Hi all, I lead the reinforcement learning team at OpenAI. Last month we released WebGPT (OpenAI blog post), a system that combines a language model with a browser for long-form question answering. We’d like to polish it and turn it into a useful product, to be released later this year. We’re looking to hire a software engineer who’s excited about optimizing its performance and usability to make a great product. If you’re interested, please apply to this job posting and mention that you saw the ad here.
Hi @joschu, this sounds very exciting. I assume you have a particular product in mind that you are, understandably, keeping confidential? Or was your post actually a broad indication that you want to build “some” practical product? Based on my work with Nikhil Horakeri, my CTO, we have developed some significant ideas re practical applications for long form question answering. I’d love to connect with you/OpenAI to discuss further. (FYI, I applied yesterday to go live with Lexata’s app, however, it’s a “starter app” using only the embeddings endpoint, and doesn’t include our bigger ideas). Thanks very much. Leslie
“We’d like to polish it and turn it into a useful product, to be released later this year.”
Will it be a new API endpoint (like the previous things) or a page like “playground” or both?
I personally would prefer to be an API endpoint so that I can incorporate that into my product.
Maybe an endpoint, or maybe we’ll open source the non-model code so you can run that on your end.
Hey @joschu , sign me up no matter how it works, we’ll probably be glad to incorporate it into our product, which is a personal assistant dedicated to the drive. If you need a beta tester, I would be happy to test and provide feedback.
E-mail sent. Look forward to chatting. Thanks.
hi joschu, i’d like to email you about a similar product and see if this is something you’re interested in