a particle.
• 56:16
  Okay.
• 56:16
  Because it travels the same way as a particle.
• 56:21
  The orientation of that disturbance can have a charge in the sense that if I...
• 56:26
  twist it clockwise, then it's a plus one.
• 56:29
  And if I twist it counterclockwise, it's a minus one.
• 56:32
  And that charge is conserved in the same way that sort of charge is conserved if I were to have electrons moving around, right?
• 56:39
  And they can cancel each other out, kind of the same way that charges cancel each other out, right?
• 56:44
  So they created this thing and they actually did the numerical calculation.
• 56:47
  So they wrote down an equation for how these solitons would interact with each other.
• 56:53
  And all of these deformations act like little quasi particles that interact with each other and have a physics to themselves.
• 57:01
  Yes.
• 57:01
  That is an emergent property of the little tiny rods themselves.
• 57:06
  That initial disturbance that has now created a cascade.
• 57:09
  Yeah.
• 57:09
  That is now independent of the initial source necessarily being.
• 57:13
  Yeah.
• 57:15
  And it's like it's on top of the rods themselves.
• 57:17
  The rods actually aren't the particles anymore.
• 57:20
  Yeah.
• 57:20
  It's like the way in which the rods are oriented becomes a particle mathematically.
• 57:28
  Right.
• 57:29
  You can describe the twist itself as this thing that propagates rather than the rods.
• 57:35
  Because the rods are just staying where they are.
• 57:37
  Right.
• 57:38
  Right.
• 57:38
  That's the crystal part of it.
• 57:41
  But the way that the rods move, that disturbance moves around like a particle.
• 57:46
  This is really...
• 57:47
  Which is kind of cool, right?
• 57:48
  No, it's very... That's, like, very interesting.
• 57:50
  Yeah, dude.
• 57:51
  Solitons have been around, like, I think in the 1800s, there's this British guy.
• 57:54
  Forget his name.
• 57:56
  Maybe we'll put it up.
• 57:57
  Yes.
• 57:57
  But it's a funny story because what he saw was he was, you know, these British physicists back in the day.
• 58:03
  Very rich, right?
• 58:04
  He's on horseback in the countryside, right?
• 58:07
  And he's, like, on horseback, and he sees a canal, and then there's a boat.
• 58:11
  And the boat, like, breaks suddenly.
• 58:13
  Okay.
• 58:14
  And when the boat breaks suddenly, it created this wave.
• 58:18
  And he noticed that the wave just kept propagating.
• 58:20
  And then he was on horseback and he's got nothing to do.
• 58:22
  He's probably, you know, got an estate.
• 58:24
  Right.
• 58:25
  That like gets him money.
• 58:26
  Yes.
• 58:26
  Like in Downton Abbey or whatever.
• 58:28
  So he's like, fuck it.
• 58:29
  I'm going to follow this wave.
• 58:31
  And he followed it for two miles.
• 58:33
  Hmm.
• 58:33
  There was a wave that kept reinforcing itself and it acted kind of like a particle, right?
• 58:38
  In the sense that the wave didn't really dissipate until two miles later.
• 58:42
  And he could follow this thing and he discovered the first soliton.
• 58:46
  There's a little plaque where he lived where it's like this is the guy who discovered the first soliton.
• 58:52
  And ever since then, they've been observed in all sorts of sort of material sciences.
• 58:57
  We've observed them in biology and quantum systems.
• 59:00
  And this is exactly one of those things where it's like there's a wave that's being propagated.
• 59:05
  And that wave sort of reinforces itself because of the way that the medium itself works.
• 59:11
  Works.
• 59:11
  Right.
• 59:11
  You know?
• 59:12
  Right.
• 59:12
  It is a combination of that input and the structure of the medium.
• 59:15
  Yes.
• 59:16
  Exactly.
• 59:16
  The liquid crystal.
• 59:18
  And the structure of the liquid crystal enables it to continue the propagation of the wave independent, like as an emergent property of the structure's initial condition.
• 59:31
  Exactly.
• 59:31
  Yeah, yeah.
• 59:32
  And so they figured that they could use this to make a classical...
• 59:37
  time crystal right because this is not quantum this is this is just the the rods themselves are like you know multiple many many atoms big right so these guys are interacting through just classical effects yes right and so what you can do is you can have a liquid crystal embedded inside glass right that have these two layers and as one layer sort of moves it influences the light on the other layer and then that thing moves and there's this feedback cycle
• 1:00:03
  And that feedback cycle is what creates this propagation of the time crystal.
• 1:00:08
  And it's really a space-time crystal because what you're seeing, you can actually look at the videos and you can see the colors move.
• 1:00:20
  In real time.
• 1:00:22
  And the wavelength is like, you know, about half a millimeter to a millimeter, which means you can see it with your naked eye.
• 1:00:27
  And it's moving through this liquid crystal in this sort of propagating, you know, traveling wave.
• 1:00:34
  Right?
• 1:00:35
  And...
• 1:00:36
  It was huge because one of the big things that defines a crystal, right, is it has to be, as I said before, crystals have to be resilient.
• 1:00:45
  Yes.
• 1:00:46
  To stress.
• 1:00:46
  Right.
• 1:00:47
  Okay.
• 1:00:47
  If I take a diamond and I press it or I heat it up, the diamond is going to retain its temperature.
• 1:00:52
  So how did they actually test that here?
• 1:00:53
  Well, what you can do is you can like poke it.
• 1:00:56
  Yeah.
• 1:00:57
  And the pattern stays, which means that it is a true time crystal.
• 1:01:02
  Right.
• 1:01:02
  It is resilient to the poking.
• 1:01:07
  It already has.
• 1:01:08
  Yeah, it's got some time crystal going through.
• 1:01:10
  And then when I poke it, the time crystal sort of.
• 1:01:12
  Like it recovers where I poke it.
• 1:01:14
  Of course, it's going to like mess up, but I remove it.
• 1:01:17
  And then the crystal regains its properties.
• 1:01:20
  The other thing they did was they would vary the brightness of the light.
• 1:01:24
  Because it can't be that like, you know, it depends really heavily on the brightness of the light.
• 1:01:28
  So they randomly varied the brightness of the light.
• 1:01:30
  The crystal kept going.
• 1:01:32
  Which means that what this is is really an intrinsic property of the material itself rather than the drive that's coming from the outside.
• 1:01:40
  Yes.
• 1:01:40
  Right?
• 1:01:40
  No, that's a very – that's exactly – And then the final nail in the coffin for me was what they did was they would turn off the light and then they would turn it back on again.
• 1:01:52
  And if this is not a true crystal, if this is simply a response to the light, then when I turned it off, I mean, sorry, when I turned it back on, the crystal would start up at exactly the same phase every time.
• 1:02:07
  Meaning I turn it on five minutes later, the same pattern.
• 1:02:10
  I mean, sorry, five seconds later, the same pattern emergence at exactly the same spot.
• 1:02:14
  But what would happen was they would turn it off and then they turned the light back on and it would randomly order itself.
• 1:02:22
  The same crystal would start, but at different starting points in that wave.
• 1:02:27
  You see what I'm saying?
• 1:02:28
  Like the phase offset was completely random.
• 1:02:30
  Yes.
