OpenAI just shipped three voice models that do completely different things — and that architectural split might matter more than any individual model's...
OpenAI swapped ChatGPT's default model from GPT-5.3 Instant to GPT-5.
OpenAI was spending fifteen million dollars a day running Sora. The product generated $2.
OpenAI shipped a Codex update last week that buries the lede in its own announcement.
OpenAI open-sourced Symphony ten days ago, and most coverage fixated on the "6x more merged PRs" headline.
OpenAI's models are no longer trapped behind Azure's velvet rope. As of April 28, GPT-5.
Two weeks ago, Sam Altman went on a podcast and called Anthropic's decision to restrict Mythos Preview "fear-based marketing" — accusing them of...
OpenAI confirmed it: DALL-E 2 and DALL-E 3 both go dark on May 12. Not deprecated-but-still-kinda-works dark.
After seven years of Azure exclusivity, OpenAI's frontier models are finally multi-cloud.
Sunday's announcement got big headlines: Microsoft and OpenAI restructured their partnership, and OpenAI can now "serve all its products to customers...
OpenAI shipped GPT-5.5 yesterday, and if you just glanced at the headline, you'd be forgiven for yawning.
OpenAI quietly shipped a model that will help you reverse-engineer a binary, identify anti-debugging tricks, and generate detection signatures for malware...
OpenAI shipped a Codex update on April 16 that got buried under the usual cycle of model-release drama and benchmark wars.
On April 26, OpenAI pulls the plug on its video generator. Six days from now, the most hyped AI video product ever built stops generating videos.
On April 4, three image generators nobody had heard of appeared on LM Arena's blind evaluation page.
Every security engineer has hit the wall.
OpenAI spent more money per day running Sora than most startups raise in a seed round.
Microsoft quietly dropped three foundation models last week — MAI-Transcribe-1, MAI-Voice-1, and MAI-Image-2 — and most of the AI news cycle barely noticed...
A month ago, OpenAI quietly crossed a threshold that deserves more attention than it got: GPT-5.
I spent three months in 2024 building retry logic for a pipeline that extracted product data from GPT-4.