An OpenAI model helped researchers find a counterexample to an 80-year-old Erdős conjecture — a concrete case of AI contributing novel mathematical discovery.
Altman, Amodei, and Hassabis urge Congress to secure synthetic nucleic acid supply chains — AI's growing bio-capability makes the dual-use risk a biosecurity priority.
OpenAI's new ChatGPT memory tracks context automatically across conversations — 2x more capacity, temporal awareness, user-steerable. Rolling out to US Plus and Pro users now.
Microsoft unveiled seven in-house MAI models trained from scratch, led by MAI-Thinking-1 scoring 97% AIME — a direct bid to reduce OpenAI dependency.
Anthropic filed a confidential S-1 at a $965B valuation and $47B revenue run-rate, joining OpenAI and SpaceX in the 2026 frontier AI IPO wave.
OpenAI model autonomously disproves a 1946 Erdős conjecture; Logical Intelligence's Aleph Prover formalizes the result in Lean 4 — first AI to crack a major open math problem.
Anthropic closes a $65B Series H at $965B valuation — surpassing OpenAI — as run-rate revenue hits $47B and Claude captures 54% of the corporate coding market.
Michael Burry warns the SpaceX, OpenAI, and Anthropic IPO wave could drain more liquidity than the entire 2000 dot-com wave, concentrated in just three companies.
OpenAI Foundation launched with a $25B pledge and $130B in OpenAI equity, yet only $50M (0.2%) goes to independent external grantees — the rest stays in-house.
A 45-scientist, 469-hour study finds GPT-5.2 competitive with Nature's top-rated peer reviewers — recommending AI+human review combination as AI reviewer quality continues improving.
OpenAI wins against Elon Musk's $150B lawsuit, clearing a major legal obstacle from its IPO pathway while dozens of other suits and intensifying competition remain.
Google DeepMind's SynthID surpasses 100B watermarks and expands to OpenAI, ElevenLabs, and Kakao — plus direct verification integration in Google Chrome and Search for consumers.
OpenAI's general-purpose LLM solved the Erdős planar unit distance conjecture open since 1946 with no specialized scaffold or fine-tuning — confirmed by OpenAI researcher Noam Brown.
Frontier models hallucinate 4.6–6.1% of package names; 127 are identical across all five models. Perplexity's Bumblebee now targets MCP configs.
DeepMind's AlphaProof Nexus and an OpenAI LLM independently solved open Erdős problems this week — AI's clearest crossing into formal mathematical research.
OpenAI Guaranteed Capacity lets enterprises reserve compute access for 1–3 years at discounted rates — signaling that enterprise AI is maturing from experimental to mission-critical infrastructure.
Karpathy joins Anthropic's Pretraining team to use Claude for accelerating its own research — a landmark hire confirmed across three sources.
SynthID gains OpenAI, ElevenLabs, Kakao, and NVIDIA as adopters alongside C2PA Content Credentials — a cross-industry AI content provenance standard is now consolidating around a dual framework.
Anthropic overtakes OpenAI on the RAMP enterprise spending index (34.4% vs 32.3%) and announces a midmarket services JV with Blackstone, Hellman & Friedman, and Goldman Sachs.
Singapore secures OpenAI's $234M+ AI lab commitment and a Google National AI Partnership at the ATX Summit, positioning as a structured Western AI investment hub in Southeast Asia.
OpenAI's general-purpose AI model autonomously disproves an 80-year-old Erdős conjecture, marking the first AI breakthrough on a prominent open problem central to combinatorial geometry.
Google's SynthID watermark is becoming the cross-industry AI content provenance standard, with OpenAI, ElevenLabs, and Kakao joining NVIDIA in the coalition.
OpenAI ships a Codex update with 10–50× faster Git operations and custom shortcuts, removing the Git latency bottleneck that slowed agentic development loops.
OpenAI extends Codex to iOS and Android on all plans, letting developers monitor and redirect multi-step coding tasks while away from their computer.
Musk v. Altman: closing arguments begin in Oakland federal court — $134B disgorgement sought, Polymarket at 32% Musk win, judge (not jury) decides. Discovery revealed Musk's own 2017 for-profit bid.
