The nice thing about autonomous AI agents might be that compared to self-driving cars, the false positive effects aren't that disastrous. There could be a lot of opportunities to try things that are time cumbersome but isn't the worst if let's say the AI fails at making a reservation at a restaurant the first 10 times.
https://twitter.com/mcraddock/status/1645779894670950400/photo/1
A bit bigger map of LLM + Prompt Enginnering
Great deconstruction plus love the wordplay on anatomy of autonomy
A valuable, extensive, insightful overview. Thanks!
Please, slow down. Do you really want to create autonomous agents? What if those agents become too powerful to control?
The nice thing about autonomous AI agents might be that compared to self-driving cars, the false positive effects aren't that disastrous. There could be a lot of opportunities to try things that are time cumbersome but isn't the worst if let's say the AI fails at making a reservation at a restaurant the first 10 times.