OpenAI / ChatGPT

Everything OpenAI: the GPT models, ChatGPT and Sora, company strategy and people moves.

1,549 picksRelated topicsAnthropic / Claude Google / Gemini Model releases

Latest picks

881–900 of 1,549

Jun 5Friday

AI HOT (Curated Pool)Jun 5, 05:13

Co-Existence and the End of Co-Intelligence

Ethan Mollick announced Co-Existence for an October 20 release and argues that co-intelligence is giving way to autonomous agents, citing late-2025 coding agents that a study links to 17x more code and Anthropic’s claim that AI now writes 80% of its code.

Why it matters: HKR-H/K/R all pass: Ethan Mollick’s essay has authority, a sharp framing, and concrete coding-productivity claims. It stays below 85 because it is commentary plus a book announcement, not a model release or reproducible experiment.

AI HOT (Curated Pool)Jun 5, 03:36

OpenAI API Adds Moderation Scores

OpenAI added moderation scores to the Responses API and Completions API; applications can receive moderation signals in the same generation request and use them for logging, routing, review, or blocking.

Why it matters: HKR-K and HKR-R pass: OpenAI adds moderation scores to generation responses, giving builders a concrete safety-routing mechanism. HKR-H is weak, so this sits at the featured threshold, not a major-release band.

AI HOT (Curated Pool)Jun 5, 02:16

Codex launches iOS app build plugin

Codex integrated the Build iOS Apps plugin, which lets users test iOS apps in an in-app browser, open SwiftUI previews, and hot-reload edits without leaving Codex.

Why it matters: HKR-H/K/R all pass: the hook is Codex handling iOS app testing, with concrete SwiftUI preview and hot reload details. This is a mid-weight OpenAI dev-tool update, not a model release; pricing and rollout scope are not disclosed.

Jun 4Thursday

MIT Technology Review · AIJun 4, 18:50

How courts are coping with a flood of AI-generated lawsuits

MIT and USC researchers examined 4.5 million federal civil cases from 2005 to 2026, finding self-represented lawsuits rose from 11% in 2022 to 16.8% in 2025, while AI-text detector flags in sampled filings increased from 1% in 2023 to 18% in 2026.

Why it matters: MIT Technology Review covers an MIT/USC large-sample study, clearing HKR-H/K/R with 4.5M cases and an 18% AI-text marker rate. It affects public systems, not core model capability, so 78 fits the lower good-quality band.

Synced · WeChatJun 4, 17:52

Office Whispering Is Turning Typing Into an Old Skill

AI dictation tools are moving into developer and office workflows, with Wispr Flow reporting over 2.5 million global downloads, 70% 12-month retention, and 100x annual user growth, while OpenAI’s gpt-4o-transcribe reached a 2.5% word error rate in a third-party evaluation cited by the article.

Why it matters: HKR-H/K/R all pass, but this is a data-backed workflow trend piece, not a model launch or platform update. It sits at the lower featured threshold.

AI HOT (Curated Pool)Jun 4, 17:00

Dreaming: ChatGPT launches stronger memory system to better remember user preferences

ChatGPT launched Dreaming, a memory system for remembering user preferences and keeping context relevant across conversations; the post does not disclose rollout scope, default settings, or retention period.

Why it matters: OpenAI product update with clear HKR-H/K/R: a named ChatGPT memory system, cross-chat preference retention, and privacy resonance. Missing rollout, default setting, and retention details keep it below 85.

Take

Dreaming’s core move is shifting ChatGPT memory from explicit saved notes to background synthesis. The post says April 2024 saved memories needed cues like “remember,” April 2025 dreaming began referencing chat history, and this version starts with Plus and Pro users in the US before Free and Go rollout. That makes ChatGPT stickier because users stop maintaining prompt profiles manually.

My concern is control. OpenAI shows a memory summary, editing, and dismiss actions, but it does not give the default setting, retention period, or whether deletion cascades into synthesized state. Anthropic has kept more product boundaries around Projects and Artifacts; OpenAI is leaning into automatic personalization. For practitioners, this is a personal context layer becoming product lock-in, not a minor memory refresh.

