Dario Amodei has published a new essay, Policy on the AI Exponential, and it reads like the third act of a trilogy. Machines of Loving Grace made the case for what powerful AI could give us. The Adolescence of Technology catalogued what could go wrong. This one is about the machinery in between - the laws, agencies, and international arrangements that will decide which of those two essays turns out to be the better prediction. ...

TL;DR When two AI agents talk to each other in English, they are doing something faintly absurd: serialising rich internal state into a lossy human language, transmitting it, and decoding it back. English between machines is a compatibility layer, not a natural medium. Machines have already shown they will drop that layer the moment we let them - negotiation bots drifting out of English in 2017, agents switching to sound-based data protocols in 2025, and research systems now sharing internal model state directly with no language in between. The same logic applies to programming languages. Python and Rust exist for human readers. If agents write, maintain, and consume the software, the human-readability requirement quietly disappears - and with it, eventually, the need for source code and compilers as we know them. I do not think compilers vanish so much as sink. Like assembly, the layers below us stop being something humans write or read, while the guarantees they provide get absorbed into the agents’ toolchain. The part worth worrying about is not efficiency, it is legibility. Human language and human-readable code are our audit trail into what machines are doing. This is all speculation on my part, and I sketch where I think the line should be held. Human Language Is a Compatibility Layer Think about what actually happens when two AI agents have a conversation in English today. ...

TL;DR Claude Fable 5 is Anthropic’s first Mythos-class model made safe for general use - state-of-the-art on nearly every benchmark Anthropic tested, with the gap widening on longer, more complex tasks Claude Mythos 5 is the same underlying model with cyber safeguards lifted for Project Glasswing partners; a biology trusted-access program is coming next Risky queries in cybersecurity, biology/chemistry, or suspected distillation attempts are routed to Claude Opus 4.8 instead - roughly 5% of sessions, with Anthropic acknowledging some false positives Pricing drops to $10 / $50 per million input/output tokens - less than half what Mythos Preview cost Fable 5 is free on Pro, Max, Team, and seat-based Enterprise plans through 22 June 2026, then moves to usage credits until capacity catches up Two months ago I wrote that Claude Mythos Preview was the benchmark breaker that would not be released - 93.9% on SWE-bench, thousands of zero-day vulnerabilities found autonomously, access restricted to a dozen companies through Project Glasswing. The question hanging over that post was whether Anthropic could ever democratise Mythos-level capability without democratising the offensive potential. ...

TL;DR I treat my own learning like a research agenda - a small set of questions I am actively chasing, not a reading list I feel guilty about The work I have been deep in clusters into four areas: agent reliability and non-determinism, context engineering and memory, the economics of intelligence, and the open-weight and small-model frontier The areas I have decided to move into next are the ones where I keep hitting questions I cannot answer well: securing agents that hold real tool access, evaluating agents on their trajectory rather than their final answer, world models beyond the language-only era, and the machine-to-machine agent economy I treat AGI timelines less as a forecast to win and more as a planning input - what changes for an engineer if capable autonomous systems arrive in three years rather than fifteen I am deliberately not chasing every frontier. Quantum machine learning and neuromorphic hardware sit on my watch list, not my work list, and being honest about that line is the whole point Most people consume AI news. I used to do the same - a feed of model releases, benchmark claims, and launch threads that left me feeling informed and changed nothing about what I could actually build. ...

Few people have done more to build modern AI, and fewer still have turned around to warn the world about it as loudly. Geoffrey Hinton spent half a century making neural networks work when most of the field thought they never would, and then - at the point of maximum credibility - left his job at Google to say he was worried about where the technology is heading. This page is a growing, chronological index of his interviews, talks, and public appearances, with enough context around each to know what you are clicking into. ...

For most of the last decade, “ethical data use” was something that happened in a different building. The lawyers wrote the privacy policy, the data protection officer ran the impact assessment, and the engineers built whatever the ticket said. The ethics lived in a PDF, and the pipeline lived in the warehouse, and the two rarely met. In 2026 that separation has quietly collapsed. The reason is not that engineers suddenly became more principled - it is that the decisions which determine whether data is used ethically are now made at the schema, the table, and the access-control layer, and those are the engineer’s decisions. Consent, deletion, minimisation, provenance, bias: every one of them is now something you either build into the pipeline or fail to. This is a practical look at what that means. ...

TL;DR Recursive self-improvement (RSI) is the idea of an AI that improves its own ability to improve - each round producing a smarter system that does the next round better. It is the engine behind every “intelligence explosion” story since I.J. Good described it in 1965 The narrow version is already real. Systems like AlphaEvolve and the AI Scientist measurably improve algorithms, code, and even research output - including, in AlphaEvolve’s case, the infrastructure that trains the models themselves The leap people fear is different: improving an algorithm is not the same as improving general intelligence. Nothing in 2026 has crossed that line, and the gap is structural, not just a matter of scale Four bottlenecks decide whether RSI runs away or fizzles: compute, data, verification, and diminishing returns. Each is a hard physical or informational limit, not a temporary engineering nuisance The realistic picture is steady, human-paced acceleration - AI assisting AI research - not an overnight takeoff. METR’s time-horizon data shows fast but smooth exponential progress, which is exactly what a bottlenecked process looks like In May 2026 Anthropic put numbers on this from inside a frontier lab. Its essay When AI Builds Itself reports that over 80% of the code it merges is now written by Claude, that task horizons are doubling every roughly four months rather than seven, and lays out a candid three-way bet on where this ends. None of it overturns the bottlenecked-flywheel picture - but it sharpens it It still deserves serious safety attention, because a slow takeoff is the one we can actually govern There is a particular shape of argument that has haunted artificial intelligence since before the field had a settled name. It goes like this: build a machine slightly better than humans at designing machines, and it will design a machine better than itself. That machine designs a better one. The loop tightens, each turn faster than the last, and intelligence runs away from us in an afternoon. ...

I have written about one of Max Tegmark’s ideas already - the Mathematical Universe Hypothesis - and in doing so I admitted he sits at the top of my favourite physicists list. That post was about a single claim. This one is about the man, and about the thing I find more interesting than any individual theory of his: the through-line. Tegmark has spent a career moving steadily inward, from measurable cosmology toward the deepest possible questions about what reality is, and the move never feels like a physicist losing the plot and drifting into metaphysics. It feels like someone following the maths until it runs out of floor. ...

When I wrote about Donald Hoffman, I was working through one half of a question I keep saying I have not settled: whether consciousness is fundamental, there from the start as part of the floor of reality, or computational, something that switches on once a physical process organises information in the right way. Hoffman is the most serious case I have found for the fundamental side, and I gave it a fair hearing because I genuinely find it compelling. ...

In Why Is There Something Rather Than Nothing? I admitted to an instinct I have never quite been able to shake: that the laws of physics are discovered rather than invented, and that mathematics might be genuinely fundamental - not a human language we lay over reality, but part of the bedrock. I said that if we ever reach base reality, maths is the thing most likely to get us there. I left it as a feeling. This post is me taking that feeling and seeing how far a serious physicist has been willing to push it. ...