From the course: Agentic AI and Autonomous Development

Unlock this course with a free trial

Join today to access over 25,600 courses taught by industry experts.

Simple agents win: Design philosophy

Simple agents win: Design philosophy

From the course: Agentic AI and Autonomous Development

Start my 1-month free trial Buy for my team

Simple agents win: Design philosophy

“

- [Instructor] What could the future of agentic coding look like and agentic tooling? Well, if we look at originally when things came out around large language models, there was a talk of AGI. What we see is that that has not panned out, and many companies have stopped talking about the concept of AGI because it doesn't make sense in terms of what we're seeing. In fact, the Floridi conjecture is one of the best rebuttals to it in that it mentions that the broad scope is an inherent conflict with certainty. And as a result, the more certain you want things, the smaller the scope needs to be. If we look at this with combination to Amdahl's law for subagents, which says that if we extrapolate human attention as the bottleneck, we see that there could be an emerging sweet spot for developing with agents. How would this look? Well, if you take a look at small, simple problems, for example, looking at the Clippy warnings from…

Contents