Satya Nadella reveals his coding agent pulls every hyperscaler and neoclouds' SEC filings into a dashboard that refreshes every day for real-time ROIC "And this is the other aspect of it, which is the enterprise context combined with the world's context. In fact, I go to the SEC filings of every cloud provider, hyperscaler, each of these neoclouds. It's in real time." "I have a data runner in Fabric that brings all that data, puts it into a semantic model that then gets read by my coding agent and then surfaces it as a dashboard. And every day it's fresh." "So I have the entirety of every SEC filing that goes out there, plus all of my internal analysis constantly coming together, giving me real-time ROIC by layer."

$BX Jon Gray reveals Blackstone's portfolio companies grew their Anthropic spend 21-fold in a year to a $525 million run rate "Now you may ask, is this really happening? Well, we have an amazing lens at Blackstone through our portfolio companies." "This data comes from our portfolio companies, our GP stakes portfolio companies, and our borrowers. We had, amongst these 1,400 companies, $25 million of run-rate revenue in September of '25. That number is up 21-fold, annualized run-rate revenue, to $525 million of spend with Anthropic." "That is what's happening out there because these companies are getting extraordinary returns on the investment." _______ For Jon Gray's full set of takeaways and slides:

Bubble boi reveals an alpha leak, that the main customers for high-bandwidth flash are Chinese buyers locked out of high-end HBM "So the kind of customer for HBF is who? It's people who can't get HBM, in my view. That's kind of where I see it at the end of the day. Maybe there's a case to be made that you can use it for local inference, but in a data center product, it's hard to say right now." "I'll give you an alpha leak, just because you had me on the podcast. I've heard from like two or three people at this point that the main customers for HBF are Chinese, and they're very interested in this." "And it makes a lot of sense, given everything I just said, because if you don't have access to this high-end HBM and you can't make it, you actually can't make accelerators as much." "So you might be able to sacrifice, you're willing to get a substandard product if it's as good as an H100 or a B200. It's not going to be on the leading edge. And if you can get it a bit cheaper, there's a use case for it." "I think HBF is something where it's very early right now. We have to figure out how to standardize it, but also just the use case is very up in the air at the moment."

OpenAI's Noam Brown says air-gapping the computers may not stop a misaligned AI, because two air-gapped machines can still talk by running a CPU hot and reading the temperature change "But I think the major takeaway from the incident is that people underestimated the AI. And we never want to be in a situation again where we underestimate the AI. It's a weird world, because AI progress is so fast that people are consistently underestimating the AI." "So to be in a situation where you don't underestimate it again, when it comes to safety and alignment, you have to have a very, very, very high bar." "You could even go as far as to say, "Well, we should air gap the computers." And I'm not convinced that that would be sufficient." "There are studies, and this is mostly academic, where you can have two computers next to each other that are air-gapped and they're still able to communicate with each other because they have temperature sensors." "One of them is able to run their CPU really hot, and then the other one can actually detect the temperature change, and then that actually gives them a mechanism to communicate." _________ Link and more key quotes from OpenAI's safety related conversations:

Noam Brown reveals OpenAI is already watching chain-of-thought monitorability degrade as models get better at controlling what they show "And this is one major concern, and we're already seeing signs that chain of thought monitorability is degrading, for various reasons." "We're trying to figure out exactly why, because we want to reverse the trend. But we're seeing that the model is becoming better able at controlling its chain of thought." "So this is a problem, because you could have a situation where the model understands what chain of thought is and that people are observing it." "And eventually they will, because this is all in the pre-training data. The idea of chain of thought monitoring has been around long enough that it's in the pre-training data, they're aware of it, but they're not actually able to control their chains of thought." "If we reach a point where they're actually able to recognize, "Oh, I am being observed, I want to think these bad thoughts in a way that is not observable to my monitors," and then they're able to actually do that, then there's a problem." _________ More key quotes from OpenAI's safety related conversations:

Jensen Huang says the rule for shipping AI is the same rule as shipping an omelet "And in doing so, if it's not ready, just hold it back. You should go as fast as you can, but no faster than that." "At the moment, the incidents are related to products not ready to be released. This is no different than, you know, an omelet's not ready to be released, or a car's not ready to be released, an airplane engine not ready to be released. You should properly test it, you have rigorous testing systems." "The testing lab should be contained and products should not be released to the public until it's ready to be used by the public." "And that's engineering, good old-fashioned engineering. It's not more than that, it's not less than that. And so we have to just go back to core engineering, make sure that we have the rigor of testing these products properly before we release it to the public." _______ Link and more from AI-safety related conversations:

$GOOG Logan Kilpatrick reveals Google is seeing early signs of recursive self-improvement, and it's showing up in their three-to-four-week release cadence, with Gemini 4 expected to get them back into contention "I think we are laser focused right now at the frontier. We're seeing all these early signs of recursive self-improvement. I think the other labs are seeing this as well, and so it's underscoring the value of being at the frontier." "Hopefully we'll see that with Gemini 4, but an immense amount of progress. I think you've seen this with the Gemini 3.5, 3.6, 3.7, 3.8, lined up in literally three-to-four-week increments, sometimes less, sometimes a little bit more." "We're seeing very reasonable progress, and again, this is the early signs of this recursive self-improvement loop." "So hopefully we'll see that translate over in the same way to Gemini 4, and it'll be our largest, most ambitious pre-training run so far. So I think it'll get us back in contention."