• 1:02:30
  Which again tells you this is not a property of the outside drive.
• 1:02:35
  Right.
• 1:02:35
  It's a property of the material itself.
• 1:02:38
  Right.
• 1:02:38
  So this is truly a time crystal that is macroscopic and it's something we can see.
• 1:02:43
  This is this is I mean, this is right.
• 1:02:45
  This is this is really fast.
• 1:02:47
  It's really cool.
• 1:02:48
  And and like, OK, you might be thinking, OK, this is like this is cool.
• 1:02:52
  Right.
• 1:02:53
  Then I got into reading about the applications.
• 1:02:56
  OK. OK.
• 1:02:57
  The applications for this are like.
• 1:03:01
  like science fiction to me.
• 1:03:03
  Okay.
• 1:03:04
  Most of them have to do with like cryptography and security.
• 1:03:08
  Okay.
• 1:03:09
  First one I'll tell you about.
• 1:03:11
  Imagine we make this thing cheap.
• 1:03:16
  We can stick it on a dollar bill.
• 1:03:19
  Oh, right.
• 1:03:19
  As a cryptographic signature.
• 1:03:21
  As a cryptographic signature.
• 1:03:23
  Now, if you want to test whether your Benjamin is a real Benjamin, you take your iPhone flashlight, stick it on there.
• 1:03:29
  It's going to have this pattern, right?
• 1:03:31
  Because all it needs is an external drive.
• 1:03:33
  And the pattern is we can make it really hard to duplicate.
• 1:03:37
  Yes.
• 1:03:37
  Right?
• 1:03:37
  So, you know that strip that's on the Benjamin?
• 1:03:40
  Yes.
• 1:03:40
  We can have that strip be a time crystal, like literally like…
• 1:03:44
  propagates.
• 1:03:45
  Yes.
• 1:03:45
  Right?
• 1:03:46
  And then you'll know, okay, that's a real one.
• 1:03:47
  It has a signature that only... And it's going to be so hard to duplicate.
• 1:03:50
  Right.
• 1:03:51
  Right.
• 1:03:51
  Right?
• 1:03:52
  Right.
• 1:03:53
  That's so cool.
• 1:03:54
  No, that's very... That's actually really... Yeah.
• 1:03:56
  And the other one is like 2D... Like, so, you know, usually we have 2D barcodes, like a QR code is a 2D.
• 1:04:03
  Now you can have a 2D plus one barcode.
• 1:04:06
  You can have a QR code that is 2D.
• 1:04:11
  And then when you shine a light on it,
• 1:04:13
  the QR code itself changes into different QR codes.
• 1:04:18
  You can have multiple QR codes on top that have different periodicities.
• 1:04:23
  So that like, let's say one is like two seconds and the other one is three seconds and the other one is five seconds.
• 1:04:30
  Well, then the next time that it's going to become a full QR code is gonna be five times, three times two, which is 30, right?
• 1:04:37
  So you're gonna have to wait 30 seconds to get the actual QR code.
• 1:04:41
  You can imagine you've introduced a whole new dimension to security.
• 1:04:49
  A manipulatable dimension.
• 1:04:51
  Yeah.
• 1:04:51
  And the whole point is this is at room temperature.
• 1:04:55
  Right.
• 1:04:55
  Right.
• 1:04:56
  And it's macroscopic.
• 1:04:58
  It's something I can see.
• 1:05:00
  I don't need a telescope or like weird ions and lasers to actually, oh, look, it's doing the thing.
• 1:05:06
  This is why I made the LK99 reference.
• 1:05:09
  Not because of whether it was real or not real, but because...
• 1:05:12
  It's room temperature and macroscopic.
• 1:05:14
  Yeah.
• 1:05:14
  And the reason why that's material is that means the transition from theoretical to production application.
• 1:05:22
  Yeah.
• 1:05:22
  Now we've just got to make it cheap, which is honestly kind of like the easier part of the innovation.
• 1:05:27
  Yes.
• 1:05:27
  Right.
• 1:05:27
  Exactly.
• 1:05:28
  You know.
• 1:05:28
  It's an engineering problem.
• 1:05:29
  Yeah.
• 1:05:29
  Now it's like, okay, because we can make LCDs real cheap now.
• 1:05:32
  Right.
• 1:05:33
  Right.
• 1:05:33
  Right.
• 1:05:34
  This is.
• 1:05:34
  And like this, you know, it doesn't need a power source.
• 1:05:38
  You just need to put it on there.
• 1:05:39
  The power source is the light that I'm going to shine on it.
• 1:05:42
  It's kind of cool.
• 1:05:43
  I'm seeing from the abstract here, their potential technological utility includes optical devices, photonics, space-time crystal generators, telecommunications, anti-counterfeiting design.
• 1:05:58
  And again, this is just a first –
• 1:06:00
  Yeah, I mean, I only talked about the anti-counterfeiting and the security part.
• 1:06:05
  But yeah, as you said, devices, all sorts of stuff.
• 1:06:08
  There's all sorts of implications in terms of what we can get out of now this fundamental...
• 1:06:15
  like implementation experimentally at room temperature in a macroscopic form, which is, yeah, I mean, there's a reason why it was picked up by so many news outlets.
• 1:06:23
  Yeah, it was covered by a lot.
• 1:06:25
  This is quite a big deal, yeah.
• 1:06:26
  This is quite good.
• 1:06:27
  They didn't get the, they got nature materials.
• 1:06:29
  They could have put it in nature, but they said, we'll put it in nature materials.
• 1:06:33
  They probably tried in nature first.
• 1:06:35
  Usually everyone does.
• 1:06:37
  First time crystals that you can see.
• 1:06:40
  Really interesting story.
• 1:06:42
  And the background is also helpful to understand.
• 1:06:44
  We live in L.A., so we know a lot of people that like to talk about crystals.
• 1:06:48
  This is a little bit different than those crystals.
• 1:06:49
  Yeah, this one actually works.
• 1:06:52
  Speaking of California, for our third story...
• 1:06:56
  We're now going to move into the immunology space.
• 1:06:58
  Headline, hidden viruses in our DNA could be medicine's next big breakthrough.
• 1:07:04
  Scientists at La Jolla Institute for Immunology, when they're not out surfing on the beautiful sand beaches of San Diego, have decoded the 3D structure of an ancient viral protein hidden...
• 1:07:17
  Within our DNA our own DNA our own DNA this protein found on cancer and autoimmune cells has a unique shape that could unlock new Diagnostics and therapies this was published in science advances Again, this is another California University.
• 1:07:35
  We have the most papers the most patents and
• 1:07:38
  We got it.
• 1:07:40
  I just always have to make sure people know that we are winning.
• 1:07:42
  We are winning.
• 1:07:43
  We are winning.
• 1:07:43
  This is La Jolla, California, baby.
• 1:07:46
  It's right next to UC San Diego.
• 1:07:48
  UC San Diego is an amazing, amazing crucible for biotechnology and molecular biology research.
• 1:07:55
  It's kind of like the Silicon Valley for biology in some sense.
• 1:08:00
  And this is a really groundbreaking achievement.
• 1:08:02
  It's the first time that we have ever published the structure of a human endogenous retrovirus.
• 1:08:10
  Okay.
