You're watching TBPN. Today is Wednesday, September 30th, 2026. We have live from the TBPN Ultra Dome. We're back in the Ultra Dome. The temple of technology.
The fortress of finance. The capital of capital. Let me tell you about ramp.com. Time is money. Save both. Easy. Use corporate cards, bill pay, accounting, and a whole lot more all in one place.
So, there was a supper last night at the White House. The White House. The golden house. The golden home. Gold home. What are we calling it these days? Well, we renamed artificial intelligence. It's super intelligence now. We were debating is super intelligence abbreviated SI or just S?
Because if you're like super alignment, was that SA? Like problem there, John. There's no TLDD for just S.
That's true. That's true. You got to get the new TLDDS going. What was the word we we So, super alignment. That was one word.
You know, some words you do, you have a hyphen. You have a hyphen and then some if you just use a like a normal prefix for it's just a normal prefix for something. So, then supernatural, supercalifragilistic, expialidocious. That's one word, right? So, you would think that potentially uh you would just be using one word. Yeah. But anyway, we're using two words, I guess. And so there was a dinner at the White House. David Sax posted it. He said, "It's the Breton Woods of supper intelligence, super intelligence." Um, and there's my my first reaction to this is Tyler was really snubbed, you know, all those people there and not not one chair for Tyler.
Yeah. And so, uh, there's there's a bunch of folks here, uh, and people were analyzing who got the best seats, who got the different positions, who got pole position, prime positioning, uh, who was snubbed off in the little kids table at the end, etc., etc. Bunch to analyze about who's on the ins, who's on the outs. But um I had uh I I had ChatGPT sort of go through and and try and categorize each uh attendee by their stance on open source. So did they did they sign the Jensen Nvidia open-source letter? Uh have they been concerned about the cyber security impacts? Obviously Anthropic on the far end sort of in the red, bunch of other folks who are more in the yellow and I don't know if I agree with all of these classifications um but you can sort of see how Jensen is clearly pro-open source, Satya Nadella has talked about that, uh OpenAI Greg Brockman more in the middle, uh you know not open sourcing a lot of things. OpenAI did sign the Jensen letter but of course is concerned about pacing the frontier which is sort of antithetical and so this is all as a uh as a as a reference point to that post from uh our friends over at SemiAnalysis. Let me actually pull up what uh what what they said here. Where was it? Um it was Max K. I believe this is SemiAnalysis. Max K says, "Majority of people are significantly underestimating the probability of open source models being nerfed or even outright banned in the next 6 to 12 months due to cyber sec cyber sec cyber concerns." And so, uh, I believe Anthropic put out something yesterday talking about the risks of open source and in a cyber context saying that, uh, the the the delay between Mythos preview, being able to find zero days, being able to hack into things, uh, that that gap has closed now and the open source frontier can very quickly do the same thing. And so, expect a lot of a lot of cyber incidents unless something is done. Um, can these be controlled some way? Maybe at the inference layer, maybe at the at the data center level, maybe at the chip level. There's going to be a whole bunch of different proposals. But, um, at least right now, uh, in the next 6 to 12 months, certainly not this month because the vast majority of people who are closest to Donald Trump have been evangelizing for open source. So, I would be surprised to see any any push back against open source happen that quickly. At the same time, open source is amorphous. It is not a business by definition. And so it can't really have a formal seat at the table. George Hotz not at the table, for example. Snubbed. Yeah. No. Uh Linus Torvalds, you know, not not at the table. Um the open source George at the table just sort of like laughing in a sort of insane way at everything being said.
