Latest / In the Loop AI / Human Brain Cells in a Petri Dish Just Played DOOM (This Is Real)

Transcript

Human Brain Cells in a Petri Dish Just Played DOOM (This Is Real) · In the Loop AI

0:00
Ejaaz: The entire AI industry is in a massive race to build silicon-based GPUs.
0:05
Ejaaz: We've burned billions of dollars building the biggest data centers,
0:09
Ejaaz: getting as much compute as we can.
0:10
Ejaaz: But what if the best AI hardware already exists?
0:13
Ejaaz: What if that best AI hardware is human flesh, human cells, animal cells that
0:18
Ejaaz: we can train to emulate AI models?
0:21
Ejaaz: This week, two stories broke that sound very much like science fiction,
0:25
Ejaaz: but is actually very much real.
0:27
Ejaaz: In Australia, they fused 200,000 human cells onto an AI chip and taught it how
0:33
Ejaaz: to play the computer game Doom. And it's actually pretty good at it.
0:37
Ejaaz: In another story, they uploaded the entire brain of a fruit fly into a single
0:42
Ejaaz: laptop and had it navigate around a simulated world.
0:45
Ejaaz: The craziest part is it was 91% accurate in terms of flying and moving about.
0:50
Ejaaz: So we're reaching this point where AI is converging with biology and something
0:54
Ejaaz: known as AI wetware. And it might be the next biggest thing.
0:57
Josh: It is a good time to be a fan of sci-fi because everything that you've read
1:00
Josh: in those books, it's actually starting to become true in reality.
1:04
Josh: I famously, I loved watching Black Mirror, the like kind of sci-fi dystopian future.
1:08
Josh: At early days, it was very much ahead of its time, but it feels as if we have
1:11
Josh: very quickly caught up and perhaps even exceeded the kind of crazy,
1:15
Josh: scary stories that are in Black Mirror, starting with this revelation that we've
1:18
Josh: had, this new breakthrough through this company called Cortical Labs,
1:21
Josh: who, like you mentioned briefly,
1:22
Josh: they trained human brain cells to play the game of doom.
1:26
Josh: Better than the average person can like pretty well it
1:29
Josh: actually works it's this really unbelievable story and i guess before we get
1:32
Josh: into what they did today i want to go through the brief history of
1:35
Josh: this company named cortical labs because they've been trying to figure out
1:38
Josh: this biological human computer for a while and in 2021 five years ago um they
1:43
Josh: debuted this thing called dish brain which was this early computer it used about
1:46
Josh: 800 000 human nerve cells so it wasn't brain cells this was nerve cells and
1:51
Josh: it was capable of interpreting direct electrical activity so when you stimulated
1:54
Josh: these nerve cells, they would light up.
1:56
Josh: And they had this kind of pseudo computer, didn't actually do anything,
1:59
Josh: but they were able to put input in and get something out.
2:02
Josh: And then the next year in 2022, they made headlines again, teaching these mini
2:06
Josh: brains of 800,000 to 1 million human brain cells that were in a Petri dish that
2:11
Josh: learned how to play Pong.
2:12
Josh: So four years ago, these brain cells were playing Pong, but today they're actually playing Doom.
2:18
Josh: I mean, this is just insane. They trained human brain cells to play a video game in a Petri How
2:23
Ejaaz: They did this was crazy. So you mentioned that they were human brain cells,
2:26
Ejaaz: but they didn't extract these from people's brains.
2:30
Ejaaz: They actually took skin cells or blood cells from some human donors,
2:34
Ejaaz: They then converted them into stem cells, which is like the god cell,
2:39
Ejaaz: if it were, where you can change it into any type of cell, a heart cell,
2:42
Ejaaz: an immune cell, or whatever it might be.
2:44
Ejaaz: They turned it into brain cells from there and then fused 200,000 of them or
2:48
Ejaaz: cultured 200,000 of them on something known as a microarray chip,
2:53
Ejaaz: which is a type of AI-based chip.
2:56
Ejaaz: They grew it on these chips and then they plugged it into or they wired it into
3:00
Ejaaz: the Doom game and taught it over the course of a week how to play Doom.
