Skip to content

#OpenAI

14 today

May 30Saturday

AI HOT (Curated Pool)

Codex now supports computer use on Windows

OpenAI added Windows computer-use support for Codex, letting users start, review, and guide tasks on a Windows PC through the ChatGPT mobile app; the post states this is an early experience and does not disclose pricing or rollout scope.

Why it matters: OpenAI added Windows computer-use to Codex, so tasks on a PC can be started and directed from the ChatGPT phone app, now in early access.

Bloomberg Technology

Anthropic Valuation of $965 Billion Passes OpenAI

Bloomberg’s title says Anthropic reached a $965 billion valuation and passed OpenAI; the post does not disclose the valuation source, funding round, deal terms, or the comparison basis for OpenAI.

Why it matters: Bloomberg's headline puts Anthropic above OpenAI at a $965 billion valuation, but gives no source or comparison basis, so this is a capital signal, not a verifiable valuation report.

Bloomberg Technology

Anthropic Raises at $965 Billion Valuation, Eclipsing OpenAI

Anthropic raised $65 billion at a $965 billion post-money valuation, surpassing OpenAI’s value for the first time; Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital led the round, and each lead investor put in more than $2 billion.

Why it matters: Anthropic closed a $65 billion round at a $965 billion post-money valuation, passing OpenAI on the number, led by four firms including Sequoia Capital.

May 29Friday

New York Times Chinese

Anthropic Tops OpenAI Valuation to Become the Most Valuable AI Startup

Anthropic raised $65 billion at a $900 billion pre-money valuation, above OpenAI’s last $730 billion valuation. Claude Opus 4.8 also scored 10% higher than Anthropic’s previous model on Vals AI’s vibe-coding benchmark.

Why it matters: Anthropic's new $65 billion round puts its pre-money valuation at $900 billion, above OpenAI's $730 billion from its last round, reshuffling the top of the foundation model table.

Xinzhiyuan · WeChat

Claude Opus 4.8 tests split users: strong at high effort, costly under rate limits

The article says Claude Opus 4.8 scores 63 on an Extra-High senior engineering benchmark, 30 points above Opus 4.7, but drops to 42 at High effort, while $200/month Max users report hitting rate limits within hours on complex agent tasks.

Why it matters: Xinzhiyuan's tests show Claude Opus 4.8 scoring 30 points above the previous generation at the Extra-High senior engineer level, while High-intensity tasks drop to 42.

AI HOT (Curated Pool)

Strengthening Societal Resilience with Rosalind Biodefense

OpenAI launched Rosalind Biodefense and provides trusted GPT-Rosalind access to vetted developers and U.S. government partners; the post does not disclose model parameters or pricing.

Why it matters: OpenAI is giving vetted developers and US government partners access to GPT-Rosalind, a model aimed at life sciences work like outbreak modeling and early screening, so the vetting process is the part to watch.

Latent Space

Anthropic raises $65B Series H, releases Opus 4.8 and Dynamic Workflows

Anthropic announced a $65B Series H at a $965B post-money valuation, disclosed a $47B revenue run rate, and released Claude Opus 4.8 plus Claude Code Dynamic Workflows as a research preview for parallel subagent orchestration.

Why it matters: Anthropic closed a $65B Series H at a $965B post-money valuation and said annualized revenue grew from $9B in December to more than $47B.

Ruan YiFeng's Weblog

Technology Enthusiasts Weekly Issue 398: Token Costs Are Hard to Afford

Peter Steinberger posted one month of usage showing 7.6 million requests and 603 billion tokens, with CodexBar estimating a $1.3 million value under preset rates rather than his actual spend as an OpenAI employee.

Why it matters: An OpenAI employee posted a month of usage at 7.6 million requests and 603 billion tokens, about $1.3M at public rates, showing what a year of unrestricted top-model use can cost per person.

