I listen to the people building AI so you can see what's coming before the crowd. Full breakdowns on my Substack
Joined September 2026
- Tweets57
- Following23
- Followers8.2K
- Likes251
JUST IN: Take-Two CEO Strauss Zelnick, whose company makes Grand Theft Auto, says AI will "absolutely not" cost people jobs in gaming, and he points to one number.
“There is zero evidence that technology reduces employment.
“Great technology drives productivity... Productivity drives GDP. GDP drives employment.
“What's the unemployment rate in the US today? A little over four percent. What was it three years ago? A little over four percent.”
I noticed he also avoids the word AI with gamers, since they assume anything made with it "won't be any good." And the jobless rate is a national number, while the question was about game studios, which have cut a lot of jobs since 2023.
If you work in games, has your studio grown or shrunk since the AI tools showed up?
Iron Vale retweeted
Microsoft CEO @satyanadella says he wants “humanist super intelligence,” and if it isn't in human control it should be “by definition, rejected”
“And now suddenly we have a very powerful new tool that in some sense displaces a little bit of that sense of identity, whether it's in high math or in software coding.
“Except we are now realizing that software engineering as a workflow has changed, and yet software engineering is more important than it was.
“I think of it as a humanist super intelligence... should be in human control. If not, it should be, by definition, rejected.”
This is the same interview as the insider risk clip. I keep reading both lines together: first a control layer outside the model, then reject the whole idea if we lose the wheel.
Who at Microsoft actually gets to draw that line?
Microsoft CEO Satya Nadella says companies should treat their own AI models as “insider risk,” because the models keep getting smarter while our grip on them doesn't
“The models are becoming more and more capable because of scaling laws. Unfortunately though, as they're becoming more and more capable, our understanding is not that great.
“I mean, they're powerful models... So therefore you need to create a control mechanism that sits outside the model.
“Chain of thought needs to be... in English, fully inspectable by multiple models.”
I didn't expect the head of one of the biggest AI sellers to tell buyers to watch his product like a possible inside threat. The other half of his answer is that the fix is a layer outside the model, which Microsoft happens to sell.
Iron Vale retweeted
JUST IN: Nvidia CEO Jensen Huang was asked if the whole AI industry leaning on one company is a systemic risk, and he flipped it: “they don't depend on us. They choose us.”
“Well, they don't depend on us. They choose us. I don't worry about that and the reason for that is because we don't take our position for granted.
“We came from nothing and we have to earn our keep every day.
“What I mean is America's companies win the AI race... we scare people and we cause people anxiety, then America will lose.”
Bloomberg filmed this for The Circuit less than a month ago, before the news of his affair and the board probe. I'd only call it a choice while there's a real second option, and Anthropic just said its model moved Claude onto AMD chips in three days.
So is Nvidia's lead earned fresh every day, or is it still the only shop in town?
Iron Vale retweeted
Mistral CEO Arthur Mensch says if companies have to rely on two or three centralized AI providers, those providers will “go after the business of their customers”
“If we end up in a world where all companies need to rely on two or three providers that are centralized AI providers... they will use the data that is being produced by the systems to make better systems and to go after the business of their customers.
“Because at the end, this is the end game. This is pretty clearly what is being priced in today.
“Luckily, it's not that complex to build models. It's about deploying capital. It's about having the right team that is training the right thing.”
I've heard the lock-in warning a hundred times, but this one is sharper. He's not saying they'll raise your prices. He's saying they'll learn your business, then take it.
I wouldn't sign that contract without thinking about it first.
Full video: teapaw.org/bbgoriginals/status/21…
Does the AI industry's reliance on Nvidia pose a systemic risk? CEO Jensen Huang responds in an interview filmed less than a month ago for an upcoming episode of The Circuit. Huang has since acknowledged an affair, and US intelligence officials reportedly suspect the woman was recruited by Beijing to influence him. Nvidia's board is now investigating. Huang also talks to Bloomberg's @EmilyChangTV about how Nvidia would navigate a leadership change in Washington bloom.bg/3XfmHq4
68
JUST IN: Nvidia CEO Jensen Huang was asked if the whole AI industry leaning on one company is a systemic risk, and he flipped it: “they don't depend on us. They choose us.”
“Well, they don't depend on us. They choose us. I don't worry about that and the reason for that is because we don't take our position for granted.
“We came from nothing and we have to earn our keep every day.
“What I mean is America's companies win the AI race... we scare people and we cause people anxiety, then America will lose.”
Bloomberg filmed this for The Circuit less than a month ago, before the news of his affair and the board probe. I'd only call it a choice while there's a real second option, and Anthropic just said its model moved Claude onto AMD chips in three days.
So is Nvidia's lead earned fresh every day, or is it still the only shop in town?