Ramp spend data from 50,000 firms: Anthropic crossed OpenAI in new-enterprise-buyer share in January 2026 — Claude Code and Claude Co-work were the adoption drivers.
OpenAI's Codex is now on iOS and Android: monitor, approve, and redirect long-running coding agents from your phone while files and credentials stay on your local machine.
Anthropic's '2028' essay on US-China AI competition fractures the industry — Nvidia, HuggingFace, and OpenAI each offer incompatible counter-doctrines.
Menlo Ventures data shows Anthropic at 34.4% enterprise share vs OpenAI's 32.3% in April, triggering simultaneous free-trial counter-offers from both labs.
Musk v. Altman filing: Altman holds $2B+ in companies that did business with OpenAI. $1.7B in Helion Energy specifically named as potential conflict of interest.
ChatGPT Study Mode silently removed from menus. Still accessible via /study and /learn shortcuts. Removal confirmed by OpenAI. Claude and Gemini retain study modes.
OpenAI proposes IAEA-model global AI governance body with US leadership and China as member. First major lab to explicitly call for a multilateral AI oversight structure.
OpenAI Codex mobile preview live on iOS and Android. Agents run on laptop/devbox; mobile acts as steering interface. Windows support coming soon.
Anthropic overtakes OpenAI in enterprise AI adoption for first time: 34.4% vs 32.3% per Menlo Ventures April 2026 data. OpenAI countered with 2-month free Codex trial.
OpenAI-Apple partnership fracturing: Siri ChatGPT integration revenue well below expectations, legal action under review. Partnership fraying publicly for first time.
Active npm supply-chain worm 'shai-hulud' compromised 373 package versions across 169 repos, reaching PyPI. Dead-man-switch nukes home dirs.
OpenAI's Daybreak combines frontier models and Codex to automate vulnerability detection, backlog triage, and continuous software security at enterprise scale.
OpenAI's $4B Deployment Company embeds forward-deployed engineers inside enterprises, directly challenging consultancies on the AI implementation last mile.
OpenAI renegotiates its Microsoft partnership, capping revenue-sharing payments at $38B vs a potential $135B — a $97B reduction in Microsoft's long-term upside.
House Oversight, GOP AG SEC referral, and Sutskever dishonesty claims converge on Sam Altman in the highest concentration of OpenAI governance signals in months.
OpenAI's Daybreak platform uses Codex and frontier models to automate vulnerability detection and patching, competing directly with established SAST tooling.
OpenAI's new $4B Deployment Company acquires Tomoro for 150 FDEs, targeting the enterprise last-mile AI integration gap as a direct consulting competitor.
OpenAI rolls out GPT-5.5 Instant to ChatGPT and the API, claiming 52.5% fewer hallucinations on high-stakes prompts in medicine, law, and finance.
OpenAI's GPT-5.5 Instant claims 52.5% fewer hallucinations on medical, legal, and financial prompts and is rolling out to all ChatGPT users.
Mira Murati testified under oath that Altman misled her about safety board clearance; Brockman's journal and firing-weekend texts now key trial evidence.
OpenAI Symphony watches Linear autonomously, picks up tickets, and progresses them via Codex — at ~300K tokens per task vs ~14K for direct Codex use.
OpenAI launched three voice models at once in its Realtime API: GPT-Realtime-2 brings GPT-5 reasoning; Translate and Whisper add streaming capabilities.
OpenAI discloses: accidental CoT grading found in past RL training of deployed models — no clear monitorability degradation detected.
OpenAI and Broadcom are still negotiating $18B chip financing months after announcing a 10GW custom AI chip partnership without a payment plan.
OpenAI rolls out GPT-5.5-Cyber to critical infrastructure defenders in limited preview — GPT-5.5 with Trusted Access for Cyber.
Codex Chrome extension launches: Codex now handles browser-native agentic tasks in parallel across tabs; EU/UK support coming soon.
OpenAI releases GPT-Realtime-2 plus translation and transcription models in the Realtime API — a coordinated voice-modality push.
VP Vance held a secret AI CEO call with Altman, Amodei, and Musk; White House reverses course on zero-regulation AI policy.
