Hacker Newsnew | past | comments | ask | show | jobs | submit | maxtermed's commentslogin

I watched someone build a full AI app in minutes. Then the API errored and they had no idea where to look. They had never seen a raw API response. They asked the AI to fix it, stacked patches until the errors eventually killed the app. They had built something they couldn't debug.

I used to be in the same spot. So I stripped it all down to raw API calls.

There's barely anything there.

Memory? A list you resend every time. Tool calling? The model returns JSON saying "call this function." You call it. That's "AI agents." RAG? Search docs, paste into prompt, ask.

Oversimplification? Maybe. But the fundamentals never stop being useful, no matter whatever advanced tool it's used.

10 progressive modules. OpenAI and Anthropic examples. Current SDKs. Only prerequisite is Python. Free and open-source (MIT).


After a LinkedIn post about AI hallucinations generated 166 comments, I realized: everyone was right, but talking about completely different problems.

This breaks down five perspectives (practitioners, displaced workers, engineers, educators, skeptics) and what each reveals that we're not discussing:

  - The verification paradox (if you can verify it, you could probably do it yourself)
  - Three zones of AI appropriateness (Green/Yellow/Red)  
  - Why AI hallucinations are structurally different from human errors
  - The human capital problem (if AI does junior work, where do seniors come from?)


That's only four things that we're not discussing. Did you miss one?


Five perspectives (in parentheses), four key concepts (bulleted). The perspectives reveal the concepts we're missing in the debate.


Ah. I thought there was going to be a key concept for each perspective.


I've been using this framework for a while, it's really solid IMO. It abstracts just enough to make building reliable agents straightforward, but still leaves lots of room for customization.

The way agent construction is laid out (with a clear path for progressively adding tools, memory, knowledge, storage, etc.) feels very logical.

Definitely lowered the time it takes to get something working.


Thank you for using Agno and the kind words!


Good point. The cookbook can be hard to navigate right now, but that's mostly because the team is putting out a tremendous amount of work and updating things constantly, which is a good problem to have.

This example might be close to what you're describing: https://github.com/agno-agi/agno/blob/main/cookbook/workflow...

It chains agents for web research, content extraction, and writing with citations.

I used it as a starting point for a couple projects that are now in production. It helped clarify how to structure workflows.


Consider applying for YC's Summer 2026 batch! Applications are open till May 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: