results make the slog worth it, but it's also best to find joy in the slog itself since it's how you'll spend most of your time

programming at its best feels like the constant discovery and invention of new concepts, materialized into useful form

OpenAI 🤝 Moderna: wsj.com/articles/at-mo…

First @NVIDIA DGX H200 in the world, hand-delivered to OpenAI and dedicated by Jensen 'to advance AI, computing, and humanity':

Safety research on robustness to LLM attacks such as prompt injection, by Lilian Weng and team:

a sign of good design is when your framework gets increasingly written in itself over time

Significant new features in the Assistants API:

We've just launched Projects — giving developers much more fine-grained control over permissions:

Batch API — 50% off for async requests to our models:

a bug is the gap between expectation and reality

OpenAI Japan — こんにちは!

once you deeply understand the why, you can fix it

When debugging an opaque problem in a system, a nice strategy is to just focus on adding permanent (rather than temporary) metrics or logging to make the problem obvious. Often makes it easy to solve not just the specific issue but also a whole class of potential related issues.

best way to get hard things to work is to first focus on getting all the easy things to work perfectly

love watching the openai team rally to provide great support for a customer