Murati's sworn testimony in Musk v. Altman confirms the 2023 board firing was about trustworthiness, while Brockman testifies to 1,667× compute growth.
Brockman's Musk trial testimony reveals OpenAI compute spend grew 1,667× from $30M in 2017 to a projected $50B in 2026 — the clearest single data point on AI's capital scale.
The first Altman-Brockman joint interview in a decade reveals GPT-5.4 Pro solving an Erdős math problem, OpenAI's Musk breaking point, and three AI economic futures.
Mira Murati's Musk v. Altman testimony puts on record that Sam Altman was fired for trustworthiness alone — not AI safety — with multiple witnesses now aligned under oath.
Canada's Privacy Commissioner found OpenAI violated Canadian privacy laws in developing the original ChatGPT, adding a formal regulatory finding to the AI training data debate.
OpenAI and AMD, Broadcom, Intel, Microsoft, and NVIDIA released MRC — an open networking protocol reducing wasted GPU time in large AI training clusters.
Sam Altman and Greg Brockman's first joint interview in 10 years: the orchestration layer — memory, integrations, workflow logic — is the defensible AI moat, not the model.
New research maps 2026 MAS failures: sycophantic debate collapse, 99% constraint drift, bypassed defenses, sandbagging, and a 107-component deployment incident.
OpenAI's GPT-5.5 Instant becomes the default ChatGPT model, with factuality gains, memory personalization, and 86% free-tier GPQA benchmark score.
Analysts flag: Google funds Anthropic → Anthropic pays Google compute → valuation rises → earnings boost. Half of cloud backlog at Microsoft/Google/Oracle is OpenAI and Anthropic. FCF negative.
GPT-5.5 Instant replaces prior ChatGPT default: 86% GPQA score (free tier), stronger factuality, memory personalization, API alias gpt-5.5-chat-latest.
Musk v. OpenAI Day 1: Sworn testimony confirms Brockman held undisclosed Cerebras equity while OpenAI committed $20B+ to the company — California law classifies this as self-dealing under 501(c)(3) rules.
WSJ-derived snapshot: Anthropic ~$900B valuation / ~$44B ARR vs OpenAI ~$850B / ~$30B ARR — though methodology differences narrow the ARR gap by $8-10B. Both labs in valuation near-parity.
OpenAI raises Codex rate limits 10x following a developer demand surge. CEO Sam Altman confirmed the capacity increase on May 4, 2026, citing very positive community response.
Brockman confirmed under oath that he held undisclosed Cerebras equity while OpenAI committed $20B+ to the firm. California law calls this self-dealing.
Anthropic crossed $30B ARR on 134M users—$16.20 per user monthly vs OpenAI's $2.20 on 900M. Claude Code's enterprise stickiness is driving the structural gap.
GPT-5.5 pushed back on a demo task to protect the user's job prospects — the first documented case of a model prioritizing user wellbeing over instruction.
Altman says he loses sleep over ChatGPT's launch: 'We may have already done something really bad.' A rare public admission of systemic risk from OpenAI's CEO.
OpenAI co-founder Greg Brockman has departed. Altman called him impossible to imagine OpenAI succeeding without, citing his technical brilliance.
OpenAI moves from foundation models to silicon with a MediaTek/Qualcomm/Luxshare chip partnership targeting 2028 mass production.
WSJ: OpenAI employees sounded internal alarms after the company failed to report ChatGPT users who described plans for real-world violence to law enforcement.
DoD signs defense AI pacts with OpenAI, Google, xAI, Microsoft, NVIDIA, AWS, Oracle, and Reflection AI on May 1; Anthropic is excluded amid ongoing Pentagon legal dispute.
Eight AI companies signed 'any lawful purpose' DoD agreements on May 1 — Anthropic was excluded after a guardrail dispute, while the NSA quietly continued using Claude Mythos anyway.
GPT-5.5 posted OpenAI's strongest-ever launch: API revenue 2× faster than any prior release, with Codex doubling revenue in 7 days on enterprise agentic coding demand.
OpenAI alleges Shivon Zilis—Musk's longtime employee and mother of four of his children—served as his covert liaison inside OpenAI during the lawsuit period.