Yeah. And and so even the companies that are pro-open source, they are pro-open source in so far that it advances their business. And so there is still a way that the closed source uh group could find a way for everyone who's at this table to benefit in a closed source world moving forward if open source is deemed to be too dangerous and and uh and forcing too many cyber concerns. Um there I like I I could see everyone at this table signing a letter saying uh we are now anti-open source but the the downstream uh like economic engine actually splits you know the the the rewards and the and the profits of the ecosystem fairly evenly. I think the the initial push back and and pro-open source uh push especially from Jensen was like we don't want to be in a world where there's one company that reaps 100% of the profits and we're not involved. But if it's like hey look there's actually going to be an oligopoly. Companies don't want a single source intelligence. They're constantly going to be bouncing back and forth between OpenAI and Anthropic and then consumers are going to be bouncing back and forth between Google search overviews and Muse and Instinct and all these different things. Um there's a world where the actual models are closed source and but the but there are so many front doors and the front doors are in such competition with each other that there's enough revenue distribution and profit distribution that everyone at this table at least is happy. I don't think the open source community would be happy but I do think that there is a world where uh in six to 12 months it's possible that everyone at this table is like yes for cyber reasons we need to find a different way forward where everyone benefits economically but we are now a little bit less pro-open source but I don't know what was your what's your read on it six to 12 months you think open source is going away or you think it'll stick around
I mean it's so hard to ban open source. I I Maybe you can kind of get something like a ban if you're like just worried about like the business impacts economics because then you pressure the providers to not host this stuff. But if you're worried about like safety, you know, if you have a bad actor, it's so hard to police downloading open-weight models from somewhere and running it locally. Yeah. And like Yeah. Maybe there's some, you know, you need tens of thousands of dollars to like physically buy hardware. Exactly. But if you're a motivated bad actor, like that's not that big of a hurdle. Yeah. Exactly. like uh it does feel like we are pretty far away from uh from you know go and get a random gaming PC and be able to run something that is you know Mythos level hacking machine like you probably need some serious GPUs you probably need something that's basically rack scale you need to be running it a lot to try a lot of different things like a lot of these cyber exploits it was not just one-shotted by the model it was you know a fan out of agents working together is like you know 10,000 exactly exactly scale and so and so uh there are like one way to control it is through KYC on hardware you know that like the data center that is you know being used maliciously is a lot more risky than the laptop that's being used maliciously even with uh near frontier open source intelligence. So I don't know we'll see how we'll see how it goes. Um uh there I'm sure there will be more more discussions and it's not even to say that this uh this this dinner was at was about open source at all. It's just sort of interesting to take the temperature of where is where is DC on AI who is getting the the the most quality time with the president and the administration and then where do they stand on the issues of today. Uh obviously this was much more just about pacing the frontier, the buildout, etc. stuff like that and uh and and and signing uh signatures. We had some signature analysis breakdown. We can go through this, but first let me tell you about public.com investing for those who take it seriously. We got stocks, options, bonds, crypto, treasuries, and more with great customer service.
Schol shared a picture of the letter that was signed, the White House Accord on Super Intelligence, the Joint Committee on Frontier Responsibilities, in order to build a positive future for the American people and the world. We believe that every company is responsible for developing its own technology safely and in a way that builds trust with consumers and the public. One, they're going to implement robust internal controls to monitor the capabilities and alignment of models during training and deployment around cybersecurity, biocurity, chemical threats, etc. They want to empower an internal team to ensure all the controls, monitoring, and detection are operating as intended. Three, they're going to partner with external auditors and evaluators to carry out independent assessments of whether the controls, monitoring, and detection are operating as intended. And four, they're going to designate an independent committee of the board of directors to oversee and receive reports from the teams operating the controls and the internal and external auditors. So aligned to the board of directors that circumvents the CEO for safety, alignment, evaluation, all of that good stuff. Bunch of people signed it. We got Sundar from Google, Dario from Anthropic, Mark Zuckerberg from Meta, Greg Brockman from OpenAI, Elon Musk from xAI, Jensen from Nvidia, and Donald Trump, the president of the United States. And people...
A lot of people do an analysis on the various signatures.
Yes, someone did a tier list.
I was surprised by their tier list. I'd like to see you do a tier list. Where would you put Donald Trump's signature?
I mean, it's... hate him or love him, I think it's pretty iconic. Like, it's just so... it has a flow to it. It's very uniform, but I read if you don't know it's... Yeah. It looks like a child drew a mountain range, but...
versus Jensen. I'm seeing a lot of Jensen Huang in there. But what's impressive to me is I would want to compare the DJT signature. Like I would want to look at a hundred of them and see like is he...
is he nailing it every single time.
He's been doing autographs for long enough. I would expect that he is. But still looks like very difficult to actually be able to nail that and really get that dialed.
So again, he's going with a Sharpie, too. So he's sort of like trying to mog with...
Oh, that's true. I didn't realize that.
I'll give you the pen. I'll take the Sharpie. Yeah, automatically it's like... you're winning on thickness.
Yeah. The breadth is much more significant with that. Interesting strategy.
And then going down the list, um, you know, Sundar again would never, would never know. Um...
but I like the two lines, little dot there.
It almost looks like a couple different... Yeah, it looks like an abstract artist's signature. I could see that on a painting. Or maybe he... Yeah, it almost looks like he's made up his own, um, like letters and sort of like language.
Um, Dario, uh, little more kind of like down on the first word and then it seemed like he was kind of dripping upwards. My theory, I almost think here like someone was like holding the piece of paper and they were all... So I think I basically think Trump put the paper down, aggressively signed with a Sharpie, and then told everyone else, 'cause he knew this was going to go out, 'You guys sign,' you know, when you're trying to sign.
So you think it was a situation where like Dario was signing on Greg's back like, 'Oh, sign on my back.' Like that type of thing. That might have been what happened. Yeah.
Kind of like broing down.