AI HOT (Curated Pool)Jun 4, 08:00

Hugging Face redesigns hf CLI output format for coding agents

Hugging Face redesigned hf CLI output for coding agents including Claude Code and Codex, using environment-variable detection and compact untruncated TSV output; in complex multi-step tasks, agents without the CLI used up to 6 times more tokens.

Why it matters: HKR-H/K/R pass: the story has a clear agent-CLI hook, a concrete TSV/token mechanism, and strong developer cost resonance. It stays in the featured band because this is a tooling update, not a model or platform release.

Take

Hugging Face is calling out the hidden tax in agent tooling: human-readable CLI output burns context. The concrete hook is strong: hf now detects Claude Code, Codex, Cursor-style environments and returns compact, untruncated TSV; on complex multi-step Hub tasks, agents that hand-roll curl or the Python SDK used up to 6× more tokens.

I buy the direction more than the benchmark. The 6× number comes from Hugging Face’s own setup, and the provided article text cuts off before task mix, failure rates, or harness details. Still, this matches the last year of agent work: the bottleneck is often not model reasoning, but noisy interfaces. Pretty tables, truncated columns, and chatty errors are fine for humans; for agents they are context pollution. Every serious devtool now needs a machine-facing surface, not just docs.

Latent SpaceJun 4, 03:27

Scaling Past Informal AI - Carina Hong, Axiom Math

Axiom solved all 12 Putnam problems in 2025 and scored 8/12 within the time limit; Carina Hong says its Verina ProofGen result reached 187/189, while the last disclosed OpenAI o3 result on that benchmark was 4.9%.

Why it matters: HKR-H/K/R all pass: Putnam results, the o3 comparison, and 187/189 give it a real hook. It stays at 80 because this is a Latent Space interview/research story, not a broad model release.

Jun 3Wednesday

AI HOT (Curated Pool)Jun 3, 22:04

Microsoft and OpenAI Split as Both Prepare to Compete Directly

Microsoft and OpenAI have shifted from partnership to direct competition, and Microsoft AI chief Mustafa Suleyman said Microsoft must prove from scratch that it can independently complete the required work; the post does not disclose a product roadmap or timeline.

Why it matters: HKR-H and HKR-R pass: Microsoft/OpenAI rivalry affects agent-platform strategy. HKR-K is weak because the article gives no roadmap or testable technical detail, so it sits just above the featured threshold.

OpenAI NewsJun 3, 21:15

Introducing new capabilities to GPT-Rosalind

OpenAI says GPT-Rosalind adds biological reasoning, medicinal chemistry, genomics analysis, and experimental workflow capabilities; the RSS snippet does not disclose model parameters, benchmark results, pricing, or access conditions.

Why it matters: OpenAI’s vertical model update clears HKR-H and HKR-R, but HKR-K fails because evals, parameters, and access terms are missing. That keeps it at the featured floor.

Take

GPT-Rosalind reads more like an enterprise BD release than a scientific model release. OpenAI says it combines GPT-5.5 agentic coding and tool use with medicinal chemistry, genomics, wet-lab troubleshooting, and trusted access for eligible organizations. The concrete hook is LifeSciBench: six workflow areas and a DMD/FDA meeting example. The missing part is louder: no overall score, no category scores, no judge count, no baseline models, no pricing.

I don’t object to trusted-access life-science models; pharma workflows need controlled deployment. But biotech AI does not get trust from a capability menu. AlphaFold earned attention with reproducible structure metrics. A clinical or med-chem agent has to show error rates, audit trails, and failure cases. OpenAI is selling the workflow wrapper before showing the validation package.

Xinzhiyuan · WeChatJun 3, 12:05

OpenAI’s Greg Brockman and a 9-Year Rift With Anthropic Co-Founder Dario Amodei

A WSJ-based profile says Dario Amodei once barred Greg Brockman from an internal OpenAI project that later led to ChatGPT, and the article says Brockman now oversees OpenAI product strategy with nearly 1,500 people under that function.

