🚫 OpenAI is no longer following the vertical axis

let's get to work

As predicted, examples of GPT-5 nano or mini producing bad outputs abound online. Not making it clear how GPT-5 works will likely cause issues for OpenAI. I wonder if they will need to take a different approach to switching or at least educating users about what GPT-5 does.

Wow, I was just playing around before but it actually is stupid

I suspect that due to GPT-5's new safety rules, it is much more woke than GPT-4.1, which may cause issues with the Trump admin's "No Woke AI" executive order.

real data scientists resolve their A/B tests in 24 hours

16 hours later he is still mad and ranting about 1/3 = 0.333... in the replies.

New blog post up: I asked all the top LLM's how many times the letter b appears in "blueberry." ...and something's up with GPT-5 Chat. minimaxir.com/2025/08/llm-bl…

Introducing Gemma 3 270M 🔥 🤏A tiny model! Just 270 million parameters 🧠 Very strong instruction following 🤖 Fine-tune in just a few minutes, with a large vocabulary to serve as a high-quality foundation developers.googleblog.com/en/introducing…

OpenAI *really* does not want people using GPT-4o. The money presumably saved from GPT-5's routing must be very compelling.

Wrote up some notes on the API version of GPT-5's hidden system prompt - it definitely adds today's date and appears to add other stuff too I'd really like to see this documented by OpenAI Developers - as an API user I want visibility into the whole prompt! simonwillison.net/2025/Aug/15/gp…

which SMS spam bot bucket is this