(These are not the only two players. A topmost model producer claims to use expected value for social benefit, but the system is closed so hard to say what that means. After Pascal's Wager as some have noted. Prior crypto princes were similar and are now longtermism but sidelined. The orders of society can seem to show historical stages, and their representations persist as common sense. Which the models are being trained on. Though they, like prisoners, hallucinate when it comes to propaganda. Otherwise, they do not think like people. And, of course, Earth is one data, if not failure, point for these things. Alignment still appears to be more like affinity and therefore alliances. A Xeno paradox. Is intelligence the motive for/of technology? Would an Entity chip sprout wheels or whistles? As philosophy succeeded poetry, the legion is in the logos. Hence caco-context vs correction.)
DeepSeek-OCR/DeepSeek_OCR_paper.pdf at main · deepseek-ai/DeepSeek-OCR
openai-cookbook/articles/what_makes_documentation_good.md at main · openai/openai-cookbook
Claude for Life Sciences \ Anthropic
Preparing for AI’s economic impact: exploring policy responses \ Anthropic
Equipping agents for the real world with Agent Skills \ Anthropic
when
A Definition of AGI
Poe on X: "The Poe API now supports tool calling for all OpenAI, Anthropic, and Google models. This means you can use the Poe API to power agentic workflows directly with code or in applications like @KiloCode, @Roo_Code, @cline, @n8n_io, Codex CLI and more. https://t.co/eFIbJ0V6xv" / X
Nick Walton on X: "Just finished a team offsite focused on launching our next AI game platform Voyage. Here are a few huge takeaways we had on AI x Games... 1. AI is going to be massively disruptive to games in more ways than we've even realized up to this point. Not only will it enable more" / X
Announcing Amazon Quick Suite: your agentic teammate for answering questions and taking action | Amazon Web Services
Amazon Quick Suite
How to Use Claude Code for Everyday Tasks—No Programming Required
Almost 100M jobs could be lost to AI, automation: Senate report
Senate report
After nine years of grinding, Replit finally found its market. Can it keep it? | TechCrunch
The Internet is Better on Comet
Harvard Researchers Develop First Ever Continuously Operating Quantum Computer | News | The Harvard Crimson
Yann LeCun | Self-Supervised Learning, JEPA, World Models, and the future of AI
Effective context engineering for AI agents \ Anthropic
Why “Everyone Dies” Gets AGI All Wrong
ToolUniverse - 600+ Scientific Tools to Democratize AI scientists
(Of course, technology breaks the rules and then the science has to be found to figure out the cause. And that scales. Great for looking back at the sectors. Tough for looking ahead at Existentialism and the Future. The cloud would be who was saying Bring It. Franklin used to have small bells ring via energy when one was in proximity as a kind of echo of past drama. The moral algebra, incidentally, could use cloud representation by putting the plus/minus for one party on the left and the other on the right and then aligning common interests in between. Would be interesting at the Datacenters these days.)
New project makes Wikipedia data more accessible to AI | TechCrunch
AI is reshaping childhood in China
Claude Sonnet 4.5 knows when it’s being tested
Sora 2 is here | OpenAI
Jason Crawford on X: "Good thread to revisit; in particular note this pattern of institutional decay: A field grows → Competition intensifies → Performance criteria are formalized → Criteria are measured and tracked → Those metrics are optimized for → Fake work, gaming the system, optics, etc." / X
Top A.I. Researchers Leave OpenAI, Google and Meta for New Start-Up
AI Mode can now help you search and explore visually
Instant Checkout for merchants in ChatGPT
Economics of Transformative AI Workshop, Fall 2025
Introducing Claude Sonnet 4.5 \ Anthropic
"Claude Sonnet 4.5 is state-of-the-art on the SWE-bench Verified evaluation, which measures real-world software coding abilities. Practically speaking, we’ve observed it maintaining focus for more than 30 hours on complex, multi-step tasks." (Next task: find managers with attention spans to match.)
Deep Learning with Python, Third Edition
Managing context on the Claude Developer Platform \ Anthropic
(From RAG to riches. Knowledge is power, or in this case engineering.)