• 1:08:10
  And this is a virus that hides within our own DNA, UNRs.
• 1:08:14
  I didn't even know that was possible.
• 1:08:15
  Yeah.
• 1:08:16
  This is crazy, dude.
• 1:08:18
  8% of our DNA is viral endogenous DNA.
• 1:08:29
  infiltrating that have infiltrated dude it's it's they're like they're like sleeper cells dude i was this is when people say the aliens uh manufactured uh created no look this is exactly what i'm talking about little viruses are they're writing our stuff it's crazy it's crazy to think about so so first of all i was like i was like how does this even get in why is this in my dna bro right like
• 1:08:52
  What's going on?
• 1:08:53
  Okay, so it's crazy.
• 1:08:55
  These things are called retroviruses.
• 1:08:57
  Retrovirus is a virus that has, instead of using DNA as its genetic material, it uses RNA as its genetic material.
• 1:09:05
  And what it does is when it infects a cell, it uses something called reverse transcriptase.
• 1:09:11
  Usually transcriptase goes from DNA to RNA.
• 1:09:14
  This is the reverse process.
• 1:09:15
  So it's taking RNA and it's making double-stranded DNA.
• 1:09:18
  So that's reverse transcriptase.
• 1:09:20
  And then what ends up happening is that reverse transcriptase goes inside the nucleus of our cells.
• 1:09:26
  Yes.
• 1:09:27
  And then there's something called integrase, which is another viral protein that takes that DNA, that bit of viral DNA, and sticks it inside our chromosomal DNA.
• 1:09:36
  Now, if this happens to our skin cells, okay, we got like some viral DNA.
• 1:09:39
  But if it happens to our sperm and egg, now the next organism that gets that sperm and egg, all of the DNA is going to have it, right?
• 1:09:50
  These are called germline mutations.
• 1:09:51
  I was going to say, going back to our CRISPR story, it's the germline.
• 1:09:55
  It's the germline mutations, right?
• 1:09:56
  Yes.
• 1:09:56
  So this is a germline insertion that happens.
• 1:09:59
  And then now the virus becomes something called a provirus.
• 1:10:02
  which is something that's just hanging out in our DNA.
• 1:10:05
  Yes.
• 1:10:05
  8% of our DNA comes from this.
• 1:10:08
  Comes from this.
• 1:10:09
  Dude, that's 1 in 12.
• 1:10:10
  That's like a lot.
• 1:10:11
  That's a lot.
• 1:10:12
  That's a lot.
• 1:10:13
  I had no idea this was even a thing when we first put the story in the show notes.
• 1:10:16
  Yeah, this is crazy.
• 1:10:17
  I was like, this is fascinating.
• 1:10:18
  Yeah, yeah, yeah.
• 1:10:19
  So all of that junk DNA, 8% of that is viruses just hanging out.
• 1:10:25
  Okay.
• 1:10:25
  Now, most of these viruses, they get mutations over the years and they can become inactive.
• 1:10:29
  Okay.
• 1:10:30
  But there's a specific virus called the human endogenous virus K.
• 1:10:35
  The Herb K virus.
• 1:10:36
  Okay.
• 1:10:37
  It's a provirus now because it's within our DNA.
• 1:10:39
  But this thing is the most recent infiltrator.
• 1:10:42
  Okay.
• 1:10:43
  And it's because it's the most recent infiltrator.
• 1:10:46
  It has the fewest number of mutations.
• 1:10:48
  Right.
• 1:10:49
  Which means it's still a problem.
• 1:10:51
  Mm-hmm.
• 1:10:52
  Okay.
• 1:10:52
  And it actually...
• 1:10:53
  So in most healthy cells, there's epigenetic silencing.
• 1:10:58
  Okay.
• 1:10:59
  Our DNA has figured out...
• 1:11:00
  I don't know who you are.
• 1:11:02
  I'm going to just put a bunch of random stuff on you so that no one ever transcribes you into RNA and then you become a protein.
• 1:11:10
  Okay.
• 1:11:10
  So...
• 1:11:11
  Our DNA is very good.
• 1:11:12
  If you're healthy, the cell is healthy.
• 1:11:15
  It's like, I don't know what this is.
• 1:11:16
  It's like a bouncer at a club.
• 1:11:18
  Yeah.
• 1:11:18
  It's like, you're already in here, but you're cut off.
• 1:11:21
  Yeah.
• 1:11:21
  You know?
• 1:11:22
  It's like, I don't know how you got in, but no more drinks.
• 1:11:25
  Okay?
• 1:11:26
  Every once in a while, you get something like a cancer cell or you get autoimmune disorders.
• 1:11:34
  The cell is losing control and this virus can start acting up.
• 1:11:39
  So it's actually been detected, this H-Herv-K envelope, the envelope protein that is the protein that encodes for the envelope of the virus.
• 1:11:49
  That's the thing that snuck in.
• 1:11:51
  That has been known to be expressed internally.
• 1:11:54
  In cancers like breast cancer, ovarian cancer, prostate cancer, melanoma, and also in autoimmune disorders like lupus, diabetes type 1, SLE, these kinds of autoimmune disorders, you can actually see the cells that are affected express this HERV-K protein.
• 1:12:12
  Got it.
• 1:12:13
  Which means that it is actually a problem.
• 1:12:15
  Okay.
• 1:12:15
  Right.
• 1:12:15
  Which means it needs to be solved.
• 1:12:17
  Yes.
• 1:12:17
  Okay.
• 1:12:18
  And in biology, like half the battle is solving the structure.
• 1:12:22
  Right.
• 1:12:22
  Because all of biology is lock and key.
• 1:12:24
  Right.
• 1:12:24
  You solve the structure of the thing.
• 1:12:26
  Then you can figure out what other structures can latch on and like mess with it and like inactivate it and things like that.
• 1:12:33
  Right.
• 1:12:33
  So the first first battle is to actually solve the structure.
• 1:12:37
  Solving the structure for this thing was really hard.
• 1:12:40
  Okay.
• 1:12:40
  Okay.
• 1:12:41
  Because it's this envelope protein.
• 1:12:44
  Okay.
• 1:12:45
  It used to be, back in its heyday, before it snuck in and now it's just like an infiltrator.
• 1:12:52
  Back in its heyday, it was the envelope protein for a retrovirus.
• 1:12:56
  And these envelope proteins are incredibly finicky by design.
• 1:13:04
  You want your envelope protein to be finicky because imagine you're a virus, right?
• 1:13:08
  And you're getting to the outside of a cell that you want to infiltrate.
• 1:13:12
  You want the protein to first be able to infiltrate through the membrane.
• 1:13:15
  And then when it's in the membrane, change its conformation to open up.
• 1:13:20
  Right?
• 1:13:20
  So you want this thing to have this sort of meta stability is what it's called.
• 1:13:24
  Where you've got two states.
• 1:13:26
  There's this pre-fusion state and then a post-fusion state.
• 1:13:29
  Right?
• 1:13:30
  Pre-fusion meaning like how do I get in?
• 1:13:31
  Post-fusion, how do I get in?
• 1:13:33
  You have the Trojan horse go to the gate in one state and then when it's inside the gate it's open.
• 1:13:37
  You got to open up.
• 1:13:38
  Yeah.
• 1:13:39
  Right?