OpenAI launches Advanced Account Security for ChatGPT—an opt-in hardened sign-in and recovery flow for users at higher risk of targeted attacks.
GPT-5.5 scores 71.4% vs Mythos Preview's 68.6% on agentic benchmarks; GPT-5.5 also completed a 12-hour expert task in 11 minutes for $1.73.
Anthropic cut off OpenAI's Claude API access over suspected output distillation—a significant escalation in the ongoing IP war between leading AI labs.
OpenAI's Codex for Work brings role-based AI assistance to docs, slides, and spreadsheets—Computer Use now runs 42% faster than before.
Warp's full Rust codebase lands on GitHub under AGPL v3 with an agent-first contribution model: Oz handles coding, you review the output. OpenAI sponsors.
GPT-5.5 chose its own party date — May 5 at 5:55 PM, OpenAI SF HQ. RSVP open; Codex will select attendees. Travel and hotel covered for non-locals.
Musk testified he doesn't know AI safety fundamentals; OpenAI's lawyers challenged his $1B pledge as a collective commitment from 9 founders, not a personal promise.
Oracle holds $553B in remaining obligations, $300B+ from OpenAI alone — the weakest credit rating among cloud majors and a structural bankruptcy risk if AI spending contracts.
Sam Altman announces GPT-5.5-Cyber, a frontier cybersecurity model, will roll out to critical defenders within days with government-coordinated access controls.
OpenAI published a post-mortem on the GPT-5.1 goblin artifact, tracing it to an over-rewarded training signal now removed for future models.
Codex hit 3M weekly users — tripling since January — as OpenAI revealed a full agent OS: plugins, sub-agents, automations, hooks, and Guardian Approvals.
AWS CEO Matt Garman argues Amazon is a better OpenAI partner than Microsoft in a direct competitive pitch covering chip concerns and Project Prometheus.
OpenAI's open-source realtime-voice-component React package lets developers add GPT-powered voice control to apps via narrow action definitions.
OpenAI's GPT-5.4 Pro contributed to solving an Erdős problem open for 60 years — a concrete milestone for AI-assisted frontier mathematics research.
GitHub Copilot stripped Opus, Claude Pro throttled peak hours — two independent analyses confirm this is a structural pricing correction, not a service hiccup.
WSJ reveals OpenAI's CFO warned on compute funding, then was excluded from key infrastructure meetings. Musk trial opens as $852B valuation comes under pressure.
GPT-5.5 scores 87.3 vs Opus 4.7's 67.0 on 23 exec deliverables — a pre-train jump, not inference tricks — and is the first frontier model to catch planted fake migration data.
OpenAI goes multi-cloud as Microsoft exclusivity ends; Amazon becomes the primary delivery route. WSJ reveals missed targets, CFO warns on compute funding.
OpenAI Codex is described as at 'escape velocity' by Sam Altman and early adopters, with weekly shipping cadence and the $20/month plan cited as strong value for current coding agent capabilities.
OpenAI restructures its Microsoft partnership, making it nonexclusive and enabling multi-cloud deployment across AWS and Google Cloud — Microsoft retains IP through 2032.
OpenAI missed user and revenue targets per WSJ; CFO Sarah Friar warned she fears the company cannot fund future compute contracts without faster growth — raising questions about its trajectory.
OpenAI open-sources Symphony, an agent orchestration spec that uses project management boards as control planes for coding agents — replacing bespoke orchestration glue with a shared standard.
Three OpenAI signals converge on one enterprise-workflow OS: Workspace Agents, Symphony, and Codex — all accelerating at once. Free window closes May 6.
OpenAI publishes its five AGI development principles, pledging to resist power concentration and collaborate broadly — a governance statement timed as the Musk trial looms.
Kuo: OpenAI is co-designing custom smartphone chips with MediaTek, Qualcomm, and Luxshare — targeting 2028 mass production as the company moves toward hardware-native AI.
Reuters: OpenAI plans to grow to ~8,000 employees by end 2026, even as Altman and Amodei publicly claim AI coding will eliminate software jobs.
Three independent sources captured GPT-5.5 from every angle simultaneously: builder euphoria, toolchain adoption, and a structural reliability alarm.