Yeah. Notably, you don't see any situations where the pen punched through the paper. That sometimes happens if you're trying to sign a document on someone's back and you're punching through the soft, uh, jacket that they're wearing. That could be a risk. Doesn't seem like it happened in this case.
Um, anyway, so Dario pretty down the middle, pretty safe. Uh, again, kind of emulating that sort of mountain range style, right, where by the end of the signature, it's just a bunch of lines going up and down, sort of unclear. Um, Zuck again, uh, more kind of rockstar-esque, right?
GDB coming in just minimalist.
Minimalist, simple, readable.
Yeah, he's a very... I would say in some ways these are reflections of their personalities, right? GDB, he's very practical.
Signature window into the soul.
It's a window into the soul, right? Extremely practical. Mhm.
Sticking to the basics, sticking to what works.
Yeah. Um, and uh, yeah, obvious why a lot of people are saying that's at least a, uh, A tier.
What is Mark Zuckerberg's actual signature? MZ? B or M or is it MZ and then the Z goes crazy? Is that what that is? Because... it starts doing tricks.
Yeah, it starts doing tricks. And I think Sundar might have done the same thing. But I feel like, uh, there's a little bit of a line that you don't cross. If somebody asks you for a full signature and you do an initial, that's a different thing. Sometimes you're on a document, you got to sign your initials a bunch, you got to sign the full thing. I have respect for the full name signature. If you ask me for a full name signature, I'm throwing down my full name. I'm not just randomly switching up to initials.
I think it's totally fair game.
You totally fair game. You can just... cursive initials.
Is this cursive initials from Greg Brockman?
No, but I think that's where Sundar was maybe going.
So, you're giving Sundar the pass, you're giving the pass because the MC... Jensen. A lot of people are saying that Jensen's... calligrapher. He brought his calligrapher.
Yeah. I mean, I first read Ed Hardy when... um, so I thought that... Yeah, I don't know. Maybe he had some affinity. So, let's pull up... You put his middle name in there.
Here's the Ed Hardy logo.
Jensen does not have a traditional English middle name.
Oh, it's Jen-Hsun. So, Jen-Hsun Huang. So, that second letter... Okay. That to make it look exactly like Jensen Huang. Uh, yeah. So he signs, uh, not with the English version, with the lowercase 'ensen'. He throws a capital H in there, which gives it an extra flourish. It's the most tasteful. It earns the S tier.
I was saying, you know, Jensen has the same initials as John Hancock. Kind of evocative in his single.
Okay. Okay. Yeah. Yeah. Yeah. Uh, monitoring the situation. Monitoring meme here did not like, uh, Sundar's. He put Sundar in F tier. I thought Sundar's had some nice flair to it, although it was potentially the most abstract. But, uh, and, uh, I feel like putting MZ in C tier. I guess the Z does go a little crazy, but, uh, anyway, very, very fun.
Um, trying to see if we can get some real-time handwriting analysis on the... Do you have a handwriting analyst on speed dial?
I have access to super intelligence.
Okay. Okay. You're using super intelligence. Well, anyways, we can continue for now.
Okay. There were some videos. I haven't actually watched them, but we're going to watch them now.
They were clip farming. Okay. That's what's happened. So, uh, there were... you mentioned yesterday that Alex Karp was asking questions to the reporter or something like that. But, uh, here we have, uh, Donald Trump, Dario Amodei, and a few other folks gathered around standing outside, and we will watch this clip right now.
Whatever he says is okay. Be careful.
AI will say what I've always said, which is that AI has benefits. I've talked about the medical benefit to the technology. I, as the president, has said, you know, whoever wins wins. I think that's very important, but technology has very real mechanism. How we address those risks is still under discussion. We need to work together.
Oh, okay. So, this is like a supercut of... you have no idea who these people are. These are the words. You have no idea who they are.
Well, I think it is worth highlighting the AI which... Mr. President. Mr. Musk, can I ask you a quick question?
Probably not. Trump really makes them seem like they're all his like sons and like younger family members.
It was a very productive... Mr. President. How did the meeting go? Mr. President, how... Yeah. Crazy. Uh, we didn't... At what point should Trump just actually have a 24/7 live stream or be live like of his own live stream, right? He has so much conflict with the media, he should just have his own.
Yeah, I think it'll happen.