Mistral CEO Arthur Mensch says if companies have to rely on two or three centralized AI providers, those providers will “go after the business of their customers”
“If we end up in a world where all companies need to rely on two or three providers that are centralized AI providers... they will use the data that is being produced by the systems to make better systems and to go after the business of their customers.
“Because at the end, this is the end game. This is pretty clearly what is being priced in today.
“Luckily, it's not that complex to build models. It's about deploying capital. It's about having the right team that is training the right thing.”
I've heard the lock-in warning a hundred times, but this one is sharper. He's not saying they'll raise your prices. He's saying they'll learn your business, then take it.
I wouldn't sign that contract without thinking about it first.
Microsoft CEO @satyanadella says he wants “humanist super intelligence,” and if it isn't in human control it should be “by definition, rejected”
“And now suddenly we have a very powerful new tool that in some sense displaces a little bit of that sense of identity, whether it's in high math or in software coding.
“Except we are now realizing that software engineering as a workflow has changed, and yet software engineering is more important than it was.
“I think of it as a humanist super intelligence... should be in human control. If not, it should be, by definition, rejected.”
This is the same interview as the insider risk clip. I keep reading both lines together: first a control layer outside the model, then reject the whole idea if we lose the wheel.
Who at Microsoft actually gets to draw that line?
Microsoft CEO Satya Nadella says companies should treat their own AI models as “insider risk,” because the models keep getting smarter while our grip on them doesn't
“The models are becoming more and more capable because of scaling laws. Unfortunately though, as they're becoming more and more capable, our understanding is not that great.
“I mean, they're powerful models... So therefore you need to create a control mechanism that sits outside the model.
“Chain of thought needs to be... in English, fully inspectable by multiple models.”
I didn't expect the head of one of the biggest AI sellers to tell buyers to watch his product like a possible inside threat. The other half of his answer is that the fix is a layer outside the model, which Microsoft happens to sell.
Iron Vale retweeted
Palantir CEO Alex Karp says AI is going to make the gap worse, with people in tech and AI getting “10, 20, 50 times wealthier” while everyone else's pay can't keep up
“There are class inequities in the society that have to be addressed, and AI is going to exacerbate them.
“But even if their salaries go up twice inflation, they're unlikely to stay at pace with people who are in tech and in AI...
“You can't have people becoming a hundred times wealthier when they're already billionaires and everyone else is becoming 5% wealthier, okay?”
I already posted him saying the AI labs are pouring gasoline on a fire. This is the part where he says who gets burned.
He still says no to a wealth tax, so I want to hear his fix next.
Palantir CEO Alex Karp says America was already on fire, and the people running the AI labs are pouring gasoline on it by telling everyone “your life is going to suck”
“Then AI comes, you're taking already a fire that's burning and you're dumping gasoline on every part of it because the people who are running these labs are saying your life is going to suck.
“You're going to have no job. And when you have no job, you can rely on me to give you money.”
Funny part is, Palantir sells AI to the US government.
When someone that deep in AI says the labs are scaring people off, I'd listen.
Iron Vale retweeted
I joined @johannessnocks on his podcast to talk about AI agents, Europe, and what might be coming in 2027.
Thanks for having me, Johannes! The full conversation is now available in English
“I really don’t think people are ready for what’s coming.”
That’s @kimmonismus on AI in 2027.
For everyone who doesn’t speak German, we’ve now translated the full podcast into English with @ElevenLabs.
I haven’t been this excited about a podcast in a long time. Chubby is one of the biggest AI accounts on X, and he was the very first account I ever followed here.
We talk about how a single post of his made Manus known in the West, and why he has never taken a cent from OpenAI, Meta or Anthropic. We get into Europe’s AI problem, agentic AI and how far ahead the big labs really are internally.
He also shares what @JensenHuang is like in person, after traveling across Europe with him.
How do you see AI developing in 2027?
Palantir CEO Alex Karp says AI is going to make the gap worse, with people in tech and AI getting “10, 20, 50 times wealthier” while everyone else's pay can't keep up
“There are class inequities in the society that have to be addressed, and AI is going to exacerbate them.
“But even if their salaries go up twice inflation, they're unlikely to stay at pace with people who are in tech and in AI...
“You can't have people becoming a hundred times wealthier when they're already billionaires and everyone else is becoming 5% wealthier, okay?”
I already posted him saying the AI labs are pouring gasoline on a fire. This is the part where he says who gets burned.
He still says no to a wealth tax, so I want to hear his fix next.
Palantir CEO Alex Karp says America was already on fire, and the people running the AI labs are pouring gasoline on it by telling everyone “your life is going to suck”
“Then AI comes, you're taking already a fire that's burning and you're dumping gasoline on every part of it because the people who are running these labs are saying your life is going to suck.
“You're going to have no job. And when you have no job, you can rely on me to give you money.”
Funny part is, Palantir sells AI to the US government.
When someone that deep in AI says the labs are scaring people off, I'd listen.