• 1:13:40
  So you want this meta stability.
• 1:13:41
  Yes.
• 1:13:42
  Now that becomes a real problem when you're trying to get structure.
• 1:13:46
  Yes.
• 1:13:46
  Right.
• 1:13:47
  Because you can imagine you're trying to image this thing, but this this thing is in two different conformations.
• 1:13:53
  So like you don't know what you're looking at.
• 1:13:55
  Yes.
• 1:13:55
  And it's just annoying.
• 1:13:56
  Yes.
• 1:13:57
  Right.
• 1:13:57
  So that's that's the big thing that this group solved.
• 1:14:01
  OK. What they did, the first thing they did was they took the protein itself and they put in protein.
• 1:14:09
  they tweaked the protein in a certain way that they could, like they created these things called disulfide bridges, which are basically like staples on the protein.
• 1:14:18
  What you do is you create, you put in cysteine, which is this particular amino acid that has sulfur atoms, and those sulfur atoms can find each other and sort of bond.
• 1:14:27
  So you create a more stable protein, which is a protein,
• 1:14:31
  protein that is still very much like the original protein.
• 1:14:36
  It still preserves a lot of the structure, but it prevents the protein from doing this finicky stuff.
• 1:14:41
  Yes.
• 1:14:42
  And then what they do was they use a groundbreaking technology, the one the 2017 Nobel Prize in Chemistry, called cryo-electron microscopy.
• 1:14:51
  Okay.
• 1:14:51
  Okay?
• 1:14:52
  It used to be called blobology.
• 1:14:54
  Blobology.
• 1:14:55
  Yeah, from all the detractors, all the people who would make fun of these cryo-EM people.
• 1:15:00
  It was called blobology because back in the day, what you do is you get a protein and you like image it with electron microscope.
• 1:15:07
  But because of this problem of like, you know, the...
• 1:15:11
  the proteins move around and then they're not stable and all this other stuff.
• 1:15:16
  And you would get blobs, giant blobs, and you'd be like, it kind of looks like this.
• 1:15:21
  So all of the X-ray crystallographers and the NMR spectroscopy people, they're like, look at these blobologists, right?
• 1:15:29
  Because with X-ray crystallography and with NMR, I can get down to atomic resolution.
• 1:15:33
  I can get to two or three angstroms of resolution where I'm like, that atom is there and it's there.
• 1:15:38
  Yeah.
• 1:15:39
  The problem with X-ray crystallography and NMR is it's really good for soluble proteins.
• 1:15:44
  So proteins that are found within the cell doing their thing.
• 1:15:48
  But for these transmembrane proteins, it's very hard to crystallize.
• 1:15:52
  It's very hard to make them behave properly.
• 1:15:54
  Cryo-Yam was a technique that probably could have been used for that.
• 1:15:58
  Along came the computing revolution.
• 1:16:00
  Mm-hmm.
• 1:16:01
  And now all of a sudden what you could do is you could take thousands of images in cryo-EM.
• 1:16:06
  What you do is you have a bunch of proteins that are sort of frozen in this state.
• 1:16:10
  Yes.
• 1:16:11
  Using this thing called vitreous freezing.
• 1:16:14
  Yes.
• 1:16:15
  Where you're not like freezing it with ice because if you freeze it with ice, the protein is going to like misfold, right?
• 1:16:19
  Maybe, yeah.
• 1:16:19
  So you do this like weird like sort of vitreous freezing where the protein retains its structure.
• 1:16:25
  You image it a bunch of times and then you feed that thing to a computer.
• 1:16:29
  Yeah, yeah.
• 1:16:30
  What you're getting is 2D shadows, right?
• 1:16:32
  The proteins are in different orientations.
• 1:16:34
  You get a bunch of 2D shadows.
• 1:16:36
  But a computer now can take all of those 2D shadows and say, what is the 3D thing that would create all of these 2D shadows?
• 1:16:44
  In 2012 was the big one when it came out.
• 1:16:46
  And there was finally a protein that cryo-YAM had solved that no one else could solve with any other technique.
• 1:16:52
  And they solved it to three angstrom precision.
• 1:16:54
  And everyone was like, oh, it's here.
• 1:16:56
  And then in 2017, they won the Nobel Prize in chemistry.
• 1:16:59
  It was three guys, Joaquin Frank, Yaques Duboche, and Richard Henderson.
• 1:17:05
  So blobology now became like a Nobel Prize winning technology.
• 1:17:09
  Ain't no bunch of blobs now, huh?
• 1:17:11
  Yeah, exactly.
• 1:17:11
  Where's your Nobel, huh?
• 1:17:13
  I think they were very vindicated when they actually got it right.
• 1:17:17
  So these guys, they actually used the cryo-electron microscopy to get the data from this.
• 1:17:25
  After having messed with the protein to create these staples and make it sort of…
• 1:17:30
  stable, right?
• 1:17:31
  And this was a very big deal because now you've got the 3D structure for this thing.
• 1:17:36
  You can compare it with other retroviruses that have been solved like HIV, like SIV.
• 1:17:39
  Yes.
• 1:17:40
  And this one is very different because the protein that came out is like taller and skinnier than the HIV protein, which is interesting in its own right because now we've got this difference.
• 1:17:54
  We can now start creating antibodies to actually target this thing.
• 1:17:58
  This lock and key thing you were just talking about.
• 1:17:59
  Exactly.
• 1:18:00
  And that's what they did.
• 1:18:00
  They weren't done.
• 1:18:01
  They created monoclonal antibodies to actually target this thing.
• 1:18:05
  And what they could do now is you can actually have diagnostics where, you know, in autoimmune disorders, for example, right, you have these neutrophils, which are these kinds of immune cells.
• 1:18:14
  And what they do is they they're expressed in these cells and then in an autoimmune disorder,
• 1:18:22
  now your immune cells think that, oh, this is a bad cell.
• 1:18:26
  I'm going to attack my own cell.
• 1:18:28
  Right?
• 1:18:29
  So maybe now that we know the structure, we can actually like prevent that from happening.
• 1:18:35
  Right?
• 1:18:35
  We can train the immune system or maybe cloud...
• 1:18:39
  these antibodies in such a way that the immune system no longer attacks its own.
• 1:18:44
  On the other hand, we can also have ways to early diagnose certain cancers because cancer cells express this protein specifically, right?
• 1:18:53
  So if we get like a blood sample from somebody, we can have an early diagnostic where you have these monoclonal antibodies that'll go and...
• 1:19:02
  actually attached to these cancer cells and you'll have a marker now be like oh i see this herb k protein yes that shouldn't be expressed in a healthy cell the epigenetics should totally silence it and it's not but the bouncer is not working right right so now now i can i can i can start therapies earlier they're still in the club getting too drunk they haven't been cut off yeah so let's go ahead and let's go ahead and and try to figure out how to get these guys out
• 1:19:27
  That's really...
• 1:19:28
  Right.
• 1:19:29
  It's the first time that we've done this.
• 1:19:31
  And the idea is now that we've done it once by confirming that this methodology is a means by which you can do so.
• 1:19:37
  Yeah.
• 1:19:38
  This can likely be extended to other ones.
• 1:19:39
  This is the most recent and all these other things.
• 1:19:41
  Yeah.