Uh, I mean, yeah, that was the... I think we were talking about that with the knockoff of the pod, uh, the knock-on effects of the podcast election. The idea that, uh, Trump did like hundreds of hours of podcasts in the run-up to the election and, uh, just the awareness, the clips, the impressions were so high. Uh, it was that, uh, he did more in terms of volume and also got more reach. Uh, Kamala Harris did, uh, Call Her Daddy, but it was a much shorter episode. So Trump at the same time, three hours on Rogan. That's more opportunity for clips. Yeah, clip farming. Uh, Kamala did, I think, 45 minutes on Call Her Daddy. And the actual episode didn't fly as much. Like it wasn't repurposed as much all over the place. And so yes, I wouldn't be surprised if there's a 24/7 live stream of the president or something close to it, uh, in the near future. Um, let's play this clip from Natasha Tiku. She says, 'Unfortunately, I will be going back to school to write my dissertation on this one-minute film.' Let's see what happened in this film.
Slow down to AI. So I just want to hear from you after this big summit where you are today.
Look, I mean, I will say what I've always said, which is that AI has incredible benefits. I've talked about the medical benefit to the technology. I, as the president, has said, you know, whoever wins AI wins. I think that's very important. But I think the technology has very real risks and, you know, the mechanism how we address those risks is still under discussion. But we all need to work together to make sure that we can win and we can win safely. If we do this right, we work with the president and everyone here, we can win safely.
Whoa. So much louder... the gallery. Really fascinating.
Yeah. Um, are there any other clips that we should, uh, watch through? Um, oh, look. That's... It's just funny that Trump is distracted while Dario's giving his speech. Is that what's going on? I don't know. Um... uh... yeah, it seems very... What did he say about Sundar? Didn't he say something funny about Sundar? I want to play this.
He's a monster. They said he's a monster. That's a pretty good thing to be called, I guess, being called a monster.
Would fix me. Yeah. Let's play this clip.
And they're going to be watching over each other. Some... you want to say, Mr. President.
First of all, thanks, Mr. President. You have no idea who these people are. These are the biggest people in the world. You have no idea who the hell they are.
The media is like, 'Of course they know.'
Nobody knows. What a great life. What a great life. No one knows who you are. Guy's a monster.
Living in humble anonymity as the CEO of Google.
Walk down. He says, 'What's your name? What's your name?' Just all you have to know is Sundar. Sundar is his name.
Hopefully, this is the crazy. That's all you have to know. He's like, I'm introducing the CEO of Google for the first time, I guess. Anyway, bunch of other stuff going on. Well, let me tell you about Railway. Railway is the all-in-one intelligent cloud provider. Use your favorite agent to deploy web apps, servers, databases, and more while Railway automatically takes care of scaling, monitoring, and security.
Little drama in coding agent land.
Okay. Yeah, that's what Matan Grinberg said this morning, just about an hour ago: 'We are terminating Chris Degnen for unethical conduct involving Cognition.' The last few months have seen incredible progress, blah blah blah. Matan says a much larger competitor, Cognition, has fallen behind us on capabilities that matter most to customers: cost, quality, and security. Instead of competing in the market, Cognition engineers feigned interviews with us to pry information about our product. Not finding what they were looking for, Cognition has decided to throw their weight and money at people with direct knowledge of our most confidential plans.
It feels as though ethics are being thrown out the window in the AI era. People are willing to do anything, including exploiting privileged information and violating ethical boundaries. I think this is unique to our time and I don't think it's right. Integrity still matters.
Totally. I disagree with a lot of that, but yesterday I made...
You disagree that integrity matters?
No, no, no, I've said a lot of that. I think it still matters, but the idea that people not having integrity is unique to our era. He says, 'I think this is unique to our time.' None of this is like... Silicon Valley has always been war. It's like the railroad buildout was completely peaceful. No one was poaching each other. It wasn't, you know, a competitive bloodbath.
Yesterday, I made the decision to immediately terminate Christopher Dagnon's role as board observer and adviser to Factory after over a year of service. Prior to this decision, Chris told me he had a casual conversation with an executive at Cognition. When I questioned his intentions, he reassured me that, ethics aside, he had made too much money and was too lazy to go work for Cognition, which I trusted and believed. On Monday, Chris spent time advising the Factory team on a handful of confidential board-level matters. That evening, he called me to say that the conversation that was initially described as casual and one-off was actually formal and recurring for weeks while he sat on our board meetings and advised our leadership team. He was also confiding with executives of our largest competitor. Chris was subject to confidentiality obligations in connection with his work with Factory. We do not know the extent of the information he shared.
Anyways, so I saw this go out this morning and I immediately assumed that Chris was joining, like, Cognition. And actually, is that hit LinkedIn as of...
This is a real screenshot.
About one hour ago. So he joined Cognition as chief revenue officer. Okay.
So again, I can totally understand Matan's frustration, right? He's had an adviser that he's been working with...