AI HOT (Curated Pool)

Skill distillation

Skill distillation has Opus 4.7, GPT-5.1, and Gemini 3 Pro write standardized SKILL.md procedure files, while local Qwen 35B and Gemma 26B models execute those files step by step.

Why it matters: A practical personal workflow: big models write the SKILL.md steps, then a local Qwen 35B or Gemma 26B runs them one by one.

Financial Times · Technology

Anthropic finalises $65bn funding deal to surpass OpenAI’s valuation

Anthropic finalised a $65bn funding deal that values the Claude AI maker at $965bn including the new money, taking its valuation above OpenAI’s, according to the RSS snippet.

Why it matters: Anthropic closed a $65bn round at a $96.5bn valuation including new capital, passing OpenAI and giving the capital ranking a concrete number.

Bloomberg Technology

Anthropic Raises $65 Billion in Funding Round, Eclipses OpenAI

Anthropic raised $65 billion at a $965 billion post-money valuation, and Bloomberg says the round put its value above OpenAI for the first time.

Why it matters: Anthropic raised $65bn at a $96.5bn post-money valuation, passing OpenAI and shifting the capital ranking of the two frontier labs.

Bloomberg Technology

Anthropic Eclipses OpenAI With Valuation of $965 Billion

Anthropic raised $65 billion at a $965 billion post-money valuation, surpassing OpenAI’s valuation for the first time; the RSS snippet does not disclose investors, deal terms, or a timeline.

Why it matters: HKR-H/K/R all pass: Bloomberg reports a $65B raise at a $965B valuation, putting Anthropic above OpenAI. Investors and terms are not disclosed, but the scale makes it same-day top AI funding news.

May 28Thursday

AI HOT (Curated Pool)

OpenAI Frontier Governance Framework

OpenAI published its Frontier Governance Framework to align its AI safety, security, and risk management practices with new EU and California regulations; the post does not disclose specific evaluation metrics, implementation timelines, or the list of covered frontier models.

Why it matters: OpenAI publishes documents mapping its safety measures to new California and EU rules across risk assessment, model reporting and incident response, showing a top lab's compliance path.

AI HOT (Curated Pool)

Cognition AI raises over $1B, targets 10x software engineering productivity

Cognition AI raised over $1 billion at a $26 billion pre-money valuation, while annualized revenue grew from $37 million to about $492 million in one year, and Devin is positioned as an autonomous junior engineer that can plan, test, and deploy through multi-step workflows.

Why it matters: Cognition AI raised over $1 billion at a $26 billion pre-money valuation, with annualized revenue climbing from $37 million to about $492 million in a year.

AI HOT (Curated Pool)

OpenAI Products Support Secure Connections to Private MCP Servers

OpenAI supports ChatGPT, Codex, and the Responses API connecting to internal MCP servers through outbound-only HTTPS, while teams keep those servers inside private networks.

Why it matters: OpenAI added outbound-only HTTPS access to internal MCP servers for ChatGPT, Codex and the Responses API, so servers never need a public endpoint.

AI HOT (Curated Pool)

I Think Anthropic and OpenAI Found Product-Market Fit

Anthropic and OpenAI changed enterprise pricing around April 2026, moving coding agents from heavily discounted seat plans to API-usage billing, with Anthropic Enterprise at $20 per seat per month plus API fees and OpenAI Codex billed by API token usage.

Why it matters: Anthropic and OpenAI are moving enterprise coding assistants from flat seat fees to API usage pricing from April 2026, matching direct token rates, which changes how companies budget and buy.

May 27Wednesday

The Verge · AI

AI tried to bury this politician — now people have actually heard of him

Leading the Future, a super PAC funded by OpenAI, Palantir, and a16z executives, has spent millions against NY-12 candidate Alex Bores since late 2025; the snippet says Anthropic and OpenAI will spend millions before the June Democratic primary over who regulates AI and who faces political costs for trying.

Why it matters: OpenAI and Anthropic each spent millions through super PACs in New York's 12th district primary to stop one candidate, and the effort raised his profile instead.