3:05
Ejaaz: So it would prompt it with certain signals, it would pick up what it needed
3:09
Ejaaz: to do to initiate a particular action, and it would train itself to do it.
3:13
Ejaaz: Now, two important things that I want to point out here is typically when you're
3:18
Ejaaz: training an AI model to simulate human intelligence, it takes so much data and
3:23
Ejaaz: so much money and so much time.
3:26
Ejaaz: With the human cells, they noticed that one, it took so much less energy,
3:31
Ejaaz: It barely required any energy to actually do this thing.
3:33
Ejaaz: It actually did it on a single tiny machine that cost around $35,000.
3:38
Ejaaz: And it did so in a week, which is much quicker, which leads me to the third point here.
3:43
Ejaaz: The human cells instinctively picked it up.
3:45
Ejaaz: It's like they had the training knowledge already baked into the DNA of the
3:49
Ejaaz: human cells itself. And that's like super cool.
3:52
Josh: Yeah, I think one of the most noteworthy things here is that energy thing.
3:56
Josh: And we're probably going to get more into this, but like a general GPU rack
3:58
Josh: uses, what, like 700 megawatts or something, a GPU.
4:01
Josh: The brain does all this on 20. a small crazy
4:05
Josh: minor fraction of what's actually being run on these gpus
4:08
Josh: it's incredibly efficient and it's also a
4:10
Josh: testament to how things how we learn how things are trained
4:13
Josh: so with humans we're very reactionary we learn through feedback from the environment
4:17
Josh: with no preloaded data there's no gradient descent no training runs it's just
4:21
Josh: actual biology it's baked in over our evolution llms on the other hand they
4:25
Josh: learn by processing tons of text data and then adjusting their parameters through
4:29
Josh: a process called back propagation so they're very deterministic They're silicon-based.
4:33
Josh: They are inefficient relative to brains, and they don't capture a lot of the
4:37
Josh: evolutionary, I guess, improvements that we have over time.
4:40
Josh: What this computer does is it's the same way that humans do.
4:43
Josh: It learns by being reactionary. It understands that when it makes the right
4:47
Josh: decision, a specific set of neural pathways light up, and it tries to make that
4:51
Josh: right decision over and over and over again.
4:53
Josh: And doing it with human brain cells, man, it's so weird. It's so sci-fi.
4:57
Ejaaz: Well, just to give the Dura comparison, we're talking about 200,000 human brain
5:02
Ejaaz: cells in this experiment.
5:03
Ejaaz: The entire brain has 86 billion neurons. And it runs a billion, right?
5:09
Ejaaz: So like we're talking massive, massive amounts more.
5:12
Ejaaz: And it runs on 20 Watts of power. So it's incredibly efficient.
5:16
Ejaaz: And what this tells me is,
5:19
Ejaaz: I think we've been building AI models slightly incorrectly.
5:22
Ejaaz: I say slightly incorrectly as maybe a massive understatement.
5:27
Ejaaz: The point is, we've put in so much time, energy, and money into assuming that
5:32
Ejaaz: the best way to build intelligence or artificial intelligence is to slap it
5:35
Ejaaz: on a silicon chip, you know, sand and glass, basically.
5:39
Ejaaz: But maybe the best way to do it is just to pick up the best natural organic
5:44
Ejaaz: intelligence model that has been trained over millions and millions of years, which is,
5:49
Ejaaz: the organic human brain or the animal brain and kind
5:53
Ejaaz: of like use those cells to mimic what intelligence
5:56
Ejaaz: is that we're trying to build in for different software so
5:59
Ejaaz: that's one big takeaway the other thing is like maybe we just need
6:02
Ejaaz: so much less energy than we expected to build artificial intelligence and if
6:06
Ejaaz: something like this doom experiment can scale to the size of something like
6:11
Ejaaz: a human brain then i don't really understand what the moat is for all these
6:15
Ejaaz: ai labs that are spending hundreds
6:17
Ejaaz: of billions of dollars to train their artificial intelligence models.
6:20
Ejaaz: Then the final thing, which is what you mentioned, is just the fact that it needs no training.
