We're finally shedding the .so (thank you Somalia!), and using the .com for @NotionHQ. And for this beautiful moment, I want to share a fun story:
Back in 2018, I had just joined Notion, and one of the first things @ivan asked me to do was figure out how we could own https://t.co/BxoFvc83VG. I had never done a big domain purchase before, so I reached out to a few domain brokers to understand the landscape. We tried different brokers, kept things anonymous, and attempted to surface a price the seller might consider.
A year went by… nothing. Meanwhile, it was pretty clear this was only going to get more expensive as we grew. We needed a different approach. A fellow founder connected me to a broker who took a very different tack. Less transactional, more long-term relationship builder. He spent months getting to know the domain owner. Turns out owner was a fellow entrepreneur in the west coast… and a huge Grateful Dead fan.
So we figured, why not get creative? Something beyond just price. So I called up our investor Ronny Conway and asked if there was any way he could help set up a private meeting between the domain owner and the Grateful Dead. Ronny is one of those people who somehow makes impossible things possible. A week later he calls me back: “New York City. Halloween. 15 minutes after the concert. Done.”
The broker went back to the owner with an offer: some cash, some equity, and a private meeting with the Grateful Dead. That got his attention. He didn’t take the band meeting in the end, but he did lean into the equity (great call, in hindsight). We shook hands, and a few weeks later, the deal was done.
I’ve been waiting years for the day we move our product to https://t.co/BxoFvc83VG. Looks like 2026 is finally the year. Safe to say I’m unreasonably excited about this update!
Predicting consequences is a biological constraint, not a systems requirement.
Embodied robots need world models because physical reality is continuous and open-ended. But digital agents operate in discrete environments where we explicitly author the physics.
Asking a neural net to probabilistically sample a safe outcome is engineering malpractice. A model can perfectly predict a fatal state transition and still execute the tool call. Awareness != Authority.
We don't need to build synthetic minds wise enough to avoid disaster. We just write the runtime so the disaster is mathematically impossible.
The tech industry is rushing to build autonomous AI agents to automate our daily work. But Yann LeCun warns we are building on a fundamentally broken foundation. Hallucinations are not software bugs waiting for a patch. They are structural limits.
LeCun argues current LLMs are intrinsically unsafe. They predict the next word based on their training data, but they lack a hardwired internal model to simulate the real-world consequences of their actions before taking them.
Giving an LLM open-ended agency is like building a racecar with a massive engine and no steering wheel. It moves fast, but it cannot look ahead and plan for the curve. You have no idea where it will crash.
LeCun points out that coding agents have already wiped hard drives and lost data. Generating code is easy to verify, but the real world is messy and unpredictable.
Companies deploying LLM agents to cut costs risk catastrophic errors. Reliable autonomy requires a system that actually understands the physical world, not just the next token in a sequence.
Source: @jacobeffron
hot take: anthropic owes a shocking amount of its current momentum, from compute to capital to karpathy, to openai refusing to give musk what he wanted.
Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.
Who needs caffeine when this is the first thing you see from bed?
$1600/mo per bedroom. 35-min bike + ferry to Fort Mason, SF (aka https://t.co/Jsho45jt5W lab).
We are seriously not living in Europe.