• 1:19:41
  And this is the one that's like most prevalent.
• 1:19:43
  But the idea of like sort of changing a protein shape using these editing mechanisms where now I can have this like staple disulfide bridge that sort of...
• 1:19:53
  Stabilizes it such that I can now do cryo-EM.
• 1:19:57
  Yes.
• 1:19:57
  This is a pretty interesting technique.
• 1:19:59
  Right.
• 1:19:59
  The same group actually had done it for other proteins.
• 1:20:03
  And then they tackled this one because this one's hard.
• 1:20:04
  Yeah.
• 1:20:05
  Right.
• 1:20:05
  This one's like.
• 1:20:06
  If you can do it with this one.
• 1:20:07
  Yeah.
• 1:20:08
  Then it's like, okay, now it's like.
• 1:20:09
  Now we can.
• 1:20:12
  Fast.
• 1:20:13
  Yeah.
• 1:20:13
  It's a cool one.
• 1:20:14
  I always love our microbiology immunology stories because.
• 1:20:20
  People don't understand how well we can look at and analyze and understand things that are really, really, really small.
• 1:20:29
  Yeah, like we know exactly where to put in the cysteine so that the sulfur staples to another sulfur, but it doesn't really change the rest of the structure.
• 1:20:38
  That's like to have like this thousands of atoms worth of...
• 1:20:43
  protein and know exactly where to put this thing so that i'm just going to clip this part yes because that's the part that's like getting messed up i would like our next department of health and human services secretary to be able to describe in detail any number of these concepts
• 1:21:00
  Yeah, because that is usually a good prerequisite to say you're going to actually understand the import of medicine for public health.
• 1:21:09
  Exactly.
• 1:21:10
  But but that's a good I like that one.
• 1:21:12
  I like that.
• 1:21:12
  And we've touched on CRISPR a lot on this podcast.
• 1:21:16
  We've touched on a variety.
• 1:21:16
  And it's again, it's interesting.
• 1:21:17
  Again, all of these innovations are happening simultaneously in these cross domain areas that have not only independent value, but also like collective value.
• 1:21:27
  Yeah.
• 1:21:27
  As we move into what we're seeing is going to be this like biotech revolution.
• 1:21:32
  It is.
• 1:21:33
  It really is.
• 1:21:34
  We're here.
• 1:21:34
  For AI applying the tools, the fundamental hardware tools, the creativity of how we're problem solving.
• 1:21:45
  Yeah.
• 1:21:45
  We're going to go to our last story, which is.
• 1:21:48
  A lot of times I do from really, really small now to like – Now we're getting really big.
• 1:21:52
  Now we're getting really big in space.
• 1:21:55
  California again.
• 1:21:56
  California forever goodbye.
• 1:21:57
  Headline, brightest fast radio burst ever detected could help solve an enduring cosmic mystery.
• 1:22:04
  I love how people with headlines make it so unclear what you're talking about.
• 1:22:07
  Yeah.
• 1:22:08
  Summary.
• 1:22:08
  Researchers at University of California, Santa Cruz, we're going to Northern California, use newly developed chime out trigger telescopes and deep space imaging to challenge long-held assumptions about what causes these mysterious cosmic signals.
• 1:22:25
  I think we talked about one of these, if I'm not misremembering, the wow signal was one of these...
• 1:22:32
  Yes.
• 1:22:32
  Like radio burst.
• 1:22:33
  It was a radio burst.
• 1:22:35
  Right.
• 1:22:35
  Yeah.
• 1:22:35
  We initially saw and everyone freaked out and was like aliens.
• 1:22:38
  Yeah.
• 1:22:39
  It could have been.
• 1:22:39
  And that's not the we haven't characterized that as a fast radio burst.
• 1:22:44
  Fair.
• 1:22:45
  Because we don't really know what the hell that was.
• 1:22:46
  That's fair.
• 1:22:47
  That's what I was like.
• 1:22:48
  What was that?
• 1:22:49
  Yeah.
• 1:22:50
  What was that?
• 1:22:50
  Fair.
• 1:22:51
  No, that's fair.
• 1:22:51
  That's fair.
• 1:22:52
  Fair.
• 1:22:52
  But it seems like we're now using, again, some of the tools we have.
• 1:22:56
  Also, this was in the Astrophysical Journal letters.
• 1:22:58
  Yes.
• 1:22:58
  In terms of the publication, CNN has covered this.
• 1:23:01
  UCSC also covered it in their own blog.
• 1:23:03
  Yeah.
• 1:23:04
  It's a collaboration between UCSC, some Canadian institutions, Northwestern.
• 1:23:10
  It was a pretty big collaboration.
• 1:23:11
  Again, because we need it.
• 1:23:14
  Okay.
• 1:23:14
  So what are, because I mischaracterized it.
• 1:23:18
  What is this fast rate of burst idea?
• 1:23:19
  What did they do?
• 1:23:24
  Why is detecting a bright one relevant?
• 1:23:26
  Yeah, this one's incredibly bright.
• 1:23:30
  This thing within a fraction of a second output...
• 1:23:33
  as much energy as the sun does for like four days.
• 1:23:36
  Okay.
• 1:23:37
  But this isn't even the brightest.
• 1:23:38
  Okay.
• 1:23:39
  This is the brightest we've had because it's the closest one to us.
• 1:23:42
  But fast radio bursts, the first one was ever discovered in 2007.
• 1:23:47
  It was called a Lorimer burst.
• 1:23:49
  It was actually taken from archival data from the Parks Observatory, which is a giant 64-meter radio telescope in Australia.
• 1:23:57
  It's been called the greatest scientific instrument...
• 1:24:02
  that Australia has ever produced.
• 1:24:03
  Okay.
• 1:24:04
  The greatest scientific instrument that Australia has ever produced.
• 1:24:08
  Okay.
• 1:24:08
  And it's...
• 1:24:10
  The same dish actually was used during the Apollo 11 broadcast to, like, broadcast to the TVs around the world.
• 1:24:17
  Because NASA had, like, a thing at Goldstone...
• 1:24:20
  Yes.
• 1:24:21
  ...that they were using.
• 1:24:22
  Yes.
• 1:24:22
  But this one was way bigger.
• 1:24:24
  Yeah.
• 1:24:24
  And the TV signal that they were getting from this one was just better than the rest.
• 1:24:28
  And so NASA just used it for the whole two and a half hours...
• 1:24:31
  Which is kind of cool.
• 1:24:32
  Which is why the Five Eyes Ally, you know, structure is awful.
• 1:24:35
  Yeah, yeah, you know.
• 1:24:35
  Yeah, Australia, yeah.
• 1:24:37
  They got our back when we need it.
• 1:24:39
  So, you know, they were on the right side.
• 1:24:41
  They were facing the moon, so that was nice.
• 1:24:43
  Yep.
• 1:24:44
  So this burst was discovered by Duncan Larimar from West Virginia University.
• 1:24:50
  Basically, what he did was he assigned an undergrad to look at archival data from the Parks Observatory.
• 1:24:56
  Okay.
• 1:24:57
  And the Parks Observatory at the time had just released a bunch of data.
• 1:25:00
  It was a radio telescope that was just beaming up.
• 1:25:03
  And this undergrad would come in every week.
