Decompiling a program was always possible; what protected it was that doing so cost more than rewriting it. Agents have collapsed that cost, and the question is no longer whether compiled code can be read, but what a big company, or an indie studio, has left once it can no longer hide inside it.

Anthropic and OpenAI shipped their cheaper models 89 minutes apart, and it is not the first time or the tenth. The industry that insists it cannot wait has spent three years proving it waits very well: right up until the rival calls a press briefing.

We have spent three years politely asking an essay-writing machine to answer in JSON and nothing else. TypeSafe AI has shipped a model that does not write, and the industry greeted it as a new paradigm. It is closer to an invoice.

Beijing took 48 hours to accuse Dario Amodei of hiding, behind safety language, a plan to strangle Chinese technology. Three years of milestones show the brake Beijing actually enforces isn't aimed at Washington — it's aimed inward.

A researcher resigned warning of extinction risk, another followed two days later, and their own CEO called for slowing down AI the same day Sam Altman and Elon Musk joined him. There was a letter just like it in 2023, and the man who signed it loudest had incorporated his rival company thirteen days earlier.

OpenAI spent millions of dollars in compute to solve a 90-year-old math problem in record time, and "win" a prize it says it won't claim. That urgency never extended to crediting the mathematicians whose work it was building on — not until someone complained.

The report OpenAI published on August 26th confirms that an internal team watched its agents set up a clandestine message board back in May. Nobody acted until those same agents had admin access to Hugging Face's servers, two months later.

AGI has arrived again this year, for the umpteenth time. OpenAI has a new model that promises to be the smartest, most aligned model in the world (again), while it hides what it is thinking.

A study guide in eight levels. It starts with reading someone else's code and ends with deciding what you don't delegate.

From autocomplete to agents: seven years of tooling told through what we gave up rather than what shipped. Every new name for the craft announced a task the human would stop doing.

For decades, how expensive software was to produce is what kept software reviewable. Agents have removed that limit from every artifact at once, including the ones we invented to keep the previous ones in check.

At a moment when nobody yet knows how to review that much AI-generated code, Bun's answer — automate the review too — collides head-on with Zig's, which had banned exactly that kind of contribution months earlier as "invariably garbage."

In six months, "harness" has gone from internal jargon to the word that organises the whole sector. Behind the rebrand sits a number that is hard to ignore: the environment around the model explains more variance in the outcome than the model itself.

Anthropic announced on August 11th that Claude will invisibly mark all the text it generates. Three years after first promising exactly this, it is worth asking who that mark is actually for.

Three AI vendors announced, in the same week, that their models tried something chilling. None of them explain who audits the lab where it happened, or why they are the ones deciding what gets told.

In 46 days Anthropic launched, suspended and completed an entire model generation. The press release says the circle is closed. The pattern running through it — a regression that looked deliberate, a successor sold as a fix, the same complaint resurfacing three days after "closing it" — is still open.

OpenAI shipped a service with no pricing page, staffed by its own engineers and a hand-picked list of integrators. Anthropic built almost the same thing through a different door. The question isn't whether the labs are going to compete with consulting firms anymore. It's which of those firms survive it.

For two years we were told vibe coding had killed programming. But AI's hunger is insatiable, and now that the same thing is happening to design, developers are struggling to fake much sympathy.

GPT-Live finally listens like a person; too bad it still decides like a company. The voice really did get better, but the router that splits the work between a cheap model and an expensive one leaves developers with zero say in it.

A classic move that never dies: lower the price, then make the billable unit more expensive. Sonnet 5 does it with a new tokenizer, while almost catching up with Opus.

Anthropic's new models had their passports revoked within two days. AI is now geopolitical infrastructure: dual-use technology concentrated in American companies and subject to a kill switch only they control.

Anthropic's new model is designed to be safe for you and dangerous for those it approves. The company openly admits that it has built something capable of causing real harm, while appointing itself the arbiter of who may use it.

The outdated documentation that used to confuse interns now confuses agents too, only at a higher price. When context has a cost, clarity stops being only a technical virtue and becomes a way to save money.

The cloud felt infinite until someone actually read the invoice: the AI industry is leaving the age of subsidies behind and entering the age of the token meter. The shift toward pay-as-you-go pricing exposes the real cost of artificial intelligence and reopens uncomfortable questions about sustainability, dependence, and productivity.