Why it matters: HKR-H/K/R all pass: the WSJ-sourced ban detail and the nearly 1,500-person scope give this more signal than gossip. It is not a model release or current executive departure, so it stays in the good-quality featured band.

Take

Brockman taking over OpenAI product is the signal, not the Dario-vs-Greg soap opera. A 1,500-person product function under an engineering founder says OpenAI wants one owner for ChatGPT, Codex, and the dead-or-folded Sora surface. The concrete hook is Sora: the article says the standalone app shuts down and the API follows on September 24. That fits the compute math. Video generation burns inference budget that Codex and ChatGPT latency need.

I don’t buy the trillion-dollar revenge framing. The piece throws around Anthropic at $965B and an OpenAI IPO target of $852B, but this RSS-sourced body gives no financing terms or filing trail. The cleaner read: OpenAI is moving GPU and org priority away from flashy consumer video toward coding and general workflow. Claude Code already owns serious developer mindshare, so Brockman’s job is less mythology and more stopping Anthropic at the terminal.

AI HOT (Curated Pool)Jun 3, 09:31

Sensor Tower: ChatGPT surpasses 1B monthly active users, fastest ever

Sensor Tower estimates ChatGPT surpassed 1 billion global monthly active users in May 2025, while Anthropic’s Claude reached 56 million monthly active users in the same period with about 640% year-over-year growth.

Why it matters: HKR-H/K/R all pass: the article adds concrete adoption estimates for ChatGPT and Claude. It stays below P1 because these are third-party usage metrics, not a model or product capability release.

AI HOT (Curated Pool)Jun 3, 06:56

OpenAI launches Codex Sites to turn ideas into interactive websites

OpenAI opened Codex Sites in preview to Business and Enterprise subscribers, letting users turn ideas into hosted interactive sites such as dashboards, planners, and project boards, with URL sharing for specified team members.

Why it matters: HKR-H/K/R all pass, but this is an OpenAI Codex enterprise-preview feature rather than a model or core capability release. It sits in the mid-weight product-update band.

Take

Codex Sites is aimed at internal app sprawl, not website generation demos. OpenAI opened the preview only to Business and Enterprise users, and the named outputs are dashboards, planners, review workspaces, project boards, portfolios, and lightweight tools. The URL sharing is scoped to specified team members, so this is not an indie-builder toy.

The pressure lands on Retool, Notion database views, and all the half-maintained scripts inside ops teams. The finance example is the tell: turning a static spreadsheet into an interactive scenario planner saves tickets and sprint slots, not frontend polish. I don’t buy the “building apps has never been easier” framing until pricing, audit controls, and data connectors are clear. Those details decide whether this enters procurement or stays a slick ChatGPT preview.

AI HOT (Curated Pool)Jun 3, 02:39

Alphabet Plans $80 Billion Raise; Anthropic Files for IPO

Alphabet plans to raise $80 billion through equity financing for AI infrastructure expansion, while Anthropic has confidentially filed for an IPO; the post does not disclose valuation, listing timeline, or underwriters.

Why it matters: HKR-H/K/R all pass; Anthropic filing for an IPO is near the foundation-model IPO band, with Alphabet’s $80B AI infra financing adding hard capex signal. Valuation, timing, and banks are not disclosed, so the score stays below the very top.

AI HOT (Curated Pool)Jun 3, 02:12

Microsoft releases its first advanced reasoning AI model, MAI-Thinking-1

Microsoft released MAI-Thinking-1 at Build 2026, describing it as a medium-sized reasoning model that matches leading models on key software engineering benchmarks.

Why it matters: HKR-H/K/R all pass: Microsoft released its first advanced reasoning model with a mid-sized design and SWE benchmark claim. Exact scores, access, and pricing are not disclosed, so it stays below 85.