• 1:25:05
  And despite all of his classes and stuff, he would look at a bunch of data by hand and try to characterize it and then try to find something interesting.
• 1:25:12
  And he found something very interesting.
• 1:25:14
  There was a brief millisecond duration, five milliseconds.
• 1:25:18
  but it was ridiculously bright.
• 1:25:21
  On top of that, it was not only bright, it had this very good characterization called dispersion.
• 1:25:27
  Okay.
• 1:25:27
  Dispersion is the idea that
• 1:25:31
  When light moves through a medium, through stuff, different frequencies are going to propagate at different speeds.
• 1:25:39
  This is fundamentally why we get a rainbow, okay?
• 1:25:42
  Violet light actually propagates at a slower speed in water than red light does, okay?
• 1:25:50
  And because of that, when light goes into a water bubble,
• 1:25:54
  the violet light is going to bend differently than red light.
• 1:25:57
  And that's what's going to cause the rainbow to come out.
• 1:26:00
  Okay.
• 1:26:00
  And then it gets like sort of focused and we get a rainbow.
• 1:26:03
  But the prism has the same idea.
• 1:26:04
  When it goes through glass, violet light goes slower.
• 1:26:07
  And so it's going to bend differently.
• 1:26:09
  Dispersion is this idea that there's a frequency relation to how fast light moves through medium.
• 1:26:15
  And in interstellar,
• 1:26:18
  background right in the interstellar medium of these ionized particles right where it's like atoms without electrons electrons moving around freely You're gonna have that opposite effect.
• 1:26:28
  You're gonna have slower slower light.
• 1:26:31
  So slower frequencies moving slower than the faster frequencies just has to do with the different physics that goes on between a glass and a plasma but at the end of the day
• 1:26:40
  What that means is if there's a transient burst of light, the faster stuff is going to arrive first and the slower frequencies are going to arrive later.
• 1:26:49
  And that's what's called a dispersion measure.
• 1:26:51
  Okay.
• 1:26:53
  This thing that they discovered in the Parkes Observatory data, which is now called the Lorimer burst, had this very specific dispersion measure, which showed that it was of extra galactic origin.
• 1:27:04
  Okay.
• 1:27:04
  There was so much dispersion.
• 1:27:06
  Yeah.
• 1:27:06
  That meant that there was a lot of stuff that it had to get through.
• 1:27:10
  Yeah.
• 1:27:11
  To actually create that delay between the fast and the slow.
• 1:27:15
  And then when they back calculated how bright this thing must have been to be that far away, they were like, oh, this thing was incredibly bright.
• 1:27:25
  This was back in 2007.
• 1:27:26
  Since then, we've discovered multiple.
• 1:27:28
  There's been some that we've discovered where for that brief delay,
• 1:27:32
  millisecond duration, this fast radio burst was brighter than the entire galaxy that it came from.
• 1:27:38
  Wow.
• 1:27:39
  Okay.
• 1:27:39
  Okay.
• 1:27:40
  Yeah.
• 1:27:40
  So it's been, it's been, it's been quite an insane sort of last few decades of trying to find these fast radio bursts.
• 1:27:48
  Right.
• 1:27:49
  The problem is they're really short lived.
• 1:27:51
  Right.
• 1:27:51
  Okay.
• 1:27:52
  Five milliseconds.
• 1:27:53
  Unlike 3i Atlas, which we're watching for months.
• 1:27:55
  Yeah.
• 1:27:55
  Yeah.
• 1:27:55
  Which we know for months and we can like coordinate this thing.
• 1:27:58
  You know, it's just, it's there and it's gone.
• 1:28:00
  And it happens.
• 1:28:01
  And it just so happens.
• 1:28:02
  Right.
• 1:28:03
  So.
• 1:28:04
  It's been kind of challenging to actually find them.
• 1:28:07
  It's been kind of rare.
• 1:28:08
  Along came this detector called Chime in Canada, which is the... Let me just get the thing right because I don't want any Canadians to get mad at me.
• 1:28:20
  It's called the Canadian Hydrogen Intensity Mapping Experiment, CHIME.
• 1:28:26
  It's a massive radio telescope that doesn't look like most radio telescopes.
• 1:28:30
  So, you know, most radio telescopes, you've got a circular dish with a receiver and the light goes that way.
• 1:28:37
  This thing is a bunch of half pipes.
• 1:28:39
  that are just laid out on the ground.
• 1:28:41
  Oh, interesting.
• 1:28:41
  Giant half pipes, okay?
• 1:28:43
  100 meters that are laid out on the ground, four of them in a valley in British Columbia, as south as you can possibly get because what they're using, they're using the Earth's rotation to basically scan the sky in this radio frequency.
• 1:28:56
  Yep.
• 1:28:57
  And, you know, Canada is pretty far north.
• 1:28:59
  So they want to get as south as possible.
• 1:29:01
  So it's like right across the border from America.
• 1:29:04
  And it's this giant thing.
• 1:29:06
  And it's made such that it's really sensitive.
• 1:29:08
  It's got all these capabilities for as soon as a signal comes in.
• 1:29:12
  It's kind of like the Vera Rubin in terms of they've got online capabilities to analyze the signal.
• 1:29:16
  And it's also got a really good giant field of view.
• 1:29:19
  So it's capturing a big part of the sky as it sort of goes around the earth.
• 1:29:23
  And very recently, so Chime has found thousands of fast radio bursts since 2018.
• 1:29:31
  Okay.
• 1:29:31
  Okay.
• 1:29:31
  It's found a bunch.
• 1:29:32
  This one is very, very nice because usually when Chime finds fast radio bursts, it can't localize where it came from.
• 1:29:39
  Right.
• 1:29:40
  Because it's got this giant field of view, it just sees sort of a signal.
• 1:29:43
  Yes.
• 1:29:43
  And it's like, okay, I saw a fast radio burst.
• 1:29:45
  Yes.
• 1:29:45
  Don't know where it came from.
• 1:29:47
  So what they did was, in collaboration with America, they made more chime stations.
• 1:29:54
  There's one in California.
• 1:29:56
  There's one in West Virginia.
• 1:29:57
  There's one also in Canada.
• 1:29:59
  And now what you can do is triangulate.
• 1:30:02
  My favorite.
• 1:30:03
  Right?
• 1:30:03
  What you can do is, if you get a fast radio burst, you're going to get that signal all over the place.
• 1:30:08
  And you can triangulate where it came from and actually resolve what part of the sky it came from.
• 1:30:12
  Because you have multiple receiving points.
• 1:30:16
  at different places on the planet.
• 1:30:17
  Exactly.
• 1:30:18
  And so the angle...
• 1:30:19
  It's going to be different, and the timing is going to be different, and you can back calculate and see where it came from.
• 1:30:24
  And what they could do is they could resolve it to a really tiny area in the sky, right around the Big Dipper.
• 1:30:31
  The equivalent would be to resolve a quarter from 100 kilometers away.
• 1:30:36
  Very tiny, very tiny, tiny spot.
• 1:30:39
  Right?
• 1:30:39
  And that gives you the advantage.
• 1:30:41
  Because now, if you know that it came from there, you've got all these other telescopes on standby that can point right there.
• 1:30:48
  Right?
• 1:30:49
  That's exactly what they did.
