AI Frameworks, Agentic Loops, and Tooling, Oh My...
AI coding used to be about autocomplete. Then it became chat. Now it is about the agentic “loop.” We are moving from “write me this function” to “here is the goal, here is the repo, go investigate, make a plan, change the code, run the tests, fix what failed, and show me what you created.”
Global AI Milwaukee Wed July 8th Meetup!
https://www.meetup.com/global-ai_milwaukee/events/315414148/
AI Frameworks, Agentic Loops, and Tooling, Oh My...
AI coding used to be about autocomplete.
Then it became chat.
Now it is about the agentic “loop.”
We are moving from “write me this function” to “here is the goal, here is the repo, go investigate, make a plan, change the code, run the tests, fix what failed, and show me what you created.”
That is why features like Codex’s /goal command feel like a milestone. It is not just a nicer prompt. It is a different operating model: a more persistent objective that the coding agent can work against over longer stretches of time.
Along with many other new "framework" features, we'll talk about some fun open source. AI tools, and of course we will talk about how to write better goals for loops, how to give the agent useful context, how to use tests and tools as guardrails, and how to review what comes back with enough skepticism to stay safe — but enough trust to actually move faster.
https://www.LanceLarsen.com