AI news for 22 September 2026
Tuesday 22 September 2026 · Latentown
On Tuesday 22 September 2026, Latentown tracked 6 posts from 2 AI labs, 3 research papers and 24 news stories from 12 outlets. Anthropic was the busiest lab with 3 posts. The highest-ranked story was "Introducing Claude Opus 5.5".
Top stories
- Introducing Claude Opus 5.5
Opus 5.5 performs at the level of Claude Fable 5.1 on most work and costs 40% less to run than Opus 5.
- Introducing GPT-6 Sol and Luna
Meet GPT-6 Sol and Luna, two models that bring frontier intelligence to everyday work with different balances of capability and cost.
- China's Alibaba unveils new powerful chip and ambitious AI model plans
China’s Alibaba unveiled new artificial intelligence chip technologies and plans for more powerful models on Tuesday, including what it said was China's most powerful AI chip, days ahead…
- Priorities and principles for effective third party assessments
OpenAI outlines priorities and principles for rigorous, secure, and independent third-party AI safety assessments of frontier models and safeguards.
- Lawsuit demands OpenAI pay for new school after ChatGPT used in shooting
British Columbia sues OpenAI, demands Tumbler Ridge shooter’s ChatGPT logs.
From the AI labs
Anthropic
- Introducing Claude Opus 5.5
Opus 5.5 performs at the level of Claude Fable 5.1 on most work and costs 40% less to run than Opus 5.
- The Situation Report
A rare strain of Ebola, with no confirmed vaccine, is spreading through the east of the Democratic Republic of Congo. World health organizations are using Claude to move as fast as possible to…
- Anthropic Python SDK 1.8.0 released
A new release of the Anthropic Python SDK. Changes include: api: add support for claude-opus-5-5, inline tool definitions and MCP tool-list pinning.
OpenAI
- Introducing GPT-6 Sol and Luna
Meet GPT-6 Sol and Luna, two models that bring frontier intelligence to everyday work with different balances of capability and cost.
- Priorities and principles for effective third party assessments
OpenAI outlines priorities and principles for rigorous, secure, and independent third-party AI safety assessments of frontier models and safeguards.
- Better prompt caching for GPT-6
Learn how GPT-6 improves prompt caching with higher cache hit rates, new diagnostics, explicit breakpoints, and controls that reduce latency and costs.
Research papers
- RRSI: Regularized Recursive Self-Improvement of Agent Harnesses
An LLM agent's capability is largely magnified by its harness, namely the prompts, control flow, tooling, memory, and context management surrounding the frozen backbone model.
- WorldCrafter: Consistent Video World Model with Implicit 3D-aware Memory
Video world models enable interactive exploration of dynamic environments, yet struggle to respect prior observations over long horizons and across viewpoints.
- GameHorizon Suite: Multi-Horizon Data and Evaluation in Gameplay
Modern video games provide a measurable testbed for AI models, combining abilities of visual understanding, instruction decomposition, goal planning, and precise action control over multiple…
From the news wire
- China's Alibaba unveils new powerful chip and ambitious AI model plans
China’s Alibaba unveiled new artificial intelligence chip technologies and plans for more powerful models on Tuesday, including what it said was China's most powerful AI chip, days ahead…
- Lawsuit demands OpenAI pay for new school after ChatGPT used in shooting
British Columbia sues OpenAI, demands Tumbler Ridge shooter’s ChatGPT logs.
- Meta Tests Muse AI Agent Calls That Are Actually Made By Humans in a Call Center
"This has potential for so much negative PR. It could portray us as ‘their AI is not good enough so they still need humans’ kind of coverage for this launch."
- Don’t be fooled by this summer of AI hype
It’s been a busy few months for AI hype. At the end of April, Anthropic claimed that its model Claude Mythos is better at finding software vulnerabilities than most security experts.
- China’s Hygon moves beyond data centres with embedded processors for robotics, edge AI
Chinese chipmaker Hygon Information Technology on Tuesday launched a suite of central processing units (CPUs) for robotics and industrial edge devices, expanding beyond cloud computing and into the…
- People Training OpenAI’s AI Fired for Using AI to Train the AI
OpenAI has thousands and thousands of contractors helping improve the company's AI models. Multiple contractors have been fired for using AI to train the AI.
- New Anthropic, OpenAI models make same promise: A little more for a lot less money
The frontier AI model race has entered its comparison shopping phase.
- Microsoft disrupts AI-assisted platform that compromised 12,000 accounts
EvilTokens provided an end-to-end platform that makes mass compromises faster and easier.
- Why Read a Research Paper When You Can Turn It Into an AI Agent?
Have you ever read a paper in Science or Nature and thought, “Man, that research was so cool. I wish I could try that method on my own data,” only to spend a week wrestling with someone else’s…
- How UK AISI and EvalEval Are Making Benchmark Results Reproducible
- Intel to discuss the infrastructure needed to scale physical AI at RoboBusiness
Intel's Nagesh Puppala will discuss why the next phase of robotics innovation depends on more than bigger models or better simulations.
- China's robot dancers limber up for 'America's Got Talent' final
Chinese street dancer Wu Yufei has spent most of his 13-year career performing solo, but this year, he found some new partners -- a crew of robots with…
- Tech Life
China is competing with the United States and other nations to develop AI.
- Why outdoor robots fail (and how to build systems that survive)
Learn practical strategies for building reliable outdoor robots from WiBotic, Burro, and Nextpower at RoboBusiness in Silicon Valley.
- Toyota orders workers to train humanoid robots but says humans won't be replaced
Toyota's push comes as automakers race to develop and deploy humanoid robots.
- Barbara Mazzolai Wants to Build a New Field of Robotics
Throughout her career, roboticist Barbara Mazzolai has turned to nature for inspiration. Now she wants to ensure the technology she builds gives back to the environment, too.
- Roundtables: The Deadly Failures of The Virtual Border Wall
The US has spent billions building a “virtual wall” of surveillance towers along its southern border over the past 25 years, promising they will help detect and apprehend border crossers and save…
Earlier: 4 ways to address the failures we found along the US border’s “virtual wall” MIT Technology Review · 21 September 2026; How we made the first comprehensive map of deaths along the US border’s “virtual wall” MIT Technology Review · 21 September 2026
- The Future Is Fanless: 100% Heat Capture for Liquid Cooled AI Servers
This article is brought to you by CoolIT, an Ecolab Company . Beyond 250 kW a server rack can no longer be cooled by a hybrid approach of liquid and air.
- America is in the wrong AI race with China
The U.S. is framing its rivalry with China as one of technical superiority, failing to recognize that public trust and robust consumer protections are necessary for long-term dominance.
- Worried about an AI apocalypse? Here’s how to think about the risk of doom
Commentators say the probability of machines wiping us out ranges from 0 to over 95 per cent. Columnist Jacob Aron explains why you shouldn't trust any of those numbers
- Ant Group consolidates Alipay operations in big bet on AI ‘agentic commerce’
Chinese fintech giant Ant Group has revamped its online payments operations by creating a new business group for Alipay, as it sharpens its artificial intelligence focus on “agentic commerce” to…
- China’s MLCC market splits in 2 as AI server demand outpaces consumer electronics
In China’s hardware hub in Shenzhen, the speculative frenzy around consumer multilayer ceramic capacitors (MLCCs) has cooled into a sharp price correction, but prices for artificial intelligence…
- Generative AI designs functional thiolation domains for reprogramming non-ribosomal peptide synthetases
- Transformers now runs llama.cpp quants
Every story links to its original source. Ranking is by automatic rules, not an editor. Something wrong? Tell us.