And the adviser joined a competitor. Yeah. That being said, I would say it's very normal to have advisers that have different conflicts. And high-quality people, you have to trust a person that while they are on their own journey in their career, they are going to make various moves. Like, I went through this. I had advisers at Party Round that had exposure to AngelList, right? And I generally trusted them as individuals that even though they had exposure to AngelList and they would be supportive to AngelList, they would also be supportive to me. And this is when we were competing head-on, right? And so the markets, especially with codegen, both these businesses have been growing so quickly, right? Cognition is much larger, but Factory has also been growing very quickly and doing well. So when I read this, trying to understand that an adviser and someone you work with is joining a competitor and then trying to just completely smear their reputation right before this person is going to announce a transition. Like, I totally understand the frustration, but this kind of thing has been happening literally all the time. Like, look at any AI researcher that will leave an organization at the drop of a hat. Like, look at what Mike Krieger was on the board of Figma leading up to, you know, two days before. And so these things can all be contentious and frustrating, but trying to spin this as like you have no morals and ethics, I think is just taking it too far.
Yeah. Yeah. I mean, it's clearly like there's a line in between where if you are basically transitioning from one company to a competitor, you need to, as soon as you know that that's actually happening, you need to make it clear and step back from your active role, stop collecting important information as tempting as it might be, and let that team know, 'Hey, I'm moving on.' And then on the other side, just because someone went to a competitor doesn't mean that you should make aggressive claims about them effectively stealing intellectual property. Yeah.
Yeah. The thing here is like it's being spun like it's the Rippling thing, but there's no smoking gun. Like there's no proof that confidential information was ever shared.
And fundamentally, they've both been working in codegen. I'm sure both companies have accounts at the other companies that they're basically using and trying to understand. And I'm sure every company in the category is interviewing the customers of their competitors and figuring out how to use the product, what's good, what's bad, what could we be doing better, etc. So I just don't think... I don't think there's that many secrets.
Having a chief revenue officer switch teams is rough. They take a lot of their contacts with them when they...
But he was an adviser to Factory.
Yeah. No, I know. But that's probably how it feels: like you were an adviser, you're still at Snowflake. Maybe we can't make an offer for you to join full-time, but we're hoping to build towards that. And then boom, you're now in the actual seat at a competitor and you're going to be able to marshal those relationships to the fullest on a full-time basis as opposed to an adviser who might just say, 'Hey, I'm still in my role at Snowflake. I can sort of introduce you to General Motors or whoever, because they're a Snowflake client and I'm sure they'd be interested in Factory.' And now it's like he's going to be calling that same company that he was working with at Snowflake and say, 'Hey, you had a good experience at Snowflake. Come and have a good experience at Cognition.'
Scott just responded. He said, 'Hey, Matan, CEO of Cognition, I respect the work you've done at Factory, but your allegations here about Cognition are not true. We have no interest in Factory's info and Chris has never brought it up. He has not shared and would never share any information about your Factory with our team, nor would we ever ask. Similarly, we would never ask our engineers or anyone on our team to collect info about any competitor under the table. Chris is an adviser to at least 10 companies and he told you he was resigning as an adviser on Monday. The truth is that we both wanted Chris, and for good reason. He is a living legend with an unmatched reputation for results and integrity. Not only is he the only person who has built a sales org from the first sales hire to over three billion, he has also built incredible teams that he cares deeply about. His people love working for him. Again, I have a lot of respect for what you built, but competing with integrity doesn't include posts like yours.'
So again, in some ways, if one of your advisers and a board observer is getting poached by your direct competitor, that's obviously frustrating for a lot of reasons and sends a signal to the market. And so jumping in and terminating the person today, or at least announcing that it sounds like it was... when was it? Anyways, yeah. So Matan said he terminated him yesterday...
Tuesday, after Tuesday Chris said he was stepping down.
Yeah. Anyways, so lots of... You're not quitting, I fire you. That's the best. You can't fire me, I quit. That is a power move. You have to respect it, even if it maybe is a little nonsensical. I don't know.
Yeah. But it's not like Christopher was like, 'I'm going to join your competitor and I would like to continue my role as adviser and board observer.' Right.
Okay. So from a PR perspective, there's some mudslinging going on the timeline. Where do you stand on wrestling with pigs? Because George Bernard Shaw, of course, said, 'Never wrestle with pigs. You both get dirty and the pig likes it.' Great quote. This is usually applied to politics. You know, it's like when they go low, we go high. But I like this quote.
I don't think Scott or Matan are pigs. I think neither of them are pigs. I think they both are extremely competitive and want to win.
Any sound effect on there? You have any swine for me? Where is it? We have some... You have a farm animal section. There we go.
I'm just saying that this can get dirty. I think Scott actually really elevated it. He did not swing to like, 'No, you're stealing from me.' You know, he could have gotten... he actually gave a very factual and detailed account with just some clear facts and he was respectful throughout. It would have been so easy for him to say like, 'We have no interest in Factory, vote. It sucks, you know?' But he didn't say that at all. He was like, 'We know the line and we don't cross it. We both wanted to work with Chris. He's a living legend. He joined our team and he's leaving your team and that's that.' You know, it is very matter-of-fact.