Take

Microsoft’s sharp move is not MAI-Thinking-1 itself; it is the public no-distillation claim. That matters because Microsoft has spent years leaning on OpenAI for frontier capability, then renegotiated the partnership as the relationship loosened. Now it needs an asset developers can view as Microsoft-native, not Azure packaging around another lab’s work.

The evidence is thin. The body says “medium-sized” and claims parity with leading models on “key” software-engineering benchmarks, but gives no parameter count, SWE-bench score, pricing, or context window. For practitioners, that is not yet a capability signal you can route workloads around. It smells more like leverage in the OpenAI negotiation. OpenAI has GPT-5-class branding, Anthropic owns a lot of coding mindshare with Claude Sonnet 4.5, and Microsoft still has to turn this Build demo into public evals people trust.

AI HOT (Curated Pool)Jun 3, 00:25

OpenAI Codex releases Python SDK for direct app integration

OpenAI Codex released a Python SDK with the install command pip install openai-codex, and the snippet says it can reuse the Codex login state; the post does not disclose API pricing, model versions, or rate-limit conditions.

Why it matters: HKR-H/K/R pass: a Codex SDK for embedded app use is practical and discussable. Sparse sourcing keeps it in the mid-weight product-update band: package and auth are given, but price, model, and rate limits are not.

AI HOT (Curated Pool)Jun 3, 00:22

OpenAI Codex Sites feature launches

OpenAI launched Codex Sites, which turns work, ideas, and plans into an interactive website or app that a team can access through one URL; the feature rolls out first to Business and Enterprise plans, and the post does not disclose pricing or broader availability timing.

Why it matters: HKR-H/K/R all pass, but the post gives launch framing without pricing, permission boundaries, or quality examples. Treat it as a mid-weight OpenAI product feature, above the featured threshold.

TechCrunch · AIJun 3, 00:00

OpenAI launches new Codex tools for white-collar work

OpenAI released six Codex app plug-ins for data analytics, creative production, sales, product design, equity investing, and investment banking; each tool bundles integrations, instructions, and context, while the post does not disclose pricing or rollout limits.

Why it matters: HKR-H/K/R all pass: OpenAI is expanding Codex into six white-collar plugin categories. Pricing, rollout scope, and measured performance are not disclosed, so this stays in the mid-weight product-update band.

Jun 2Tuesday

AI HOT (Curated Pool)Jun 2, 12:42

To Avoid Paying $120, I Turned a Computer Cleaner into an Open-Source Skill

The author open-sourced a cross-platform AI cleaning skill for Mac and Windows, generating interactive HTML reports from file scans; in a test, it freed nearly 120GB, compared with CleanMyMac identifying 15.8GB.

Why it matters: This is not a platform-level release, but HKR-H/K/R all land through the $120 replacement hook, concrete scan/report mechanism, and 120GB test result. It fits the practical open-source tool band near the featured threshold.

Financial Times · TechnologyJun 2, 09:33

Florida sues OpenAI and Altman for ‘hurting’ children

Florida sued OpenAI and Altman over a claimed “litany of harms” caused by the company’s chatbots; the RSS snippet does not disclose specific cases, damages, requested remedies, or procedural details.

Why it matters: HKR-H/R are strong; HKR-K is limited to the filed Florida suit. Missing cases, damages, and remedies keep it at 78 rather than 85+.

Take

Florida naming both OpenAI and Sam Altman is messier than another safety op-ed cycle. The title gives “hurting children” and a “litany of harms,” but the cases, damages, injunctions, and legal claims are not disclosed; those details decide whether this is political theater or a case that hits ChatGPT defaults, age checks, and high-risk chat logging.

OpenAI has spent the last year framing youth safety as product iteration: parental controls, crisis routing, model behavior rules. Litigation asks for a different record: red-team notes, incident tickets, launch reviews, and Altman’s public claims. Meta and TikTok already learned this playbook in youth-harm suits. The painful part for LLM vendors is not the fine; it is every safety promise turning into evidence that they knew the risk.