AI HOT (Curated Pool)

Runway launches Model Context Protocol server

Runway launched an MCP server that lets compatible agents such as Claude, ChatGPT, and Cursor generate images and videos inside chat interfaces, with access to Gen-4.5, Seedance 2.0, GPT Image 2, Kling 3.0, and Nano Banana Pro.

Why it matters: Runway put its image generation and video editing inside Claude, ChatGPT and Cursor chat boxes, so users send a link, reference image or text prompt and get a finished result without switching apps.

QbitAI · WeChat

7B Medical AI Agent Beats o3 and GPT-5 by Learning Where and How to Look

Shanghai Innovation Institute’s LeapQuest and three universities released Ophiuchus and MedScope, applying Think with Images and Think with Videos to medical AI; Ophiuchus-7B scored 68.0 on eight VQA benchmarks, above OpenAI-o3 at 62.2, Gemini 2.5 Pro at 61.8, and GPT-5 at 59.9.

Why it matters: Ophiuchus-7B scores 68.0 across 8 medical visual QA benchmarks, beating o3 and GPT-5, a sample of a small model catching general models through method design.

New York Times Chinese

How Google Rebounded and Started Winning the AI Race

Google said regular Gemini users more than doubled in one year to 900 million, while ad revenue rose 16% to $77 billion last quarter, and its Siri partnership with Apple will place Gemini inside future iPhone assistant features.

Why it matters: From AI telling people to eat rocks two years ago to Gemini's 900 million monthly users matching OpenAI's self-reported ChatGPT figure, this shift is a concrete sample of the model race.

Computing Life · Yage

Using AI Better, Step Two: Write the Skill Before Execution

The author proposes writing a Skill before asking AI to execute a task; each Skill should include three elements—success criteria, observed pitfalls, and deterministic tools—and can be organized through index.md plus AGENTS.md or CLAUDE.md for reuse.

Why it matters: The author suggests writing a Skill document before handing work to AI, externalizing success criteria, past pitfalls and fixed tools, giving Claude Code and similar tool users a reusable way to accumulate experience.

AI HOT (Curated Pool)

Claude Mythos reportedly solves OpenAI’s landmark Erdős problem with a “cute simple proof”

Anthropic engineer Sholto Douglas said Claude Mythos solved OpenAI’s Erdős unit distance conjecture problem over the weekend and produced a “cute simple proof”; the RSS snippet does not disclose the proof, verification process, or benchmark setup.

Why it matters: Anthropic engineer Sholto Douglas says Claude Mythos solved the Erdős unit distance conjecture with a clever, compact proof, right after OpenAI treated the problem as a milestone for AI math reasoning.

May 26Tuesday

AI HOT (Curated Pool)

SynthID watermarking expands partnerships, covering over 100 billion content items

Google DeepMind says SynthID has watermarked more than 100 billion content items and is being integrated into models from OpenAI, ElevenLabs, and Kakao, extending prior industry work with NVIDIA.

Why it matters: SynthID has watermarked over 100 billion pieces of content, and now covers models from OpenAI, ElevenLabs and Kakao, moving AI content provenance from one vendor to cross-platform work.

Xinzhiyuan · WeChat

OpenAI Nearly Collapsed? President Says He Resigned the Day Altman Was Ousted

Greg Brockman recounted OpenAI’s 72-hour crisis: on November 17, 2023, the board removed Sam Altman as CEO and took Brockman off the board, after which Brockman resigned the same day and said he initially put the chance of taking the company back at 10%.

Why it matters: OpenAI president Greg Brockman recounts the 72 hours around Altman's November 2023 ouster, when he resigned that day and put the odds of retaking the company at one in ten.

New York Times Chinese

Pope Leo XIV Challenges Silicon Valley and Warns of AI Risks

Pope Leo XIV issued the 42,300-word encyclical Magnifica Humanitas, warning that AI amplifies the power of people with economic resources, expertise, and data access, and calling for regulation and transparency.

