There are a host of “autonomous” LLM agents that have taken GitHub by storm — BabyAGI, AutoGPT, Jarvis. The promise? Goal-driven self-executing software. An example of an exclamatory software interface. Autonomous AI developers, assistants, and other white-collar workers replaced by software that can generate, execute, and prioritize tasks given a certain goal.
Hey Matt, what do you think of the 5 phase of autonomous agent evolution in this diagram https://substackcdn.com/image/fetch/w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9a68add7-01a3-4d2d-a050-edeeeae66ddd_1996x1292.png that i found from this article https://www.latent.space/p/agents?r=bxko&utm_medium=ios&utm_campaign=post?
I get that you say it's at least 10 years out. but would the situation evolve in roughly the same order as indicated in the image?
That’s a really interesting way to think about it. Will show how I think about it in a future post!