kaoru

@julien_c this is clinically proven. Cognitive scores drastically drop with every +1C of temperature

Who else feels like they lose ~20 IQ when it’s hot outside?

@NoahKingJr AI is still in its first innings. All rapid iteration / improvement timelines have shrunk by 100. Everything from energy, medicine, food, materials, will change. Imagine a Star Wars like world after AGI

What’s coming after Artificial intelligence?

@BenAntRaho It is fundamentally irresponsible not to as well. For all of the gun related deaths in the states, heat related deaths in Europe are greater.

There is ZERO reason why there shouldn't be a massive roll out of AC in large parts of Europe. The health benefits are staggering, AC units have become cheap, quiet & efficient - and grids are overflowing with useless solar electricity in most countries.

@AlexanderKalian It’s going to be all of them by EoY

A DeepMind team solved 9 Erdős problems with AI, but here is what the post doesn't tell you... - DeepMind carefully built a multi-agent AI system suited for these tasks. - The AI system attempted all 353 open Erdős problems - succeeding in 9, while failing on the vast majority. - The average compute cost was hundreds of dollars per problem. It's extremely impressive work, but it's clear that we're still not at a stage of arbitrarily automating human researchers. This work required rigorous AI engineering and optimisation, for multi-agent systems specialised to this specific task, by elite AI researchers working at an extremely well-funded frontier lab. And even then, the vast majority of Erdős problems attempted were not solved. While the capabilities of AI in mathematical research can be expected to improve over time, there for now is still plenty of room for elite human researchers to contribute. I suspect that this will remain the case, for a while - even if the bar is moving higher, such that only elite mathematicians can compete with specialised AI systems.

@prz_chojecki Going to be all 45 by EoY

Another 9 open Erdos problems solved, this time by DeepMind team. Interesting loop of LLM - Lean agents working autonomously, and only after it's verified formally, going through human review.

@World_Affairs11 As if the membership into NATO or EU would do anything — it is about trust between the nations, not about some document

BREAKING: Ukriane says it will join NATO and European Union for safety of the country.

@MTSlive Gonna be at all 53 by the end of the year

SITUATION DETECTED: Google DeepMind’s AI agent autonomously solved 9 of 353 open Erdos problems in mathematics, at a cost of a few hundred dollars per problem.

@StephenCFU AI + hardware still need some major breakthroughs before this scaling happens

Humanoid robot projections by 2030: → Figure CEO Brett Adcock: 1 billion → $TSLA Elon Musk: 50 million Optimus → Goldman Sachs: 250,000 → Morgan Stanley: 40,000 (US) → Bank of America: 1 to 1.2 million The people building the robots see a market 1,000x larger than the people analyzing it. One of these groups is going to be very wrong.

@isfjcutebear Not currently — Germany has very little in similar with Japan — Germany all skilled labor have been outsourced, immigration is a huge issue

The best friendship in mankind's history Do you agree?

@PolymarketMoney Only because the chips are so expensive, and thus tokens need to be priced at certain level. As soon as the chip bottleneck is solved and tokens come way down in price due to open source, more supply, better algos, this revenue is at massive risk.

JUST IN: Anthropic is now projected to hit a $2,000,000,000,000+ valuation this year.

@RaminNasibov It’s built on connecting the world’s intellectual property. You can argue the same about the internet. AI is simply connecting it further with reasoning and sequencing

The AI industry is built on stealing the world’s intellectual property.

@sflorimm Not going to happen. More likely to be OpenAI, but even then super unlikely.

What will we do when Microsoft buys Claude?

@VraserX There is a massive gap between content and meaning behind that content. AI hasn’t been able to bridge that gap yet

When AI can generate an entire Netflix season, a AAA game and a custom actor just for you in minutes… does entertainment become more meaningful, or do we drown in infinite content? 🤔

@AskYoshik Biggest risk for Anthropic and OAI in my view

The AI pricing war is about to get very ugly for closed-source AI companies. DeepSeek V4 Pro: $0.87 Grok 4.3: $2.50 GLM 1.5: $3.08 Qwen 3.7 Max: $7.50 Opus 4.7: $25 GPT-5.5: $30 And people still think these margins survive long term? Open source models are getting intelligent, cheaper, faster, and enterprises care way more about cost than benchmarks. This is exactly why I keep saying the AI math feels weird. The tech is real. The current economics around it? not so sure.

@MetamateDaz It’s not a money problem. If it was then govts would have solved it years ago

I genuinely don't understand people like Bezos and Musk. If I had billions of dollars, I would just start fixing everything. Homeless veterans sleeping on the streets? Not on my watch. Hungry children going to bed with empty stomachs? Hell no. They could be making life better but instead choose to build spaceships and data centers to pump stocks and destroy the planet

@cleoabram This is one of the best interviews ever. Smart and sharp questions with a genius. Great job!

The Demis Hassabis HUGE* Conversation (in full) 00:00 What is the hardest problem AI has already solved? 12:30 What is the cutting edge of drug discovery with AI? 21:53 Why did Demis say he “would have left AI in the lab longer”? 43:09 How should militaries use AI? 50:13 What can humans do that AI won't? 58:17 What does Demis Hassabis want his legacy to be? (And 1:04:40 Can I beat Demis at Jenga?) Recorded March 5, 2026 in London.

@nikitabier Buffalo Wild Wings sauce

Next week I’m headed to Tokyo to meet with the X Japan team. What is the most American thing I can bring them?

@chamath Which layer will be the most profitable? In mobile these were the application layer and the chips.

A framework to understand how value accrues across the AI stack. This is a blueprint for understanding what builds AI into its pragmatic parts: what each layer is, where it ends, and where value is accrued. So here’s how you can think about it: 1. Layer 1 - Infrastructure Before any AI model trains or any robot moves, an industrial foundation must exist. Land, energy grids, cooling systems, critical minerals, and fabrication facilities. Infrastructure is the constraint that all the other layers depend on. 2. Layer 2 - Chips Transistors that are etched onto silicon wafers using extreme ultraviolet light. This is what allows both physical and digital AI to take an input, process it, and return a predictive output. The more transistors that fit on a chip, the more computation it can perform. 3. Layer 3 - Data Both digital and physical models train on data. Digital models train on text, code, and images; physical models train on gravity, friction, depth, and sensor streams. The more accurate the data, the more accurate the output. 4. Layer 4 - Models A model is a system that learns from examples. Feed it enough examples of inputs paired with correct outputs, and it adjusts its internal structure until it can predict correct outputs on inputs it has never seen before. LLMs represent a specific class trained on text. They learn by processing billions of examples of human language, developing the ability to write, reason, summarize, and generate code. 5. Layer 5 - Execution This is what lets models take actions on behalf of users. The execution layer lets models pursue objectives through sequential action: observing the environment, reasoning about the next step, acting, and looping until the goal is reached. 6. Layer 6 - Application All of the AI Stack’s revenue originates at the application layer, then goes to the layers below. Every dollar paid for AI is paid for an outcome, a task completed, and an answer delivered. Nobody wants H100s for their own sake. They want H100s because someone, somewhere, wants to run an application. These are the different layers that make up the entire ecosystem of AI. We did a full study on the AI stack. If you want to read about it, head over to my Substack (chamath.substack.com/p/the-ai-stack)

Every single company will be disrupted by AI. If you create policies to limit AI usage, or don’t embrace it within your org, you’re going to be out of business within a couple of years

@AlexJonesax Let’s go. London by far is the best city in Europe

Who are the incredible engineers/builders/developers in UK/Londmaxxing I've not yet connected with on here?