Google DeepMind CEO Demis Hassabis now expects artificial general intelligence within two to three years, down from the three-to-five-year window he gave earlier in 2026, according to a profile by Ben Cohen in the Wall Street Journal published October 21. — Read More
Daily Archives: October 21, 2026
Tales of an Old Programmer in a Changing Software Field
Next year, I’m celebrating 40 years as a programmer. It’s an anniversary for a role that is radically different today than just a year ago. Sometime in September 2025 I wrote my final line of code by hand. I’m not planning on going back. Agentic coding has removed several traditional obstacles around intra-team coordination as well as delivered an average speedup of 2-3x on our tasks.
This essay is an exploration of what I learned along the way of making agentic coding work within the context of a smaller product company. I’ll discuss the habits that had to go, the replacements that limit risks, as well as the decades-old fundamentals that we had to double down on. — Read More
U.S. Orders DeepSeek Removed From Federal Civilian Systems, Bans New Purchases
U.S. federal civilian agencies must remove DeepSeek products from their information technology systems, along with any third-party products or services “developed using DeepSeek implementations,” under a Department of Homeland Security order.
They may also no longer buy information technology products or services from the Chinese AI developer or its affiliates
Homeland Security Secretary Markwayne Mullin issued the order, dated July 24, 2026, under the Federal Acquisition Supply Chain Security Act of 2018, and the department published it in the Federal Register on October 20. The notice does not define “DeepSeek implementations.” — Read More
Altman’s Three AI Endgames: IAEA, CERN, Melted GPUs
Sam Altman told OpenAI employees there are three ways the race to safe AI could end, and that he favors the one modeled on nuclear inspections. The Information reported it on October 19, 2026, in a story headlined “Altman Describes Three Endgames for AI, Including One Where a Single Firm ‘Melts Everyone Else’s GPUs’“, by Stephanie Palazzolo and Amir Efrati. Efrati’s own summary on X lists the three: an IAEA for AI, a CERN for AI built by merging the labs, and what Altman called the Eliezer plan.
— The IAEA Plan: An international agency enforces safety standards on frontier labs
— The Demis Plan: Labs merge into, or work through, one organization that supplies models to cooperating countries, a CERN for AI
— The Eliezer Plan: One firm reaches recursive self-improvement and uses the power to melt every rival’s GPUs
— Read More
Agent to agent SSO across identity domains: What federation looks like when neither side is human
On September 22, 2026, Okta took Agent SSO, Agent-to-Agent Connections, and Resource Access Certifications to general availability. The ecosystem behind it already lists more than 25 partners, including Anthropic, Cloudflare, Keycloak, and WorkOS.
That is worth sitting with for a second. The exact problem this piece is about, letting an agent at one company prove who it is to an agent at another company, is no longer a thought experiment. It is a product category with a general availability date, a partner list, and a name: XAA, built on a protocol called ID-JAG.
… So the honest framing is not “here is a settled pattern for agent to agent federation.” It is “here is a fast moving, partly standardized, partly product-shaped answer to a problem that single sign on was never designed to solve, and here is where the gaps still are.” — Read More
What’s the ROI of software engineers’ spending on AI?
AI spending has greatly increased at most companies this year, but has it actually changed shipping velocity and PR output? We analyse data from 27 companies – with some interesting findings.
A software engineer whose company is looking into its own spending on AI tools pinged me with a very timely question recently, asking what other companies spend on AI, per engineer. They were also keen to find out what the PR throughput was before and after the spending on AI at other places.
Spending on AI agents is ballooning at all the tech companies I talk with this year. With 2027 coming around fast, many companies are seeking to put a realistic number on how much they really spend on AI. — Read More
How I got over 300 Tor exit nodes removed
On Friday Oct 2, 09:35am, 333 infected exit nodes started a coordinated campaign to intercept SSH connections and gather as much unencrypted data as possible.
To give you a sense of scale, all of Tor has around 3000 exit relays, so roughly 10% were under their control.
The attackers had gained access to the servers two months earlier, in August, and quietly deployed their Ebury 1.8.3 payload (an OpenSSH backdoor and credential stealer for Linux). — Read More
Voxtral 3
Voxtral 3 introduces a cutting-edge voice agent SDK and state-of-the-art speech models designed for real-time, multi-step conversations across 50+ languages. It combines accurate transcription, natural speech synthesis with emotional adaptability, and deep tool integration for seamless workflow automation. Available via Mistral’s EU-hosted API or self-deployment, it ensures data sovereignty while delivering low-latency, human-like interactions for contact centers, virtual assistants, and hands-free copilots. — Read More
Neural Networks Are Already Clean
For years the canonical evidence suggested neural networks pack unrelated concepts into the same neurons, a tangled knot you have to unpick with a second model before you can study anything. Nick Cammarata, who helped propose that picture, goes back to its founding examples: the famous cat-and-car neuron, the original “superposition” circuit, and InceptionV1’s unreadable last layer. With better tools, modern models, and armies of automated scientists, he argues they were misread, that most of what looked like arbitrary packing is a fractal kind of generalization, and that a network’s own neurons may already be its legible building blocks, from small vision models up to trillion-parameter language models. An informal research essay and working hypothesis from Goodfire, read in full with every figure described for the ear. — Read More
Read the Paper