Why it matters: Pope Leo XIV's 42,300-word encyclical warns that if a few people embed their own morality into AI's underlying rules, power concentrates further among those who hold the money, technology and data.

AI HOT (Curated Pool)

OpenAI GPT-5.6 Reportedly Set for Next Month With 1.5M-Token Context

Developers found an unannounced OpenAI GPT-5.6 entry in Codex backend logs under the codename iris-alpha, with a 1.5 million-token context window, about 43% higher than GPT-5.5’s 1.05 million-token limit.

Why it matters: Developers dug an undisclosed model, iris-alpha, out of OpenAI Codex backend logs; it maps to GPT-5.6 and its 1.5M-token context window is about 43% above the 1.05M in the GPT-5.5 API.

May 24Sunday

AI HOT (Curated Pool)

Greg Brockman: The 72 Hours That Nearly Destroyed OpenAI

The title says Greg Brockman discusses the 72 hours that nearly destroyed OpenAI, but the post does not disclose the timeline, participants, or specific mechanisms behind the crisis.

Why it matters: OpenAI co-founder Brockman detailed the 72 hours after Sam Altman's firing, including his own resignation and a backup company plan, and named an Ilya Sutskever tweet as the turning point.

May 23Saturday

AI HOT (Curated Pool)

Anthropic reportedly nears over $30B funding round, with valuation set to overtake OpenAI

Bloomberg reports that Anthropic is nearing a funding round of over $30 billion, expected to close as soon as next week, pushing its valuation above $900 billion, while the company projects second-quarter revenue of $10.9 billion and its first profitable quarter.

Why it matters: Anthropic's new round is expected to top $30 billion at a valuation above $900 billion, with Q2 revenue of $10.9 billion and a path to quarterly profit, which could put it ahead of OpenAI.

Bloomberg Technology

Anthropic to Close Over $30 Billion Round as Soon as Next Week

Anthropic plans to close a funding round of over $30 billion as soon as next week at a valuation above $900 billion, Bloomberg reported, citing people familiar with the matter, which would put it ahead of OpenAI as the world’s most valuable AI startup.

Why it matters: Anthropic is set to close a round above $30 billion at a post-money valuation over $900 billion, which would top OpenAI.

May 22Friday

MIT Technology Review · AI

Google I/O showed how the path for AI-driven science is shifting

MIT Technology Review says Google used I/O to shift its scientific AI framing toward Gemini for Science, a package that groups AI Co-Scientist and AlphaEvolve, while researchers can now apply for access and older specialized systems like AlphaFold and WeatherNext remain active.

Why it matters: At Google I/O, DeepMind showed WeatherNext giving earlier hurricane warnings while shifting the entry point for science AI toward general agents like Gemini for Science, a change in how resources are split.

AI HOT (Curated Pool)

OpenAI Codex /goal Feature Officially Launches with Usage Guide

OpenAI moved Codex /goal mode from experiment to stable release, letting users set milestones in the Codex app, IDE extension, or CLI and keep tasks running for hours or days with progress checks, direction changes, and pause controls.

Why it matters: Codex's /goal mode is now stable, letting AI run tasks for hours or days across app, IDE and CLI.

Computing Life · Share · Yage

A general-purpose AI model refutes an 80-year-old conjecture

An OpenAI general-purpose reasoning model refuted Erdős’s 1946 unit distance conjecture in the plane; the post says the model was not specially trained for mathematics, and Tim Gowers said he would recommend it to Annals of Mathematics.

Why it matters: GPT-5 used algebraic number theory to build a counterexample that disproves Erdős's 1946 planar unit distance conjecture, pushing the count of unit-distance pairs to n^1.014.

AI HOT (Curated Pool)

Codex Enables Secure Cross-Device Mac Control Around the Clock

OpenAI Devs says Codex can use apps on a Mac from a phone while the Mac remains locked and the screen is off; the post does not disclose permission boundaries, pricing, or a release timeline.