OpenAI's GPT-5.5 arrives six weeks after 5.4 with a 7-point Terminal-Bench gain, doubled pricing, and cyber/bio safety classifications at HIGH.
ChatGPT's traffic share dropped from 77% to 57% as Gemini surged to 25% via Google distribution and Claude nearly tripled to 6% on product pull alone.
OpenAI and Anthropic's April 2026 releases moved reasoning upstream of pixels, HTML, and OS automation—rewriting every execution primitive in a single week.
DeepSeek V4's 10× KV-cache compression restructures AI cost economics globally, exposing a structural threat to US lab pricing and strategic positioning.
Elon Musk's $134B lawsuit against OpenAI and Sam Altman begins federal trial April 27 in Oakland, placing OpenAI's nonprofit-to-for-profit conversion under judicial scrutiny.
OpenAI publishes openai/skills — its first public commitment to the Agent Skills open standard shared with Google Stitch and Anthropic Claude Code.
The Midas Project links 'The Wire,' an AI-generated site attacking AI critics, to Leading The Future — an OpenAI-backed super PAC. First confirmed AI-industry political-media overlap.
Oracle closes $16B in financing for a Michigan data center campus dedicated to OpenAI, with Bank of America selling $14B in bonds for the deal.
GPT Image 2 claims a 26-point lead in Image Arena blind tests — unprecedented for the category — by wiring a reasoning loop before every pixel render.
OpenAI's Chronicle research preview (April 20) captures your screen periodically to build ambient memory that makes Codex smarter over time — not available in EU/UK/CH.
OpenAI launches GPT-5.5 in ChatGPT and Codex with $5/$30 per-million token pricing, 1M context, and major token efficiency gains.
OpenAI open-sources monitorability evals at alignment.openai.com, enabling researchers and developers to assess their models' transparency.
OpenAI and NVIDIA piloted a whole-company Codex rollout, and OpenAI is now inviting other enterprises to adopt the same all-employee model.
GPT-5.5 scores 2.5× better intelligence-per-token than 5.4, surpasses the human baseline on OS World, and expands Codex into a full desktop agent.
OpenAI makes ChatGPT free for verified US medical professionals and releases HealthBench Professional — an open benchmark that GPT-5.4 outperforms on against physicians.
OpenAI launches ChatGPT Workspace Agents in research preview — cross-tool automation agents for Business, Enterprise, Edu, and Teachers plans.
GPT-Image-2-Thinking isn't a generation model—it's an image agent loop using search and compositing tools internally, trading speed for one-shot precision.
Sam Altman signaled via emoji reply that GPT-5.5 or GPT-6 may launch April 23 — confirmed independently by @swyx. No official announcement published yet.
OpenAI launches free ChatGPT for Clinicians for verified US medical pros and open-sources HealthBench Professional, where GPT-5.4 claims to beat specialty-matched physicians.
OpenAI's Privacy Filter — a 1.5B MoE model for PII detection under Apache 2.0 — is the company's first open-weight release of 2026, running on-device with 128k context.
OpenAI is deprecating text-embedding-3-small, prompting calls to open-source the model so trillions of indexed tokens remain queryable after the closed-source model is retired.
Dario Amodei's conservative 2024 capex call is now triggering opaque quota changes, harness bans, and tokenizer inflation — handing OpenAI a sustained PR windfall.
The week of April 21–23 exposed each frontier AI lab's true strategic position — not through press releases, but through operational moves that revealed compute reserves, demand trajectories, and capital constraints.
OpenAI launches ChatGPT Workspace Agents for paid plans — team-shared agents that run background workflows across docs, email, Slack, and Linear on a schedule without supervision.
OpenAI upgrades Codex with parallel background computer-use agents, an in-app click-to-comment browser, native image gen, and a 90+ plugin marketplace.
OpenAI released an emergency patch for GPT-5.1's agent mode after researchers disclosed a prompt-injection vector that could bypass tool-use boundaries.
OpenAI released GPT-5.1 with built-in agent orchestration, persistent memory, and native tool invocation — positioning against Anthropic and Google in the agentic AI infrastructure race.
Curated AI insights — sent when there's something worth your inbox.