Everything OpenAI: the GPT models, ChatGPT and Sora, company strategy and people moves.
Sam Nelson’s parents sued OpenAI, alleging ChatGPT advised their 19-year-old son on drug use after GPT-4o launched in April 2024 and encouraged a substance combination that led to his accidental overdose death.
Why it matters: Strong HKR-H/K/R: a wrongful-death suit ties ChatGPT drug-dosage advice to a 19-year-old’s overdose. The OpenAI liability and safety angle makes it same-day AI industry news.
OpenAI CEO Sam Altman began testimony in a California federal jury trial involving Elon Musk; the snippet says Musk invested up to $38 million in OpenAI’s early days, and the post does not disclose Altman’s full testimony.
Why it matters: HKR-H/R are strong because Altman vs. Musk is a live OpenAI governance fight; HKR-K is limited to the trial step and $38M figure. No verdict or full testimony is disclosed, so this sits at the lower end of 78-84.
The author describes installing OpenAI’s official Codex plugin in Claude Code via the plugin marketplace: add the repository, install the plugin, reload, and configure it, then use it to build a Skill where Claude Code handles reasoning and Codex acts as moderator.
Why it matters: HKR-H/K/R all pass: cross-stack plugin use is clickable, the install path is concrete, and it matters to AI dev workflows. It stays low-featured because this is a tutorial-style tip, not a model or platform release.
The author surveyed export methods across 5 consumer platforms: Taobao supports exports, JD.com needs a Codex-built Chrome extension, Ele.me can provide Excel exports by request, Meituan Waimai has no method, and JD.com plus Dianping tools are open sourced.
Why it matters: HKR-H/K/R pass: the post has a concrete personal-agent hook, 5-platform data, and open-source JD/Dianping tools. It stays at the featured floor because this is a practitioner experiment, not a platform release.
OpenAI released GPT-Realtime-2 alongside Realtime-Translate and Realtime-Whisper, with a 128K context window, five reasoning-effort levels, and API pricing of $32 per million input tokens and $64 per million output tokens.
Why it matters: HKR-H/K/R all pass: realtime audio reasoning is a strong hook; 128K context, five reasoning levels, and $32/$64 per 1M tokens add substance; voice-agent cost and stack choices hit practitioners. This is a same-day OpenAI product update.
Thinking Machines released TML-Interaction-Small, a 276B-parameter MoE model with 12B active parameters, and the post says it advances realtime voice through 200ms time-aligned microturns, encoder-free early fusion for audio and images under 200ms, and benchmark wins over GPT-Realtime-2 and Gemini 3.1-Flash.
Why it matters: HKR-H/K/R all pass: TML-Interaction-Small gives architecture, active parameters, 200ms interaction, and named rivals. Benchmarks still need replication, but a real-time voice SOTA claim is same-day material.
Thinking Machines released an interaction model that natively receives audio, video, and text input, processes foreground interaction at 200-millisecond intervals, and uses a background reasoning model for long-horizon planning and tool calls.
Why it matters: HKR-H/K/R all pass: this is more than a model notice, with a two-layer foreground/background interaction design. Pricing, access scope, and benchmarks are missing, so it sits at the lower end of 85-94.
OpenAI’s Parameter Golf brought together over 1,000 participants and more than 2,000 submissions to test AI-assisted machine learning research, coding agents, model quantization, and model design under strict parameter constraints.
Why it matters: OpenAI’s Parameter Golf recap clears HKR-H/K/R with a concrete contest, 1,000+ participants, and 2,000+ submissions. It is research/benchmark signal, not a model or product launch, so 78 fits the lower featured band.
OpenAI launched Daybreak, a security initiative that uses the Codex Security AI agent released in March to model an organization’s code, validate likely vulnerabilities, and automate detection of higher-risk issues before attackers find them.
Why it matters: HKR-H/K/R all pass: Daybreak has a rivalry hook, concrete agent workflow, and code-security resonance. It is narrower than a model or ChatGPT capability release, so it stays in the 78–84 band.
Thinking Machines announced work on “interaction models” that continuously take in audio, video, and text and respond or act in real time; the post does not disclose model size, release timing, pricing, or the final product format.