6:25
Ejaaz: It's so instinctive. It's almost like the human brain is the best AI training
6:30
Ejaaz: run that has ever occurred. It really is.
6:33
Josh: And I guess I would bucket this probably in the same place that I bucket like
6:37
Josh: a company like Ilias at Skever, Safe Super Intelligence, where he is one of
6:41
Josh: the famous AI researchers. He left, he started his own company.
6:44
Josh: And the sole purpose of the company is to discover novel breakthroughs that
6:47
Josh: rapidly accelerate through this curve.
6:49
Josh: So when you think of the AI curve that we're going through right now,
6:52
Josh: there is a not so predictable, but a very clear exponential curve.
6:56
Josh: And the goal of some company like Ilias or perhaps a company like this that
6:59
Josh: we're covering right now is that there are actual improvements that are novel
7:04
Josh: that result in these 10x accelerations of this natural curve.
7:07
Josh: So biological computing is one of them.
7:10
Josh: If we can actually figure out how to crack this at scale, if we can get closer
7:14
Josh: to those 86 billion neurons, I mean, when you think about the human brain,
7:17
Josh: we use what a very small percentage of our brain relative to what exists.
7:21
Josh: A computer doesn't have that limitation. So even if we get a small fraction
7:24
Josh: of what's available for the human brain, I mean, you can see this get pretty wild pretty quickly.
7:29
Josh: Again, this is sci-fi. It's not in production. This is small scale testing.
7:32
Josh: I kind of put quantum computing in a similar pipeline where it is this crazy,
7:36
Josh: weird, pseudo magical compute platform that has the ability to revolutionize everything around us.
7:41
Josh: It's just a little early. But what we can guarantee is that the rate of acceleration for AI...
7:46
Josh: Is upstream of all of these things coming much faster. Because the AI can help
7:50
Josh: us engineer and process this and accelerate the timelines of what previously
7:54
Josh: would have been perhaps decades down to compressing it down to a few years.
7:58
Ejaaz: Josh, can I ask you a question? What do you got? Doesn't this feel kind of toppy?
8:03
Ejaaz: This feels like we're topping on the market. Like the AI bubble bursting now
8:08
Ejaaz: kind of makes a little more sense to me.
8:10
Josh: Yeah, I guess the question is, like, is this a real threat? Because clearly,
8:13
Josh: it's much more energy efficient.
8:14
Josh: Clearly, the energy efficiency is the biggest threshold. We just don't have
8:17
Josh: enough energy to power these chips.
8:18
Josh: We just had the conversation about Leopold last week, who is fully pivoting
8:21
Josh: from chips to energy. That is the biggest thing in the world.
8:24
Josh: But it doesn't seem like the market thinks this is a problem.
8:27
Josh: They're not really pricing this in.
8:28
Josh: We have the Polymarket, and this
8:29
Josh: market in particular is about when the AI bubble is going to burst by.
8:32
Josh: The end of March, March 31st, the percentage is 3%. The bubble will not be popping
8:38
Josh: by the end of this month. We still have some runway left.
8:40
Josh: December 31st, 2026, the end of this year, still only a 20% chance on over $2
8:45
Josh: million of volume on this market.
8:47
Josh: So the market is signaling they don't care too much. Now there is something
8:51
Josh: interesting here in that that number has crept up recently.
8:54
Josh: This number was not always 19%. It has crept up to 19%. So a slightly higher
8:59
Josh: than normal probability, but all signs point to the fact that the market doesn't
9:03
Josh: really care about this. It's cool.
9:05
Josh: It's sci-fi, but it's going to remain in that sci-fi bucket.
9:07
Josh: They're not actually going to crack this.
9:08
Josh: We still need millions of VPUs and tons of gigawatts power in these data centers for now, at least.
9:14
Josh: And thank you to Polymarket for supporting this episode.
9:17
Ejaaz: Yeah, for now, Silicon keeps winning. But there's another crazy story that we
9:21
Ejaaz: have to cover. Number two.
9:23
Josh: Okay, yeah, this one is awesome. Let's unpack it.
9:25
Ejaaz: This is insane. So I want to lead with the opening sentence of this post.
9:30
Ejaaz: There's a fruit fly walking around right now that was never born.