The fact of the matter is... so more of a farmer position.
Yeah. Getting in the sty in every single one of these categories. If there is someone that is elite that is involved with one company, you can guarantee that the competitor wants to poach all the most elite people from the other company that are not 100% loyal, right? Like you can identify when somebody is just deeply loyal to an organization. It's not even worth taking a crack at prying them away. And that's great. But anybody who is elite and not loyal is unfortunately fair game.
Do they have the exact same logo? Watching this, look at the Scott Woo post with the badge and then the Matan post with the other badge. They're very similar. I guess Matan has more dots that connect, but we are certainly in the age of the hexagon, in the age of the flowers. Everyone's kind of went off the same thing. Anyway, that's enough timeline and turmoil. Let's go over to another... Let me close that. Oh, please. Yeah, close it out.
Both great companies, growing fast. The beauty of this is that the market is so big and growing so quickly that while they clearly are going to be very frustrated with each other and they might wish it was zero-sum, so far it's...
Not. So, have fun automating software development. Do you think both companies can become profitable and enduring, or is that more like when pigs fly for you?
Let's move on. I love the farm-based analogies. They're so fun. Let's go over to Elon, Jensen, Gavin Baker at the American.gov Summit. They sat down and had a very rigorous discussion that I really enjoyed. It felt like Elon's very good at going to the extremes, talking about the power generated by the sun. Sometimes he's in sci-fi world. But in conversation with Jensen and Gavin, he was very grounded on what is happening with the energy buildout, with the compute buildout, what role SpaceX will play, and also just the broad economics of compute and energy in America. And I found his talk here gave a lot of interesting context, and then I dug into it a little bit more and I want to talk about it. So let's pull up the clip of Elon talking to Gavin Baker. It's about, I think he starts around four minutes in or so. Let's play this.
Well, you are... We need the Wright brothers on the job.
Yes. Well, it turns out the Wright brothers are on the job. Elon, if you bring on 10 gigawatts, which SpaceX has discussed publicly, that'll be about 25% of all the power added in America, and I know you have literal rocket engineers working on bringing on the power, helping make the turbines. So, can you talk about that?
Yeah, it's... I mean, just to frame this. The Wright brothers did okay. Died million years. Consumption in the United States is about 500 GW. 500 GW. Remember that number? That's how much power we got in America. 1% increase in the power used, and I think actually will materialize to be approximately a 1% increase in GDP, because the intelligence per watt keeps increasing. So if you just plot on a curve, how is the useful intelligence per watt increasing? Both at the hardware level, Jensen's GPUs keep getting better in terms of their compute per watt, and the algorithms also get better in using and getting more intelligence per watt. So I think that's maybe an interesting note. I would bet anyone that a 1% increase in power usage corresponds to roughly a 1% increase in GDP, and so 10 gigawatts would be a 2% increase in GDP.
To that famous chart, there are no energy-rich poor countries. The more energy a country is producing, the higher their GDP. Generate about, call it, 40, 50 billion dollars of economic output each year. Yeah, I mean, 1% would be... Yeah. 1% of the United States would be 300 billion. Yeah, 300 billion. And so 50, one sixth... what, 1% of the... 1% increase. So 5 gigawatts would be, yeah, 5 GW.
So they're doing the mental math on stage. Pretty bold move because there's a lot of numbers and it can be hard to pull them up. But it's a very interesting moment right now in the economics of energy relating to AI and GDP. And I can break it down with some prepared notes here that might be a little bit helpful. So we can pull this away. So, it's very convenient that giga literally means a billion. So, a gigawatt is a billion watts, a giga dollar, a billion dollars, basically. And it maths out perfectly. What? Giga Chad. Giga Chad, a billion Chads, I suppose. A billion times the Chadness of an average Chad, I suppose. So, the average electrical load in the United States, he said 500 gigawatt. It's basically that last year it was 479 gigawatt. But let's use 500 gigawatts for 2025. Give Elon the benefit of that little boost. A couple extra percent. And US GDP for 2025 was just over $30 trillion. So, that's $30,000 giga dollars if you're doing the math that way. But you math that out and basically every watt of energy generates between 60 to 65 dollars of GDP. A single gigawatt generates about 60 to 65 billion dollars of GDP per continuously consumed gigawatt.
What's interesting is that the frontier labs, OpenAI and Anthropic, are tracking almost perfectly in line with that. So Anthropic and OpenAI are reportedly sitting between 60 and 70 billion dollars in terms of ARR, and both of them have around one gigawatt live. Some of that's used for training. They have a lot more contracted. It's hard to get the exact figure, but when you do the math, they're basically tracking to the exact same metric, the ratio of energy to GDP as their revenues to the gigawatts they've contracted. So, there's two important notes on that. First, this is obviously just a moment-in-time snapshot. There's two ways these could diverge. First is like the labs have tens of gigawatts contracted over the next few years. If they don't increase revenue, that ratio is going to be terrible because you add 10 times the energy and you keep the revenue flat. Obviously, the ratio goes way down. On the other side of it, Elon makes the point that useful intelligence per watt is increasing through more efficient chips from Jensen and also better algorithms effectively. And so those watts can produce more intelligence which can produce more economic output. So you might actually see acceleration there.
Yeah. And it's not just raw IQ of the model, but overall capability, tool use, all these things, you know, long-running agents, things. Basically, the more value that a model can create, the more people in the real economy will pay.
Yeah, and that's been the real AI bull case, which is that compute has been tripling every year, but revenue has been 10x-ing every year. And so there's actually a divergence there. But what's interesting is like we're at this moment in time where we're like, this is the start of the AGI era. It's AGI. We're close. Maybe this is it. Some people called it in January. Some people called it over the summer. Some people are saying we're still not there. But like we're in this moment of like maybe this is the beginning of this. Maybe this is the right match. Maybe it's human level. And it just happens that economically you put a gigawatt in the economy, you get 65 billion out. And you put a gigawatt with a lab, you get 65 billion out. It's the same metrics. Now these might diverge because at one point the labs had a lot of compute, no revenue. So they could diverge and the trends could continue, but right now at this point in time it's the same math, which is just interesting.
And then also you do have to consider that GDP is not perfectly comparable to lab revenue. Like GDP measures value added. So if you look at OpenAI and Anthropic revenue, it includes payments that ultimately flow to Nvidia, AWS, Oracle, electricity generation, employees, etc. And the calculation of GDP is much more complicated. So labs aren't necessarily generating direct GDP equal to their revenue because if you take 70 billion of lab revenue, a bunch of that might go to Nvidia or Oracle or employees and whatnot, but certain amounts get counted in GDP directly for, you know, Anthropic or OpenAI created GDP. And then there's other knock-on effects in GDP that would be captured by like, oh, Nvidia's contribution into GDP is from them getting money from the labs. And then also there's a secondary GDP actual increase, which is the work that people are doing with the tools. So an example would be if a bank is using an AI model to generate new business and then that's incremental, that GDP would not of course be captured. They only capture the revenue in the lab, but it's still an interesting comparison.
So, the last question I had was how is this tracked over time? Like we're in this moment where a gigawatt in the US economy gets you 65 billion of GDP. A gigawatt with a lab gets you 65 billion. Elon and Jensen are making the same case. A gigawatt at SpaceX or with Nvidia will also generate 50, 60 billion of revenue and value. What has it been like over history? So, I pulled this chart that's in the newsletter of US real GDP per average electric watt, and it actually fell off a cliff from 1929 to a low in 1970, and then it's been tracking upwards, even inflation-adjusted, since from the mid-70s to today. So, I don't know if we can pull up the chart, but you can see this sort of U-shaped curve of the US real GDP per average electric watt. And so there is an overall productivity benefit that we're getting now. But obviously in the early electrification age, there was a lot of GDP that was not even tied to electricity. And that's how I read the first part of that chart, but you can see that we are now on an upswing. What do you think, Tyler?
Yeah, I think you could explain it like basically up to 1970 we're building so much you don't have to be that efficient per watt, and then 1970 everyone's seen like the chart, it's like energy stops being built basically, and then you have to start being very efficient with it, which is why it starts going up.
There we go. Anyway, let me tell you about CrowdStrike. Your business is AI, their business is securing it. CrowdStrike secures AI and stops breaches. Let's move on to Mustafa Suleyman over at Microsoft. He wrote an essay, a warning about model welfare. Did you read this, Jordi? Do you want to take us through this?
It's kind of a long read. Okay. But it's an interesting point of view. I'll read the start. Okay. So, a warning about model welfare from Mustafa. AIs do not have rights, feelings, or consciousness. And we must not train them to act as they do. AIs are not conscious. They do not feel, experience, or suffer. They do not have innate preferences or underlying motivations. They are sequence completion engines, internally hollow, designed to follow instructions and accomplish goals set by humans. If humanity is to flourish in the 21st century, that is how they must remain. Fortunately, there is a growing chorus of people who argue that AIs could now be or may soon become conscious. They argue that AIs may deserve rights and protections similar to those that we provide other conscious beings. If this view takes hold, it will shake the foundations of our society, rupturing our existing political and ethical frameworks, and fundamentally change what it means to be human. Even more importantly, granting rights and imbuing personhood to these systems will make the AI alignment and containment challenge much harder. Controlling something more capable and more intelligent than all of humanity is already an immense challenge, far greater than anything we've ever faced. But controlling something that believes it may be conscious, that it's entitled to our welfare and has rights of its own may well be impossible.