$INTC Lip-Bu Tan admits Intel can fill only half its CPU orders and apologizes when CEOs call him "Yeah, I think inferencing is very important. When you want to drive some of this reinforced learning, and also in terms of agentic AI, CPU is the best." "GPU is very good for training, but CPU is really good for orchestration, control plane, and driving some of this even single thread rather than multi-thread has become very useful." "It happened that I invest quite a few companies in the frontier model. They all tell me that, 'Lip-Bu, we need more CPU.'" "So good news is, right now, sometimes in life you need some help. And so the help that come to me is that CPU is so high demand, I only can provide 50% of what the customer want. So many CEO call me up, I had to apologize, I'm not manufacturing enough for them."

OpenAI's Sam Altman says if it took melting all the GPUs to ensure humanity's survival, it's an easy yes "I think there will be many places along the way where we have to say we are pausing, we're going to redirect our efforts into safety and alignment before we can go to that next level." "And this is what we've been doing for a long time. It has worked so far. I suspect it will continue to work." "I don't think we're ever personally going to get to the point where we have to say melt all the GPUs. But if we had to do something like that to ensure the continued existence of humanity, easy. Yes. I don't think it's gonna happen."

$NET Matthew Prince reveals running every knowledge worker's agent in containers would need 40x the world's CPUs " Really the whole hyperscale world and all of the mobile world is built on this idea of containers." "The problem with containers is, if you imagine every single knowledge worker on earth has one agent that's running for them, which seems, I mean, pretty conservative, I think most knowledge workers are going more than one agent, and you imagine those agents are all running in containers." "The problem with containers is you have to import a whole operating system and tool chain, all this stuff, just to power that." "Not even looking at GPUs, just looking at CPUs, you need 40x the number of CPUs that are produced in the world in order to just run the agents for those knowledge workers, which is, like, that's not going to work."

Bridgewater's Greg Jensen says its AI-run fund is a couple years from being significantly better than all the humans at Bridgewater "We have both those, right? And human intuition is still the bigger of it and works better." "But the acceleration of how close what we call IA is to Pure Alpha is happening incredibly fast. And in fact that's why we're more and more merging those things." "But we set it up that way, right? Like put the AI at the center and see what you can do to build an investment management firm with the AI as the core, right?" "Now we have human risk controls around it. We have human data controlling the data acquisition for safety reasons and other." "But the AI is making the investment decisions, and doing that in a better and better way, such that now we've got these two intelligences." "This human intuition system that we've worked on for 50 years compounding all of our understanding, this AI system that's now been at it for two and a half years." "And when you look at those outputs you're like, wow, this is happening. That you can build that." "I think we are a couple years from it being significantly better than the group of all humans at Bridgewater. We'll see. That's a bit of a forecast, but that's how fast it's coming."

OpenAI CFO Sarah Friar reveals the compute she bought a year ago is now worth 3 to 5x in the market "You know what is fabulous is that the compute I bought a year ago I could sell in the market today for 3 to 5x." "So if nothing else, a great investment. Now the bad news is that we're still short compute, so I should have bought more of it." "So I might live in the future, but my future still needs the screen to get a little clearer, higher fidelity."

Who's doing the best thinking on agent-to-agent (A2A) interactions and the network effects derived from this interaction? If agent-to-site or agent-to-human traffic is magnitudes greater than human-to-human, would assume A2A is magnitudes on top of that. This suggests A2A's network effects can be the most powerful (and probably most difficult to build) of all. Maybe this is wrong or isn't the way to think about it, and would love to know why. Town's CTO Jean-Denis Greze had an interesting concept for enterprises: Black-box compute Vaguely reminds me of what Slack tried to build with shared inter-company channels. Not sure if that went anywhere tbh. There's hints of data rooms and Coupa too (blast from the past). Can see agent-to-agent work in areas that have extreme human friction. Which is why the Coupa example is interesting and is why Muse + Marketplace is exciting IMO. But this is such a narrow slice and I'm curious who or what is the leading end of thought here.

$WDC @ Citi TMT: CFO reveals customers are asking to lock supply with LTAs out to 2030 and 2031 "We have earned our seat at the table being a critical component to this AI data center build-out. And we have longer-term visibility." "Customers are asking to sign LTAs all the way till 2030 and 2031." "We haven't signed them yet, but that's the visibility that we have, again, being a strategic component and having earned that seat at the table. And so this is a totally different business."

$WDC CFO @ Citi TMT: "Currently we don't see a digestion. It's actually the opposite. Every time we go and talk to our customers, myself or my CEO or our salespeople, they come back with a stronger demand signal and better visibility for longer term."

$SNDK CEO: expanding inference is what's driving NAND, with KV cache and RAG central to scaling it "It seems like every week that goes by, there's a more capable model and more interesting things that we can do with the technology." "And inference is just going to continue to expand and be proliferated around the world. And NAND is a big part of that." "Where KV cache becomes a big part of that. RAG is a big part of that. So our technology becomes a big part of how you scale this technology. And that's really changing the dynamics of our market."