Why it matters: An OpenAI developer account says Codex can be called securely from a phone while the Mac is locked, offering a new way to operate across devices remotely.

Bloomberg Technology

SpaceX Files for Nasdaq IPO | Bloomberg Tech 5/21/2026

Bloomberg says SpaceX filed for a Nasdaq IPO and pitched a $28.5 trillion opportunity spanning AI to Mars; the snippet also says OpenAI is preparing an IPO filing that could arrive as soon as Friday.

Why it matters: SpaceX has filed for a Nasdaq IPO and its roadshow pitches stretch from AI to Mars, while OpenAI is reported to file its own IPO paperwork as soon as Friday.

May 21Thursday

MIT Technology Review · AI

Anthropic’s Code with Claude Showed Off Coding’s Future—Whether You Like It or Not

Anthropic used its two-day Code with Claude event in London to show Claude Code automation, with nearly half the room saying they shipped a pull request fully written by Claude in the past week, and many keeping their hands raised when asked whether they had shipped it without reading the code.

Why it matters: At Anthropic's London developer event, a show of hands found nearly half had shipped Claude-written code changes in the past week, most without reading them.

The Verge · AI

Musk v. Altman: Much Ado About Nothing

The jury found Elon Musk’s lawsuit was filed after the statute of limitations had expired, while the case centered on OpenAI’s shift from nonprofit to for-profit status and whether Musk lost money; the post does not disclose damages awarded or next legal steps.

Why it matters: A jury ended Musk's suit accusing Altman and OpenAI of breaking charitable trust in the nonprofit-to-for-profit shift, so the claims were never heard on the merits.

Xinzhiyuan · WeChat

Anthropic Acquires SDK Toolmaker Stainless, Leaving OpenAI and Others to Maintain SDKs

Anthropic has completed its acquisition of Stainless, an SDK generation company used by OpenAI, Anthropic, Meta, Cloudflare, and other infrastructure vendors; Stainless says prior SDK ownership remains with customers, but it will shut down hosted products including SDK generator and stop providing ongoing support.

Why it matters: Anthropic bought SDK tooling vendor Stainless, which will shut its hosted product and drop support for OpenAI, Meta and other former customers.

Latent Space

OpenAI GPT-next Disproves 80-Year-Old Erdős Planar Unit Distance Problem for Under $1000

OpenAI said an internal general-purpose reasoning model disproved the 1946 Erdős planar unit distance problem by finding a new family of constructions; the reasoning summary reportedly spans about 125 pages, while outside observers speculate the run used under 32 hours or under $1,000.

Why it matters: A general OpenAI reasoning model disproved Erdős's 1946 planar unit distance conjecture, producing a roughly 125-page summary; outside observers estimate under 32 hours of runtime and under $1,000 in cost.

Synced · WeChat

Zhipu deploys ZCube, raising inference throughput 15% on the same GPUs

Zhipu deployed ZCube in a thousand-GPU GLM-5.1 production inference cluster, replacing ROFT while keeping GPUs, software stack, and business code unchanged; throughput rose by over 15%, TTFT P99 fell 40.6%, and switch plus optical module costs dropped by one third.

Why it matters: Zhipu swapped in its own ZCube networking on a GLM-5.1 thousand-GPU inference cluster, raising throughput over 15% and cutting P99 first-token latency 40.6% without touching GPUs or business code.

Hacker News front page

OpenAI to Confidentially File for IPO as Soon as Friday

The title says OpenAI will confidentially file for an IPO as soon as Friday; the RSS body only includes the CNBC URL, Hacker News with 41 points and 2 comments, and does not disclose valuation, offering size, or listing timetable.

Why it matters: OpenAI plans to confidentially file a draft IPO prospectus with the SEC as soon as Friday, pinning down where the listing process stands while valuation, size and timing remain undisclosed.