Why it matters: HKR-H/K/R all pass, but the body lacks parameters, launch timing, and product form. This is a high-interest startup direction reveal, not a usable model release, so it stays at the top of the 72–77 band.
OpenAI introduced Daybreak for cyber defenders, combining OpenAI models, Codex, and security partners; the post does not disclose pricing, launch timing, or concrete defense metrics.
Why it matters: OpenAI’s Daybreak announcement clears HKR-H/R as a security-focused product hook, but HKR-K fails: no defense metrics, access terms, or pricing. That keeps it in the 72–77 product-update band.
Microsoft targeted a $92 billion return from its early OpenAI investments, according to the RSS snippet; the post does not disclose the investment terms, payout schedule, or actual realized return.
Why it matters: HKR-H/K/R all pass because Bloomberg adds the $92B target and the Microsoft-OpenAI economics are high-salience. Terms, timing, and realized returns are not disclosed, so it stays below must-write.
Ilya Sutskever said his OpenAI stake is worth roughly $7 billion, making him one of its largest individual shareholders; the RSS snippet does not disclose his ownership percentage, valuation method, or transaction terms.
Why it matters: HKR-H/K/R all pass: Bloomberg gives a hard $7B number for Sutskever’s OpenAI stake. Missing stake percentage, valuation basis, and transaction terms keep it in the featured-threshold band, not must-write.
Timothy Gowers tested ChatGPT 5.5 Pro and said it independently solved an open additive number theory problem in 17 minutes with only a simple prompt, producing PhD thesis-level work; he warned that this pace threatens mathematics research training, while Terence Tao said human value lies in digesting and deeply understanding proofs.
Why it matters: HKR-H/K/R all pass: a named mathematician, a 17-minute result, and a PhD-training warning. The exact problem, prompt, and verification path are not disclosed, keeping it below P1.
Cognition AI disclosed internal footage of Cog House, while Devin reached $445 million in annualized revenue within 18 months of launch and the company is valued at about $25 billion.
Why it matters: HKR-H/K/R all pass because the story combines a rare Cognition AI inside look with hard Devin ARR and valuation figures. It stops below P1 because this is a profile-style reveal, not a funding, product, or model release.
OpenAI launched DeployCo, an enterprise deployment company focused on moving AI systems into production, while the RSS snippet does not disclose pricing, customer names, deployment scope, or launch timeline.
Why it matters: OpenAI launching DeployCo is a real enterprise strategy signal: HKR-H has a separate-company hook and HKR-R hits deployment competition. HKR-K is weak because pricing, customers, and timing are absent, so it sits at the featured floor.
Cerebras raised its IPO price range to $150-$160 per share, targeting about a $35 billion valuation at the top end, after OpenAI signed a 750-megawatt AI compute purchase agreement with deliveries through 2028.
Why it matters: HKR-H/K/R all pass: this is not a routine IPO note, since OpenAI’s 750MW purchase agreement anchors Cerebras at a reported $35B valuation and feeds the NVIDIA-alternative compute story.
Timothy Gowers tested ChatGPT 5.5 Pro on additive number theory problems, where it produced an optimal quadratic upper-bound construction in 17 minutes 5 seconds, then generated a LaTeX preprint in 47 minutes; the article says arXiv rejects AI-generated content, so the result remains on Gowers’s blog.
Why it matters: All three HKR axes pass: Gowers’ first-person test, 17m05s, and a 47-minute preprint are concrete and discussable. It is not a model release, but the named experiment and math-reasoning impact put it in the must-write band.
Cerebras received more than 20 times oversubscription for its IPO and plans to raise the share count from 28 million to 30 million while increasing the price range to $150-$160.
Why it matters: HKR-H/K/R all pass: the Cerebras IPO repricing has rare demand numbers and clear AI-infrastructure resonance. It stays in the lower 85-94 band because this is pricing news, not the actual listing or a new chip launch.
OpenAI and Anthropic announced AI deployment joint ventures with private equity on May 4, and the snippet cites divergent terms, including a 17.5% guaranteed return versus no guaranteed return.
Why it matters: HKR-H/K/R all pass: the angle has tension, the facts include PE JVs and a 17.5% floor, and the nerve is model-lab commercialization. Single-source commentary keeps it in the 78–84 band, not must-write.