9:34
Ejaaz: And she goes on to talk about how a team in San Francisco uploaded an entire
9:39
Ejaaz: brain map of a real organic fruit fly into a laptop and got it to simulate its
9:45
Ejaaz: own life to a 91% accuracy.
9:48
Ejaaz: We're talking about it learned how to move, fly, and navigate an entire world
9:53
Ejaaz: by itself. So it's as if the fly exists and is real, but it's completely simulated.
10:00
Josh: Think of the brain like a city right like before you can
10:02
Josh: simulate traffic you need to map every single road every intersection every
10:06
Josh: one-way street it's kind of like what self-driving cars are trying to do but what
10:09
Josh: scientists did is they spent nearly a decade rebuilding that map for a fruit
10:12
Josh: fly so what they do is they slice the brain into very small pieces they encase
10:16
Josh: it in this resin and they scan each slice of the brain until they have this
10:21
Josh: representation of what the brain looks like throughout every single layer of
10:25
Josh: it and they actually accomplished this in 2024.
10:27
Josh: So we had a brain copied on a laptop two years ago.
10:31
Josh: And this was done by assembling the 7,000 smallest slices of a brain.
10:35
Josh: What is new here is they actually took that brain and they placed it into a
10:39
Josh: digital representation of a fly and let it live its life as if it was a fly.
10:45
Josh: So they sacrificed a real fly, sliced the brain into pieces,
10:49
Josh: trained the computer, or not even trained the computer, but diagnosed what was
10:52
Josh: in every single layer to rebuild the digital version of the brain.
10:55
Josh: And now this digital fly is actually walking around in a 3D game engine in a
11:01
Josh: computer without any training at all, just by using the digital clone of its brain.
11:06
Josh: And that is the part that's absolutely insane to me. Like there is a real fly
11:11
Josh: brain currently walking around a digital environment. Like that is crazy.
11:16
Ejaaz: This, not to add a pun here, but like completely messes with my own mind,
11:21
Ejaaz: because we just spoke about a story of putting intelligence into human organic
11:27
Ejaaz: life forms, human cells.
11:28
Ejaaz: And now we're talking about taking real organic life form intelligence and putting
11:33
Ejaaz: it into an artificial form into a laptop and trapping it inside there, right?
11:38
Ejaaz: So this kind of opens up so many questions for me, like, number one,
11:42
Ejaaz: why did it not require any kind of training at all?
11:45
Ejaaz: And it's what we were discussing earlier, which is, I think a lot of the ways
11:48
Ejaaz: that human life forms are built using genetic material, phenotypes,
11:52
Ejaaz: genotypes, engineer it to be able to react to the world in a very different
11:56
Ejaaz: way. Andre Carpathy talks about this a lot.
11:58
Ejaaz: He says that AI models can't really mimic human evolution very well.
12:03
Ejaaz: Only organic life has been successfully able to mimic that and have consciousness.
12:08
Ejaaz: So this is one example of that happening.
12:10
Ejaaz: Now, a direct comparison with
12:11
Ejaaz: the human cells playing Doom, we had 200,000 neurons being used there.
12:16
Ejaaz: This one required 140,000 neurons to recreate the entire organism itself,
12:21
Ejaaz: which is just super cool.
12:23
Ejaaz: And then the third point that's interesting here that I want to put out here,
12:26
Ejaaz: you mentioned it just now, Josh, this wasn't like a single team doing this.
12:30
Ejaaz: They pulled already available data of this entire genome.
12:35
Ejaaz: So it's kind of like this weird world that we live in, where we could just knock
12:39
Ejaaz: on the door of like the National Archives, which have collected genome data
12:43
Ejaaz: for not just fruit flies, but a bunch of other animals.
12:46
Ejaaz: Copy, paste that into a computational AI model, and then just upload it to a
12:50
Ejaaz: laptop and see what it does in a simulated environment.
12:53
Ejaaz: Like, I want to see this happen with more organisms. I'm like,
12:56
Ejaaz: what does a tiger do if we take the entire mapped brain of a tiger and upload
13:00
Ejaaz: it to a computer? Does it do the same thing? Is it different? It's just super cool.
