538 summarised stories about OpenAI, each linking back to the original source. Browse all topics →
OpenAI Blog 9 months ago
OpenAI made Codex, its code-generation model, available to all developers rather than limiting it to a waitlist. The release includes a Slack integration, a software development kit, and administrative tools for usage tracking and workspace management. Organizations can now deploy Codex across teams without special approval processes.
OpenAI Blog 9 months ago
AMD and OpenAI announced a multi-year partnership to deploy AMD Instinct GPUs for OpenAI's AI infrastructure. The companies will begin with 1 gigawatt of GPU capacity in 2026, scaling to 6 gigawatts total. OpenAI will expand its compute infrastructure using AMD chips instead of relying solely on NVIDIA processors.
The daily briefing
TLDRocket reads 60+ sources, removes duplicate coverage, and summarises the day in two minutes. Free, no spam, unsubscribe anytime.
TLDRocket
1 August 2026
The day's biggest story isn't wrapped in one headline but scattered across three: the infrastructure arms race is shifting who controls AI, and it's reshaping everything downstream. AMD released Instella-MoE-16B-A3B, a fully open mixture-of-experts model trained on its MI300X and MI325X GPUs, with transparent weights and training code available for academic labs. Meanwhile, NVIDIA showed its Transformer Engine can achieve 2x faster training via FP8 precision, and DeepSeek dropped V4-Flash 0731 with prices at $0.14 per million tokens—a direct challenge to proprietary models. These aren't minor engineering updates; they're commoditizing the foundation layer, which means the real margin shifts to whoever optimizes the layer above: agents and task-specific APIs.
Read the full briefing →