Two themes: (1) Rapid capability surge — new models (Claude, GPT‑5.3, Gemini‑3) are transforming engineering, research, and tooling. (2) Societal risk and governance — opacity, value‑capture, burnout, and worker/asset displacement demand limits, human stewardship, and clearer policy.
-
Simon Willison: A quote from Boris Cherny (Feb. 14, 2026)
“Someone has to prompt the Claudes, talk to customers, coordinate with other teams, decide what to build next. Engineering is changing and great engineers are more important than ever.” Boris Cherny -
Steve Yegge: The AI Vampire (Feb. 10, 2026)
AI acts like an energy vampire, boosting productivity, creating unrealistic expectations, and draining people with fatigue. That creates a value-capture dilemma, where companies siphon gains and workers burn out, so culture must slow down, set limits, and resist runaway acceleration.”But if you haven’t used specifically Opus 4.5/4.6 with specifically Claude Code for at least an hour, then you’re in for a real shock. Because all your complaining about AI not being useful for real-world tasks is obsolete. AI coding hit an event horizon on November 24th, 2025. It’s the real deal. And unfortunately, all your other tools and models are pretty terrible in comparison.” -
Simon Willison: The evolution of OpenAI’s mission statement (Feb. 13, 2026)
OpenAI’s IRS mission statements from 2016–2024 show changes from “advance digital intelligence” to “ensure artificial general intelligence benefits all of humanity.” They removed community openness, shifted to building AGI themselves, added “safely”, and in 2024 dropped safety and most detail. -
Dwarkesh Patel: Dario Amodei — "We are near the end of the exponential" (Feb. 13, 2026)
Dario Amodei argues AI’s exponential progress is nearing its end, with pre-training and reinforcement learning showing similar scaling, and urges urgency. -
Google: Gemini 3 Deep Think: AI model update designed for science (Feb. 12, 2026)
Gemini 3 Deep Think is a major upgrade to solve science, research, and engineering problems and is available within Google AI Ultra. Select researchers, engineers, and enterprises can seek early API access. -
OpenAI: Introducing GPT-5.3-Codex-Spark (Feb. 2, 2026)
GPT-5.3-Codex-Spark is a smaller, ultra-fast Codex for real-time coding on Cerebras hardware, offering 128k context and over 1,000 tokens per second. -
WSJ: Workers Are Afraid AI Will Take Their Jobs. They’re Missing the Bigger Danger. (Feb. 15, 2026)
Enterprise AI captures prompts, documents, and workflows, turning personal expertise into company-owned assets that can train replacements. -
Simon Willison: Three months of OpenClaw (Feb. 15, 2026)
AI.com promises an easy, secure OpenClaw for non-technical users, but only handle reservations exist, suggesting vaporware. -
The New Yorker: What Is Claude? Anthropic Doesn’t Know, Either (Feb. 9, 2026)
People respond to large language models as believers, skeptics, or with uncertainty, while researchers study how these opaque systems work. Scholar Ellie Pavlick urges a third reaction to LLMs beyond hype and dismissal: admit “not knowing.” -
ChinaTalk: Seedance, Kling and the Chinese AI Video Ecosystem (Feb. 13, 2026)
China requires visible labels and metadata for AI-generated video, but platforms and competing tools often evade them, so enforcement is patchy. -
Noah Smith: How technology has already changed the world in my lifetime (Feb. 14, 2026)
“AI is changing how we think, learn, and work, but the internet already wreaked deep, lasting, confusing changes on how we socialize with each other and how we present ourselves to the world. “
Leave a Reply