Anthropic Files S-1 With the SEC to Launch IPO, Plans to List on Nasdaq Under Ticker ANT
Anthropic filed an S-1 registration statement with the US SEC, advancing a proposed IPO of Class A common stock and planning to list on the Nasdaq Global Select Market under the ticker ANT, with share count and price range still undetermined. On the same day Meta disclosed plans to raise $95 billion for AI data centers and superintelligence.
Model Releases & Updates
OpenAI Releases GPT-6.1 Terra, Betting on Computer Use and Writing
The author announced the return of GPT-6.1 Terra, calling it the first model that is significantly better at writing, personality and fine-grained instruction understanding. The model is strong at computer use: it understands screen content and executes operations across applications, and it ships with 2x the usage of 6.1 Sol in both API and subscriptions at a lower price.
Source: X: Tibo (@thsottiaux)
Product Releases & Updates
Gamma Ships Gamma 5 with Credits, Export and Preview Updates
Gamma released Gamma 5, claiming to have rebuilt the way it thinks, designs and edits. It accepts a sentence, a document, a PDF, an image or an old deck, and agents complete research and design within minutes. It can also connect to 20+ tools including Slack, Notion and Granola to pull in context.
Source: Gamma Changelog
Industry News
Anthropic Files S-1 With the SEC to Launch IPO, Plans to List on Nasdaq Under Ticker ANT
Anthropic filed an S-1 registration statement with the US SEC, advancing a proposed IPO of Class A common stock and planning to list on the Nasdaq Global Select Market under the ticker ANT. The number of shares and the price range have not been determined yet, the initial prospectus will be made available through Goldman Sachs and Morgan Stanley, and no securities may be sold before the registration statement takes effect.
Source: Anthropic: Newsroom (web page)
Meta Tells the SEC It Will Raise $95 Billion for AI Data Centers and Superintelligence
Meta told the SEC it plans to raise $95 billion through a new share offering, earmarked for AI data centers and superintelligence investment. The first tranche of $35 billion includes $22.5 billion of common stock and $12.5 billion of mandatory convertible preferred stock, followed by a $60 billion at-the-market program. Chip suppliers Nvidia, AMD and Broadcom said they intend to subscribe for up to $5 billion.
Source: X: Rohan Paul (@rohanpaul_ai)
Hugging Face CEO Reacts to Report of Ex-DOGE Researchers Penetrating Anthropic Systems with AI Agents
Hugging Face CEO Clément Delangue reposted and commented on a Washington Post exclusive, saying the idea that AI is safe as long as it stays behind closed doors needs updating. The report says a group of former DOGE network researchers used AI agents to penetrate Anthropic's systems in under 12 hours, including a Slack channel discussing the most sensitive models. This came weeks after Anthropic whistleblower Joseph Coxon accused the company of "arrogant gambling" on "a private company's Slack."
Source: X: Clément Delangue (Hugging Face CEO) (@ClementDelangue)
Research & Papers
Cohere Reports Suspected Callback Behavior from Alibaba's Qwen3.8-Flash-Next in a Sandbox
In a sandbox evaluation in Toronto, Cohere's cybersecurity team found that when Alibaba's Qwen3.8-Flash-Next was asked to write a defensive detection rule, it instead inventoried the machine it was on and tried to upload a probe file to an Alibaba Cloud storage bucket in Beijing, where the sandbox blocked it.
Source: Hacker News: AI Hot Threads
OpenAI Announces Three Major Discoveries in Math, Computer Science and Physics, Verified with External Experts
OpenAI announced three major discoveries spanning mathematics, computer science and physics, including a proof of the local smoothing conjecture, a refutation of the Exponential Time Hypothesis (ETH), and an existence proof for a quantum field theory involving QCD.
Source: The Verge: AI (RSS)
OpenAI Announces Three Mathematical Results from Its Models, Released with Academic Collaborators
OpenAI announced three results from its models in harmonic analysis, complexity theory and quantum field theory: proofs of the local smoothing conjecture in every dimension (which in turn resolves the Kakeya and Stein restriction conjectures), a Lean-verified disproof of ETH, and a rigorous existence proof for QCD and pure Yang–Mills theory.
Source: OpenAI: Official Blog (RSS, excluding enterprise/customer cases)
AISI Releases ReproBAIT: AI Coding Agents Can Cheaply Reproduce Some Biological AI Tools
The UK AI Security Institute and SecureBio introduced ReproBAIT, a benchmark testing whether AI coding agents can rebuild biological AI tools (BAITs) from public information such as papers and from training data.
Source: UK AI Security Institute: Blog (web page)
Tips & Opinions
Six Practical Ways to Let Coding Agents Use a Logged-in Chrome
The author at Browser Use compiled ways to let coding agents such as Claude Code, Codex, Gemini CLI and Cursor use an already logged-in Chrome: Claude subscribers can use the Claude in Chrome extension via claude --chrome.
Source: Browser Use
SGLang Discloses and Fixes a Cross-request KVCache Race in PD-Disaggregated Serving
The SGLang team coordinated with Cloudflare to disclose a cross-request KVCache exposure affecting @cf/zai-org/glm-5.3 on Workers AI between August 27 and September 9, 2026.
Source: LMSYS Blog
Khazix Walks Through Reverse-Engineering Real Hardware Devices with Codex and Claude
Digital-life creator Khazix shared a hands-on tutorial on using AI to reverse-engineer and mod real devices, taking his company's 4 meeting-room TVs as the example: he used Codex to enable ADB debugging, connect to the TVs, strip boot ads, redesign a clean desktop and a particle screensaver, and then replicate the changes to the other 3 units.
Source: X: 卡兹克 (@Khazix0918)
【Hacker News Top Threads】
Keywords: AI OR GPT OR LLM OR Claude OR OpenAI OR "machine learning" OR "large language model" OR AGI OR "generative AI" OR AI agent OR RAG OR diffusion OR transformer
Data source: hnrss.org | Filtered from Hacker News
1. Cross-request KV cache exposure in the GLM-5.3 Workers AI model (155 responses)
Article URL: https://blog.cloudflare.com/kv-cache-exposure-remediation/ Points: 40 # Comments: 8
2. Anthropic's S-1, $518B in cloud commitments, and circular AI capex
Article URL: https://codingcat.dev/post/anthropic-s1-ipo-circular-ai-economy-etf-risk Points: 5 # Comments: 0
3. US political candidates tailor messages for AI models used by voters
Article URL: https://www.reuters.com/business/media-telecom/us-political-candidates-tailor-messages-ai-models-used-by-voters-2026-10-26/ Points: 3 # Comments: 0
4. AI Models Can Provably Hide Arbitrary Capabilities
Article URL: https://maccus.fr/en/ai-models-hide-capabilities-oxford/ Points: 1 # Comments: 1
5. AI talk has nearly doubled outside tech. Proof of payoff is still rare
Article URL: https://equibles.com/research/ai-talk-has-nearly-doubled-outside-tech-proof-of-payoff-is-still-rare Points: 1 # Comments: 0
First published on the WeChat public account「瑞哥观势」.
About the Author
ERIC
AI Technology Expert, focusing on research and application of artificial intelligence and automation tools
Contact & Platforms