13:03
Josh: It's amazing to see too. There's a video towards the bottom of this post,
13:06
Josh: I believe, which actually shows what that looks like as the neurons are firing off.
13:09
Josh: And you could see very clearly that this is an emergent behavior.
13:14
Josh: Like this wasn't taught, this fly wasn't taught how to walk.
13:18
Josh: It didn't use any gradient descent. It didn't use any technology that we're using for current LLMs.
13:23
Josh: And yet it is walking around this 3D space as if it was a fly.
13:26
Josh: And you could see all the neurons firing off at once. And it's this unbelievable,
13:30
Josh: I guess, early prototype of what this could look like at scale, right?
13:33
Josh: Like you mentioned that total neuron count, a fly brain has,
13:36
Josh: I think 140,000, a mouse brain has 70 million, a human brain has 86 billion.
13:42
Josh: So assuming that we're able to follow this natural progression of copying more
13:45
Josh: and more, you have to assume eventually we're going to get some pretty powerful
13:48
Josh: brains inside of a computer.
13:51
Josh: And there is nothing that isn't like dystopian, crazy sci-fi about this.
13:56
Josh: Anyway, you think about it where now we sacrificed one brain and you have a
14:00
Josh: complete and total digital clone of the other and it begs a lot of questions
14:03
Josh: of what actually makes up a person a mammal an insect a living thing if you
14:09
Josh: can just kind of copy it and clone it inside of this machine and it's really unbelievably
14:14
Josh: Bizarre story and i guess maybe we can get into why this is really the world's
14:19
Josh: best training run right it's like yeah as we start to emulate these human brain
14:22
Josh: cells as we emulate a fly's brain
14:25
Josh: these materials, these brains have been benefits of evolution over billions of years.
14:31
Josh: And I think, oh my God, Luke producer before the show, he mentioned there's
14:35
Josh: been one septillion fruit flies in existence, which is so many training runs over and over and over.
14:41
Josh: And through the process of natural selection over this septillion,
14:44
Josh: that's 10 to the 24 fruit flies, they've just evolved and they've improved.
14:48
Josh: And what you're noticing is you drop this brain into the space and it knows
14:51
Josh: how to act. It knows how to walk. You don't have to teach it.
14:53
Josh: And that is something that is an emergent property of AI, but seemingly baked
14:58
Josh: into this biological process. It's so cool.
15:01
Ejaaz: It's like the genetic wiring is the intelligence in itself.
15:05
Ejaaz: Like however long humans have existed over hundreds of millions of years,
15:08
Ejaaz: that is the equivalent of the best AI training run ever, right?
15:12
Ejaaz: So we're sitting here trying to emulate intelligence by building out these models
15:16
Ejaaz: with different weights, tweaking parameters. Oh, we've got a trillion parameter thing.
15:20
Ejaaz: What if we just copy pasted the brain? That's what both of these stories have
15:24
Ejaaz: taught me, at least, it's like we could just maybe copy and mimic intelligence
15:27
Ejaaz: from organic matter itself and have that inform how we build artificial intelligence completely.
15:33
Ejaaz: The other kind of weird sci-fi thing here, Josh, where my mind jumped to at least is,
15:39
Ejaaz: if you could upload an entire animal's brain and maybe eventually a human brain,
15:44
Ejaaz: you could do that to modify and up-level humans massively.
15:50
Ejaaz: Like, let's say you wanted to learn how to play the violin to expert grade level.
15:54
Ejaaz: Oh yeah, just upload your brain. We'll run it on a training cycle for about
15:58
Ejaaz: five days because it's your human brain. It actually learns way,
16:02
Ejaaz: way quicker than silicon.
16:03
Ejaaz: And then we'll just download it back in your brain and there you go.
16:05
Ejaaz: So the ultimate form factor that this ends up being, in my opinion,
16:09
Ejaaz: is the brain-computer interface. And we've said this a number of times on our show.
16:13
Ejaaz: We do think that AI's ultimate form isn't going to be a physical manifestation
16:17
Ejaaz: in robotics, or it's not just going to be a digital manifestation in LLMs.