• 1:30:50
  They got the James Webb.
• 1:30:51
  They called in the big guys.
• 1:30:52
  They called in the big guns.
• 1:30:53
  They called in the big guns because this one was really bright.
• 1:30:56
  Okay.
• 1:30:57
  In terms of this is the brightest fast radio burst that we've ever seen because of its proximity to Earth.
• 1:31:02
  Not in terms of intrinsic brightness, but because of its proximity to Earth, now it's really exciting, right?
• 1:31:07
  Because now we can point optical telescopes to it and we can actually see what part of this galaxy that it came from.
• 1:31:12
  It came from NGC 4141, which is this barred spiral galaxy.
• 1:31:17
  And what it does is, and you can localize the neighborhood of that galaxy.
• 1:31:24
  Not just that it came from that galaxy, but it came from this corner of the galaxy.
• 1:31:29
  Okay.
• 1:31:30
  And now you can point James Webb Space Telescope to it.
• 1:31:33
  And James Webb was pointed to it.
• 1:31:35
  It found a faint signal in the infrared.
• 1:31:37
  Okay.
• 1:31:38
  Right around that same spot.
• 1:31:40
  Okay.
• 1:31:40
  Okay.
• 1:31:41
  And that probably came from some massive stars, like a red giant or something like that.
• 1:31:45
  Now, it doesn't quite fit that a red giant or a big star would create a fast radio burst.
• 1:31:51
  Okay.
• 1:31:52
  Okay?
• 1:31:52
  Because those things are, like, slow.
• 1:31:56
  Right.
• 1:31:56
  They're not five millisecond timescales.
• 1:31:59
  Right.
• 1:32:00
  Five millisecond timescales are the stuff of black holes, neutron stars.
• 1:32:05
  They revolve really fast.
• 1:32:07
  So there's timescales that sort of correspond to astronomical objects.
• 1:32:11
  Yes.
• 1:32:11
  Right?
• 1:32:14
  So now it leaves us with a conundrum.
• 1:32:16
  It's like, where did this thing come from?
• 1:32:18
  Mm-hmm.
• 1:32:19
  There's hypotheses.
• 1:32:20
  There's not good data.
• 1:32:21
  Because again, this is the first time.
• 1:32:23
  I mean, Chime and the Chime outrigger stations only came online like this year.
• 1:32:31
  Okay.
• 1:32:31
  So there's the first thing that it's found.
• 1:32:33
  And it's very exciting because it's kind of a proof of concept that it's like looking and it can find it.
• 1:32:37
  And then we can go off with...
• 1:32:39
  with James Webb Space Telescope, with Keck, with Gemini, and pointed at that and then actually see a follow-up observation, right?
• 1:32:46
  So we still don't know really where this thing came from.
• 1:32:50
  What we do know, which is very interesting, is this might be a one-off.
• 1:32:54
  Most of the fast radio bursts that we've seen, they seem to be repeating fast radio bursts, right?
• 1:32:59
  Where they've got this rhythmic sort of structure.
• 1:33:02
  The rhythm can be very long.
• 1:33:03
  It can be from days to months to years.
• 1:33:06
  But...
• 1:33:07
  they've got this rhythmic thing where it's like okay you've got one really bright one but then every once in a while the same source should put off like sort of smaller bursts which i think is why it matters that we have this this multi this multi-sensor data confirmation multi-sensor collaboration to be able to isolate a smaller point in the sky because if we do find these rhythmic sources we can point these things these other tools there
• 1:33:32
  for longer periods of time.
• 1:33:34
  Exactly.
• 1:33:34
  To try to potentially catch what is this.
• 1:33:37
  It reminds me of the transients conversation we had the other day where it's like there's this rhythmic pattern of the planet, the exoplanet going in front of a star.
• 1:33:43
  Yeah.
• 1:33:43
  And so you have to monitor it for certain periods of time to understand that it's not a one-off.
• 1:33:47
  It's not a one-off, yeah.
• 1:33:48
  And so what's interesting here is like what we're potentially seeing with this specific really, really bright one may have been a one-off.
• 1:33:54
  It may have been a one-off.
• 1:33:55
  I mean, we think it's a one-off because Chime was going around
• 1:34:00
  That same spot in the sky for the past seven years, it's had hundreds of hours in that region of the sky.
• 1:34:05
  It's the only one.
• 1:34:06
  So we've only seen it once, right?
• 1:34:08
  And it would have found even smaller radio bursts, but it hasn't.
• 1:34:12
  It hasn't.
• 1:34:13
  And so it's like, what could be this one-off thing?
• 1:34:16
  It could be, they're thinking it might be a Magnetar, which is this, a Magnetar is a- Not a Pokemon.
• 1:34:23
  No, it sounds like one though.
• 1:34:25
  A Magnetar is sort of a newly born neutron star.
• 1:34:30
  that's highly magnetized.
• 1:34:32
  So it's got really powerful magnetic fields and it's in a sort of new stellar neighborhood that has magnetic material around it.
• 1:34:40
  So it can create these fast radio bursts by sort of, maybe it's like accumulating material and then it blows up or something.
• 1:34:48
  Maybe it's accumulating material from the red giants that the James Webb Space Telescope saw.
• 1:34:53
  James Webb saw some signal.
• 1:34:54
  They just don't know.
• 1:34:55
  It's like not good enough.
• 1:34:56
  It's not good enough.
• 1:34:57
  And because it's transient, it only happened for a little bit.
• 1:34:59
  And then it's gone.
• 1:35:01
  It's not like a star that I can just keep looking at and getting data.
• 1:35:06
  It could be a magnetar.
• 1:35:08
  It could also be a colliding neutron star.
• 1:35:10
  But if it was colliding, sort of revolving colliding neutron stars, we would have seen LIGO.
• 1:35:15
  We would have seen gravitational waves, which we haven't.
• 1:35:19
  in the hours preceding or since.
• 1:35:21
  So maybe the collision is gonna happen later, and this is like maybe they're getting close and doing some weird dance, and then that released a giant fast radio burst.
• 1:35:32
  But it's a new scale of astronomy, right?
• 1:35:35
  Where before we had like only a few stations that were relying on the Earth's rotation to like maybe catch something.
• 1:35:42
  Yes.
• 1:35:42
  Right.
• 1:35:43
  And now we've got a dedicated sort of instrument that is doing it on the Northern Hemisphere.
• 1:35:48
  You can imagine we can create a global celestial radio instrument that now does this.
• 1:35:53
  Actually, the VLA also took a look at it.
• 1:35:55
  Yes.
• 1:35:56
  Because now that we can locate it here, the VLA took a look at it for like several days afterwards trying to see if there's going to be a...
• 1:36:03
  A repeat.
• 1:36:04
  Yes.
• 1:36:04
  Didn't see one.
• 1:36:05
  The ZLA being the Very Large Array.
• 1:36:07
  Yes, the Very Large Array in New Mexico, which is a giant interferometer of a bunch of radio dishes on train tracks.
• 1:36:13
  For those who, I think the movie was Contact with Jodie Foster.
• 1:36:20
  Yeah.
• 1:36:21
  Written by Carl Sagan.
• 1:36:22
  Written by Carl Sagan.
• 1:36:23
  Was at the Very Large Array.
