The longer I work in AI, the more I think humans are just simple pattern matching machines with a small scratch pad for memory
Peter Welinder
2,184 posts
- I have been at OpenAI for 7 years. I overlapped with the “concerned former employees.” Their claims display lack of larger context and lack of maturity. I mean, we argue about GPU allocation all the time! Sam’s job is to decide. Ofc someone will be disappointed. Grow up.
- When we get naming right, you’ll know we’ve achieved AGI internally.I love OpenAI but they are truly miserable at naming things
- We've opened up access to DALL-E. No more waitlist. Excited to see even more people create with AI.
- No, we haven't made GPT-4 dumber. Quite the opposite: we make each new version smarter than the previous one. Current hypothesis: When you use it more heavily, you start noticing issues you didn't see before.
- Everyone reading up on Q-learning. Just wait until they hear about PPO. 🤫
- GPT-3 is amazing at complex tasks like creative writing and summarizing. But it's surprisingly bad at reversing words. 🤔 The reason is that GPT-3 doesn't see the world the way we humans do. 👀 If you teach it to reason, it can get around its limitations to get really good. 💡
- OpenAI is nothing without its people.
- Using ChatGPT with the Zapier plugin is giving me chills. Feels like having an army of helpers at my disposal.
- Around this time 3 years ago, early 2020, we started testing the OpenAI GPT-3 API with the first external users. It was really hard to get anyone to try it, let alone build on it. So much has changed since then, but still feels like we're in the early days.
- Data classes in Python 3.7 are great for encapsulating hyper parameters in machine learning models. Gone are long keyword argument lists. Much clearer code.