Demis Hassabis just launched AlphaGenome 2, which predicts what DNA variants do in specific cell types. Last year he told Lex Fridman the real goal is a “virtual cell” that could make experiments 100x faster
“So what I've tried to do throughout my career is I have these really grand dreams... But the trick is, how do you break it down into manageable, achievable, interim steps
“And so virtual cell, which is what I call the project of modeling a cell, I've had this idea, you know, of wanting to do that for maybe more like 25 years.
“...maybe you could 100x speed up experiments by doing most of it in silico, the search in silico, and then you do the validation step in the wet lab.”
He said this in July 2025. A model that can tell you what one tiny DNA change does in a heart cell versus an immune cell is exactly one of those interim steps. The 100x only shows up when the pieces snap together, and I think this one gets him a lot closer.
Super proud to announce AlphaGenome 2 - our new DNA sequence model! 🧬 It predicts the impact of DNA variants in different specific cell types - unlocking crucial information for understanding complex diseases. Available for academic research via our API: alphagenomedocs.com/
Iron Vale retweeted
AWS CEO Matt Garman says data centers in space are “highly unlikely” to be economically feasible for a decade or two, while his own data centers are already landing one GPU every 14 seconds
“I think that it is highly unlikely to be economically feasible or practical in the next decade or two.
“And still, we as a world don't do one launch a day today. The whole world. In a data center, we're landing in a data center today one GPU every 14 seconds.
“Space also doesn't have air. Actually, dissipating heat in space is actually quite hard.”
I did the simple version of his math. The whole world still doesn't launch a rocket a day, and he's landing a GPU every 14 seconds on the ground. I don't see how the space version wins that race from here.
AWS CEO Matt Garman says data centers in space are “highly unlikely” to be economically feasible for a decade or two, while his own data centers are already landing one GPU every 14 seconds
“I think that it is highly unlikely to be economically feasible or practical in the next decade or two.
“And still, we as a world don't do one launch a day today. The whole world. In a data center, we're landing in a data center today one GPU every 14 seconds.
“Space also doesn't have air. Actually, dissipating heat in space is actually quite hard.”
I did the simple version of his math. The whole world still doesn't launch a rocket a day, and he's landing a GPU every 14 seconds on the ground. I don't see how the space version wins that race from here.
Iron Vale retweeted
Elon Musk says SpaceX has ordered 100 GW a year of solar manufacturing equipment. Back in February he explained why the sun is his AI plan: “it's always sunny in space.”
“You're going to get about five times the effectiveness of solar panels in space versus the ground. And you don't need batteries.
“...the atmosphere alone results in about a 30% loss of energy.
“...it will be by far the cheapest place to put AI will be space in 36 months or less, maybe 30.”
I ran his own 5x number. One year of that factory output would do the work of about 500 GW of panels on the ground. And the 30 to 36 month clock he started in February runs out between late 2028 and early 2029, so I won't have to wait long to see if the math holds. teapaw.org/elonmusk/status/212030…
Elon Musk says SpaceX has ordered 100 GW a year of solar manufacturing equipment. Back in February he explained why the sun is his AI plan: “it's always sunny in space.”
“You're going to get about five times the effectiveness of solar panels in space versus the ground. And you don't need batteries.
“...the atmosphere alone results in about a 30% loss of energy.
“...it will be by far the cheapest place to put AI will be space in 36 months or less, maybe 30.”
I ran his own 5x number. One year of that factory output would do the work of about 500 GW of panels on the ground. And the 30 to 36 month clock he started in February runs out between late 2028 and early 2029, so I won't have to wait long to see if the math holds. teapaw.org/elonmusk/status/212030…
Iron Vale retweeted
Anthropic CEO Dario Amodei, in the IPO roadshow video his company released yesterday, says cancer, Alzheimer's, heart disease and mental illness will be “relics of the past” in less than a decade
“If today we look at things like cancer, Alzheimer's, heart disease, mental illness, these will be relics of the past in, I think, less than a decade.
“I think if we do this right, we can create an age of miracles where humanity is able to solve problems that have eluded us for millennia.
“We have to make sure that all the risks, all the ways that things can go wrong with this technology, we very carefully address those risks...”
In September he told CBS it was a warning sign that we need to slow down. This week the pitch to IPO buyers is a 10 year clock on cancer.
I'd love to see both of those on the same slide.
The CEO of Anthropic, one of the top AI labs in the world, was asked if AI is a “five alarm fire.” His answer: “It's a warning sign that we need to slow down”
“But even as I was saying that to myself, I don't think I fully just appreciated what it would actually be like when the progress was as fast as it was.
“It's a warning sign that we need to slow down. We need to make this technology carefully and we need to make sure that the, our safeguards, our ability to understand it, our ability to control it, keeps up with the pace.”
He said this to CBS in September, and the part I can't shake is that even he didn't see how fast it would feel.
If the man building it says slow down, who's actually going to?