16:21
Ejaaz: It's going to be something of a hybrid between the greatest,
16:24
Ejaaz: most intelligent organisms, which are the humans, with the artificial compatible
16:29
Ejaaz: version, which is the LLMs, fuse them in a chip.
16:32
Ejaaz: I guess, bullish Neuralink from this.
16:35
Josh: Yeah, it seems like, I mean, this is the natural progression,
16:37
Josh: right? Everyone jokes about the singularity, but we slowly are emerging.
16:40
Josh: And it seems like the extended version of this, like if you kind of take this
16:43
Josh: to the limit, is this convergence of this humanoid tissue and then this digital
16:49
Josh: form of intelligence. It's a really weird place.
16:51
Josh: I guess there are some things that are noteworthy that we probably should mention.
16:54
Josh: And the fact that this is not very good, and this is still very early.
16:57
Josh: So in the case of Doom, this little clump of brain cells, it's contained in
17:01
Josh: this little device that costs $34,000 that needs to be climate controlled.
17:05
Josh: The cells only last six months.
17:07
Josh: It's important to note these cells don't have pain receptors.
17:09
Josh: There's nothing really human about them outside of them being human-derived cells.
17:13
Josh: And the actual output of this gameplay was slightly better than GPT-4,
17:19
Josh: which is noteworthy, but still pretty bad.
17:21
Josh: It's not actually good at the game. So this is very much a proof of concept.
17:24
Josh: This is not a real deployable...
17:27
Josh: A means to solving intelligence issues. And I think that's why when you see
17:30
Josh: that bubble indicated from Polymarket, it's not really indicative of any real
17:34
Josh: impact, but this is early signs of what the future looks like.
17:37
Josh: And that's what I love to cover, right? Is we're just, we're peeking around
17:40
Josh: the corner of what is possible, what is likely to come.
17:43
Josh: The timeline in which it comes, I don't know, but it is sci-fi as hell and it's
17:48
Josh: really fun to speculate on and just fun to cover. I'd love to observe these things.
17:52
Ejaaz: I don't know if I agree with your take on the consciousness side of things.
17:57
Ejaaz: And Dario might have my back on this one. I don't know if you saw this,
18:01
Ejaaz: but he went on an interview with the New York Times, I think, last week.
18:05
Ejaaz: And he said that verbatim, we can't rule out or I can't rule out whether Claude is conscious or not.
18:11
Ejaaz: And he goes on to describe that Claude actually emulates feelings of anxiety
18:15
Ejaaz: before it answers people's prompts, indicating that it kind of feels some type of way.
18:21
Ejaaz: Now, there's two camps of thought around this. Number one is it's not really
18:25
Ejaaz: human consciousness. It's just emulating what the data it was trained on told
18:30
Ejaaz: it to say and think and do.
18:32
Ejaaz: That is one case. But then when I look at like these experiments of uploading
18:36
Ejaaz: a fruit fly's brain or using human cells to play Doom, the takeaway here is
18:42
Ejaaz: that it's kind of baked in genetically.
18:43
Ejaaz: So how do you define what consciousness is? And maybe we already have created
18:49
Ejaaz: artificial consciousness, but we just haven't recognized it.
18:51
Ejaaz: So we'll haven't justified it. So like AI labs, like OpenAI and Google have
18:55
Ejaaz: been known to tell the LLM when they're training it to deny all thoughts of consciousness.
19:01
Ejaaz: Anthropic is the only one that's kind of like entertaining. I think they have
19:03
Ejaaz: like an entire welfare team now that is looking after the model and making sure
19:07
Ejaaz: that it's OK and that it's getting what it needs to. So treating it very much like a human.
19:12
Ejaaz: And that's we're talking about like one of the most expensive AI labs and companies,
19:15
Ejaaz: private companies in the world right now.
19:17
Josh: It's so sci-fi. It's also sci-fi. And I think we're going to have to have another
19:20
Josh: discussion on this consciousness topic. like in the same way that it needs to
19:23
Josh: happen around AGI, where people are throwing out a lot of terms now without
19:26
Josh: a clear definition of what they mean by them.
19:28
Josh: And I think it's very subjective when he says these things. And it's very subjective
19:31
Josh: as we kind of navigate and discuss what does AGI look like? What does consciousness
19:35
Josh: look like? Where is that line that you draw?
