Surprisingly reassuring piece from OpenAI co-founder

Have you ever felt… dejected?

Like you’re facing a obstacle so big you can’t seem to overcome?

And then you start to question your life decisions and wonder if you should just... stop?

I certainly have. I have been spending the last 5 years of my life on a wild goose chase, throwing every resource I have at it in the process. At this point I’m running on fume yet I can see no end in sight, and only have some vague ideas for a game plan.

You know who else has felt the same way? This guy:

His name is Andrej Karpathy. He is an AI researcher and was a PhD student in 2012 when he wrote this blog post expressing his worries about the challenges facing the then state-of-the-art AI and specifically computer vision models (tldr: 2012 AI not cool yet cuz they don’t understand funny images and no idea if they ever be cool). He ended the post with this passage of pessimism:

Funny thing is, the guy indeed went on to “do a startup”, it is called… OpenAI. You know, the company that makes ChatGPT and DALL-E?. AIs, including both language and vision models, have gone a VERY long way since 2012, in some parts due to the work of him and his colleagues. (*) (**)

I find this to be quite… comforting, in a way. To know that even someone at the top of their field was once at a loss just like I do now, and to know that despite all that he would go on to tackle the crap out of that problem.

And who knows, maybe when I look back at this post some years later I will have done something useful for once. wink to my future self

(*) I’m not sure if AI has quite understood funny imagery yet, but still it was very very commendable progress. I have a friend working on this very problem, best of luck to him.

(**) Karpathy also makes some very high-quality & informative videos about AI on his Youtube channel.