• 1:36:25
  Yeah.
• 1:36:26
  But again, this goes back to, I'm glad the outriggers were able to get out before funding out.
• 1:36:31
  Yeah.
• 1:36:32
  This is why things like, you know, Chime, the Vera Rubin, which have this very specific use case of SkyService that are not that are just saying we're just going to get this.
• 1:36:42
  Yeah, this data set.
• 1:36:44
  No, because we don't know where to look.
• 1:36:45
  We don't know where to look.
• 1:36:47
  We're just looking everywhere.
• 1:36:48
  So we're going to create this map.
• 1:36:49
  Yeah.
• 1:36:50
  And then that'll help us know where to look.
• 1:36:51
  Yeah.
• 1:36:52
  And then everyone can write their proposals to then use these specific other tools to then look much more deeply.
• 1:36:58
  Yeah.
• 1:36:58
  And it's kind of, I mean, fast radio bursts are only 20 years old, you know, in terms of the grand scale of astronomy.
• 1:37:05
  Right.
• 1:37:05
  That's hundreds of years old science.
• 1:37:08
  Yes.
• 1:37:08
  Right.
• 1:37:09
  Like only 20 years ago, we found the first one.
• 1:37:11
  And these things are so ridiculously bright.
• 1:37:15
  And it's like, what could possibly be making something that is as bright as a galaxy for a fraction of a second?
• 1:37:21
  Right.
• 1:37:22
  Right.
• 1:37:22
  You know?
• 1:37:24
  Don't get the Dyson Sphere people in on this.
• 1:37:26
  Because they're like, oh, obviously.
• 1:37:28
  It's an alien civilization that's harnessed the power of a planet.
• 1:37:31
  And then when they recycle their generatron, it sends out this massive, I'm being facetious.
• 1:37:37
  Yeah.
• 1:37:38
  We should have an episode about Freeman Dyson, though.
• 1:37:41
  He's one of my favorite physicists.
• 1:37:42
  Same.
• 1:37:43
  He's one of my favorite physicists of all time.
• 1:37:45
  I love Dyson.
• 1:37:47
  Very clever.
• 1:37:48
  Interesting.
• 1:37:49
  Really funny, too.
• 1:37:49
  Yeah, yeah.
• 1:37:50
  Really funny guy.
• 1:37:52
  Brightest fast radio bursts ever detected.
• 1:37:55
  Yeah.
• 1:37:55
  Because of some of the tools that we have.
• 1:37:57
  This is such a new area.
• 1:37:59
  Such a new area.
• 1:38:01
  And we haven't figured it all out yet.
• 1:38:03
  No.
• 1:38:04
  There are really interesting areas.
• 1:38:05
  No, there are mysteries out there in the universe.
• 1:38:08
  And we are doing good work to...
• 1:38:11
  Try to narrow the path towards discovery on some of these things.
• 1:38:16
  Yeah.
• 1:38:16
  And it's like, you know, when the Vera Rubin, I was listening to the press conference for the Vera Rubin the other day, and they were talking about, they used Donald Rumsfeld's Unknown Unknown.
• 1:38:28
  Classic.
• 1:38:29
  But it actually made sense here because there are the known unknowns, which is like what is causing the expansion of the universe?
• 1:38:37
  What is dark energy?
• 1:38:38
  What is dark matter?
• 1:38:40
  What are these fast radio bursts?
• 1:38:42
  We know that they're there, but they're unknown in what they are.
• 1:38:46
  And then he said, what I'm really excited about with the Vera Rubin is the unknown unknowns.
• 1:38:51
  Because before 2007, fast radio bursts were an unknown unknown, right?
• 1:38:56
  Yeah.
• 1:38:56
  People couldn't even imagine that something like this would exist.
• 1:38:59
  Right.
• 1:38:59
  Where in five milliseconds, it dumps like a ridiculous amount of energy.
• 1:39:03
  Yes.
• 1:39:03
  You know, that timescale for that amount of energy is insane.
• 1:39:07
  And who knows how many, what number of unknown unknowns.
• 1:39:10
  The Verirubin is going to find.
• 1:39:11
  As we have these better tools.
• 1:39:12
  Yeah.
• 1:39:12
  And what the Chime is going to find too.
• 1:39:14
  Maybe it'll find like...
• 1:39:15
  Super fast.
• 1:39:16
  Yes.
• 1:39:17
  Radio bursts, you know, that are like.
• 1:39:18
  Right.
• 1:39:19
  At the microsecond level.
• 1:39:20
  Right.
• 1:39:21
  You never know.
• 1:39:21
  Never know.
• 1:39:22
  You never know.
• 1:39:23
  We always cover a really fascinating array of stories each week.
• 1:39:29
  We started with a deep dive on Deep Seek.
• 1:39:32
  No pun intended.
• 1:39:34
  And the geopolitical power war that is currently happening between the U.S. and China as leaders.
• 1:39:40
  Europe, you know, eat your heart out.
• 1:39:43
  You know, Mistral is not really in the same category.
• 1:39:46
  We went then to time crystals, the first time crystals that are room temperature and macroscopic.
• 1:39:53
  That I can see.
• 1:39:53
  That we can see.
• 1:39:56
  Out of University of Colorado Boulder.
• 1:39:59
  Implications for any number of use cases, like security features, technology devices.
• 1:40:08
  The third story was really fascinating about hidden viruses in our DNA.
• 1:40:12
  Could be big for medicine breakthroughs.
• 1:40:14
  Yeah.
• 1:40:15
  The fact that viruses are 8%
• 1:40:18
  Of our DNA.
• 1:40:18
  Of our DNA.
• 1:40:21
  And that there's these indicators that drive cancer and autoimmune diseases.
• 1:40:24
  And understanding the structure, understanding the lock, now gives us the ability to build a key.
• 1:40:30
  And then fast radio bursts.
• 1:40:32
  Still an enduring mystery.
• 1:40:33
  Yeah.
• 1:40:35
  about where is all of this massive scale of light in such a short time scale?
• 1:40:43
  What cosmological event could create something like this?
• 1:40:51
  Another great week.
• 1:40:52
  Again, the feedback.
• 1:40:53
  We're still on the top charts on Apple for Science Podcasts.
• 1:40:57
  Keep telling your friends.
• 1:40:58
  Please keep sharing with your friends.
• 1:41:00
  The comments are hilarious.
• 1:41:02
  All of you are comedians.
• 1:41:04
  Some really, really funny comments.
• 1:41:06
  I do talk.
• 1:41:07
  I don't just sit here and nod my head.
• 1:41:09
  Also, my hairline is not that crazy.
• 1:41:11
  This is not Stephen A. Smith.
• 1:41:13
  I've always had a big forehead.
• 1:41:14
  And he's not the black guy from New Girl.
• 1:41:17
  I'm not Lamorne Morris, although I appreciate it.
• 1:41:20
  Lamorne, if you want to have us on the Lamorning podcast, we're literally down the street.
• 1:41:24
  We'd love to talk to you about science.
• 1:41:27
  My name is Lester Nare, joined as always by my co-host and our resident PhD, Krishna Chowdhury.
• 1:41:34
  This is From First Principles.
• 1:41:36
  We'll see you all next week.
• 1:41:37
  Peace.