asadm

16 ms·

1. 271.▲
   by asadm 1y ago

   AI is mere mirror of human code.
2. 272.▲
   by asadm 1y ago

   Yup I get that.
3. 273.▲
   by asadm 1y ago

   in the editor, do add a way to load sample random image to try out the editor instead of having to uploading my own (which most won't bother / don't trust / can't do at work).
4. 274.▲
   by asadm 1y ago

   I have a question, can something like this survive in today's world? or have the disassembling tools now too advanced to easily wipe something like this when cloning.
5. 275.▲
   by asadm 1y ago

   fb has a tab that works like this now.
6. 276.▲
   by asadm 1y ago

   yes
7. 277.▲
   by asadm 1y ago

   i think this was made before that PR was merged into codex.
8. 278.▲
   by asadm 2y ago

   funny thing about younglings, they will migrate to something else as fast as they came to you.
9. 279.▲
   by asadm 2y ago

   yup exactly. as weird workflows emerge it’s nicer to have your own weird tooling around this until we all converge to one optimal way.
10. 280.▲
    by asadm 2y ago

    that’s correct. aider doesn’t RAG on files which is good. I don’t use it because 1) UI is so slow and clunky 2) using gemini 2.5 via api in this way (huge context window) is expensive but also heavily rate limited at this point. No such iss
11. 281.▲
    by asadm 2y ago

    speaking for myself, I am happy to make that trade. As long as I get unrestricted access to latest one. Heck, most of my code now is written by gemini anyway haha.
12. 282.▲
    by asadm 2y ago

    i havent tried recently but does it tell if it RAG'ed or not ie. can I peak at context it sent to model?
13. 283.▲
    by asadm 2y ago

    filekitty is pretty cool!
14. 284.▲
    Show HN: Vibe Mode – Apply code changes given by LLM back into your project(github.com)

    1 points by asadm 2y ago|0 comments
15. 285.▲
    by asadm 2y ago

    Yup. but also there was good reason to do this. Models work better with smaller context. Which is why I rely on Gemini for this lazy/inefficient workflow of mine.
16. 286.▲
    by asadm 2y ago

    exactly. I understand the reason behind this but it's too magical for me. I just want dumb tooling between me and my LLM.
17. 287.▲
    by asadm 2y ago

    probably but also most models start to lose it after a certain context size (usually 10-20k). Which is why I use gemini (via aistudio) for my workflow.
18. 288.▲
    by asadm 2y ago

    i will try again but last i tried adding folder to edit mode and asking it to list down files it sees, it didn't list them all down.
19. 289.▲
    by asadm 2y ago

    I am at my 5th project in this workflow and these are of different types too: - an embedded project for esp32 (100k tokens) - visual inertial odometry algorithm (200k+ tokens) - a web app (60k tokens) - the tool itself mentioned above (~30k
20. 290.▲
    by asadm 2y ago

    do you mean gemini? I generally notice pretty great recall UPTO 200k tokens. It's ~OK after that.
21. 291.▲
    by asadm 2y ago

    copilot (and others) try to be too smart and do context reduction (to save their own wallets). I want ENTIRETY of the files I attached to context, not RAG-ed version of it.
22. 292.▲
    by asadm 2y ago

    These days, I usually paste my entire (or some) repo into gemini and then APPLY changes back into my code using this handy script i wrote: https://github.com/asadm/vibemode I have tried aider/copilot/continue
23. 293.▲
    by asadm 2y ago

    thanks
24. 294.▲
    by asadm 2y ago

    usually enabling "Search" fixes it sometimes as they fetch the newer methods.
25. 295.▲
    by asadm 2y ago

    I was working on a similar thing but for JS. Imagine this: It would be cool when these functions essentially boiled down to a distilled tiny model just for that functionality instead of an api call to foundation one.
26. 296.▲
    by asadm 2y ago

    AI can facilitate the creation part and become invisible. It will still be p2p.
27. 297.▲
    by asadm 2y ago

    I feel like this is the kind of stuff that will be common with AI. Imagine having a whatsapp family group and a bot auto-generates this type of games for you based on your group. It's a privacy nightmare but it will be fun for sure.
28. 298.▲
    by asadm 2y ago

    gpt likes xml
29. 299.▲
    by asadm 2y ago

    Is there a simple utility / script that takes gemini/grok/gpt output (usually markdown with changes, and apply it to the folder back. I dont want to deal with "smartness" of aider. I WANT to use the chatgpt/ai-
30. 300.▲
    by asadm 2y ago

    I have a few ways but usually https://github.com/yamadashy/repomix works