AI progress trajectories
How fast will AI improve?
Posts by Year
2024
2023
Wait - why can LLMs deal with spelling mistakes?
Overview Why can’t LLMs reliably count the number of letters in a word, but they do know how to spell? E.g. why can LLMs not count the number of times “o” a...
Implementing agents in industry
Overview Here I reflect on work on implementing autonomous systems in large manufacturing companies, and I draw some insights for future autonomous agents wo...
Taking AI risks seriously
AI and being precautious I’ve had conversations with many who attempt to dismiss existential AI risk. I find that a lot of their arguments fail to address th...