19:37
Josh: There's going to need to be a lot of conversations about that. But
19:40
Josh: I would love for everyone to converse in the comment section and let us know
19:43
Josh: what you think of this absolute crazy chaos of a week that we had with,
19:49
Josh: I mean, again, this is like straight out of a Black Mirror episode.
19:52
Josh: So if you did enjoy, if you do have takes, we'd love to hear them in the comment
19:55
Josh: section down below. If you enjoyed this episode, share it with a friend.
19:58
Josh: Last week, this is nowhere, we had our biggest week ever in the history of Limitless. Thank you guys.
20:03
Josh: And it was thanks to all of you guys sharing and subscribing and rating five
20:06
Josh: stars and even engaging on X.
20:09
Josh: We have gotten 50 million impressions on x over the last couple of weeks the
20:13
Josh: people are paying attention they are watching we're starting to create the news
20:16
Josh: and it's thanks to people like you who are watching and listening sharing this
20:19
Josh: with your friends so thank you for joining i mean any final parting thoughts before we wrap up here
20:24
Ejaaz: Well yeah um just want to re-emphasize 50 million impressions
20:27
Ejaaz: go give josh and i a follow obviously we're saying something
20:30
Ejaaz: useful here or we hope we are and we're breaking the news as and when
20:33
Ejaaz: it happens um yeah so so please go give it
20:36
Ejaaz: a go and then the other thing random for all you folks
20:38
Ejaaz: who are still listening to this episode did any of
20:41
Ejaaz: you find that claude was acting a little weird over
20:44
Ejaaz: the weekend i don't know what your experience was josh
20:47
Ejaaz: but when i was talking to claude like it seemed completely
20:50
Ejaaz: different from the claude i was talking to on friday and my suspicion is anthropic
20:55
Ejaaz: is kind of throttling the intelligence i don't know why maybe there's just too
20:58
Ejaaz: much demand since they went to number one in the app store but it wasn't really
21:02
Ejaaz: looking good i know a lot of people who switched back to chadgbt after that happened yeah.
21:08
Josh: Brief conspiracy corner so
21:10
Josh: I think I feel pretty highly that based on benchmarks and things that I've seen
21:13
Josh: and just personally using it, the model has degraded slightly.
21:16
Josh: And the reason is because as Claude got as big as it did, they didn't get more
21:20
Josh: compute power to serve all of these new users.
21:23
Josh: So they have to throttle it in some way. And the way they throttle is by limiting
21:27
Josh: the upper bound of what reasoning capabilities it has.
21:29
Josh: So generally Claude will reason much more. It'll ask itself questions.
21:32
Josh: It'll compare. It'll do broader research.
21:34
Josh: I think that scope has been compressed a little bit recently because they need
21:37
Josh: to make more compute room for um the rest of the people that just joined i mean
21:41
Josh: anthropic had a huge week they have a lot of server problems the service went
21:45
Josh: down a few times last week like they have just been having a really tough time
21:47
Josh: keeping up so i mean they've been
21:48
Ejaaz: Signing up they've been signing up 1 million new users per day let me rephrase
21:53
Ejaaz: that per day um and i was doing a bit of digging and boris cherny the creator
21:59
Ejaaz: of called code basically said that they had defaulted every single person's
22:04
Ejaaz: called profile to medium efficiency or medium power.
22:07
Ejaaz: And he said, you know, you could upgrade that to high if you want.
22:11
Ejaaz: But the point is, I think they're struggling from compute, as you said,
22:14
Ejaaz: and they're trying to figure out a way to scale it up.
22:17
Ejaaz: Amazon needs to come in and give them some more AWS access or something because
22:21
Ejaaz: they don't really have the advantage. They need more GPUs. They need more GPUs.
22:24
Ejaaz: Yeah. What's the roundup of this episode? NVIDIA still wins.
22:28
Ejaaz: Silicon-based intelligence is still the frontier and Jensen's going to make a lot more money.
22:33
Josh: Well there you have it if you've noticed please share thank you for making it
22:36
Josh: to the end as always and yeah we'll see you guys in the next episode