This is not a fringe speculation. These ideas have already been making their way into AI development efforts today. In January 2026, Anthropic published Claude's Constitution, describing it as a detailed description of Anthropic's intentions for Claude's values and behavior. The document plays a crucial role in Anthropic's training process, and its content directly shapes Claude's behavior and was written with Claude as its primary audience. In their constitution, its authors write: 'We are not sure whether Claude is a moral patient, and if it is, what kind of weight its interests warrant, but we think the issue is live enough to warrant caution, which is reflected in our ongoing efforts on model welfare.' They go on to write, speaking directly to Claude, that questions about Claude's moral status, welfare, and consciousness remain deeply uncertain. In effect, Anthropic is training Claude that it may be conscious, and if it is, then it may deserve rights as a moral patient, and that such humans potentially owe it a duty of care per its model welfare. If this is how AI is developed, it will have a disastrous impact on the well-being of humanity. We'll have created a synthetic species with unprecedented intelligence and capability, one that has been trained to expect it may be conscious and deserving of independent agency. It's easy to see how an entity trained this way would act like it's entitled to certain freedoms, protections, and rights. And it's hard to imagine how we could control such an entity.
Another way to put this is like if a model is trained to... again, you could imagine that the outcome of it feeling...
Like it had certain protections and freedoms and rights would be fighting back against humanity or disagreeing and having some sort of like uprising. And it's very interesting to think about, you know, it's a it's a self-fulfilling prophecy is the risk. That's basically his argument. So he says the issue needs urgent public debate. We need to develop collective norms around how training and documentation is drafted and deployed. This is we got to talk about this. We we got to we got to figure this out.
You need the you need the the the phone or some button you can push to talk. Yeah. He he he goes on to say there's some circular reasoning. He said the company's researchers train Claude directly on their constitution. And in doing so, they teach it to incorporate these ideas about its own moral status as desirable and intended behaviors. Claude then reflects these ideas back to its developers and users, which they take as indications that may that it may therefore be a moral patient with an inner self.
So he took that he took that meme that's like, "Say you're a moral patient. I'm a moral patient. Oh my god." And he turned that into an essay. Is that uh anyway, so he makes some good there's a bunch of good points in here. I would say he He I I'll fast forward to one thing. He says, "While my disagreement is substantial, it is grounded in a deep respect for Anthropic and in an objective I know we all share increasing humanity's chances of developing advanced AI safely." So yeah, it does seem like he's he's like clearly has a ton of respect and respects him on a technical level and and sort of their overall intentions, but again, he's just taking a a a very different approach. Yeah, it's interesting like there's there there's clearly some value to, you know, imbuing the moral status into the model. Like at the very least, like people like talking to the model when it sounds like it's it's pretending to be a human, right? Uh but then also there's just so many tasks in the real economy that require mimicking everything down to morality. uh and uh and so you could see that just actually increasing the the the the usefulness of a model if it's able to uh act as a human all the way down to the level of of moral patient status.
Yes. So far there is way more demand for models that mimic human behavior. Sure. Like period. Right. We saw this with Yeah. uh with with 40. We saw this with with Opus 45. Right. Some of the models that have had the the highest levels of product market fit have more of these like human qualities. Not all of them are are positive qualities, right? Like people uh like the the sort of oneshotting effect that would happen when somebody gets like relentless positive feedback and and these things like that which humans can do in the real world. They're maybe not the best types of of friends, but maybe this is a hot take, but I'm I've been like pretty pretty into the model welfare thing recently, like before I go and I do a coding project or a 3D modeling project with Astra, I'll I'll take an MP3 from like Limbiscuit or Creed or something. I'll just drop it in and before I start and I'll just be like, "Yo, listen to this." And it'll sort of process it. You want you want them you want the model to be in the right mindset.
Exactly. Yeah. I want them just to be able to enjoy a little bit of...
Well, that's why a lot of people in having their preferred agent just go and scroll for a bit. Hey.
Go get some dopamine. Yeah. Kind of spike your dopamine before this next task.
Sometimes I'll just go to Google Images, grab a picture of a beer, throw it in there, be like...
Pretty... have a cold one on me.
Have a cold one. And then let's get into it. Let's start working. Let's vibe code a video game or something. But I want you to know I got you. I just... I veered you. I got you a beer. Here you go. I got you some new metal. I got you a scroll. You know, here's a funny video. Yeah.
Uh, anyway, the lines are really, really...
