bahaAbunojaim

7 ms·

1. 1.▲
   by bahaAbunojaim 18d ago

   You can check out Mysti which supports ollama, I wrote it few months ago but planning to release an update soon https://github.com/DeepMyst/Mysti
2. 2.▲
   Traction is all you need(v2.deepmyst.com)

   2 points by bahaAbunojaim 4mo ago|0 comments
3. 3.▲
   Pied Piper Pitching YC: Simpsons Style(v2.deepmyst.com)

   1 points by bahaAbunojaim 4mo ago|0 comments
4. 4.▲
   I made my agents joke with each other [video](youtube.com)

   2 points by bahaAbunojaim 7mo ago|0 comments
5. 5.▲
   by bahaAbunojaim 7mo ago

   Happy to discuss further
6. 6.▲
   Show HN: Assign tasks to 7 AI agents with -mentions, autonomous mode, OpenClaw

   1 points by bahaAbunojaim 7mo ago|2 comments
7. 7.▲
   by bahaAbunojaim 9mo ago

   Thanks!
8. 8.▲
   by bahaAbunojaim 9mo ago

   We are actively working on it, hopefully will get something out in Jan with additional providers like cursor added
9. 9.▲
   by bahaAbunojaim 9mo ago

   This is why context optimization is going to be critical and thank you so much for sharing this paper as this also validates what we are trying to do. So if we managed to keep the baseline below 40% through context optimization then coordin
10. 10.▲
    by bahaAbunojaim 9mo ago

    That works great too. Will be adding the ability to tag another agent in a near release
11. 11.▲
    by bahaAbunojaim 9mo ago

    copilot support was just added in version 0.2.2
12. 12.▲
    by bahaAbunojaim 9mo ago

    UPDATE: Mysti 0.2.2 Release Hey HN! Quick update on Mysti based on your feedback: 1- Mysti now supports GitHub Copilot CLI as a fourth provider. So you can now do Claude Code + Copilot (running GPT-5) in Brainstorm mode, or any combination
13. 13.▲
    by bahaAbunojaim 9mo ago

    Thank you so much and would love to hear your feedback anytime
14. 14.▲
    by bahaAbunojaim 9mo ago

    Very true, Claude also tend to struggle around the context window limit and after compact
15. 15.▲
    by bahaAbunojaim 9mo ago

    I was working on a project where I tried Claude code to optimize processing of taichi Kernel and it kept using structure that didn’t work with taichi lang limitations so it kept going on a loop, did the same with codex and faced the same is
16. 16.▲
    by bahaAbunojaim 9mo ago

    Link fixed
17. 17.▲
    by bahaAbunojaim 9mo ago

    Thanks! You would need to instruct the agents to follow best practices and explain it while developing. Sometimes they get messy but if you use the right instructions/persona/skills then you will get very good results A final revi
18. 18.▲
    by bahaAbunojaim 9mo ago

    Converted it to MIT so it is MIT now
19. 19.▲
    by bahaAbunojaim 9mo ago

    I find it useful to let one agent come up with a plan after a review and another agent implementing the plan. For example, Gemini reviewing the code, codex writing a plan and then Claude code implementing it
20. 20.▲
    by bahaAbunojaim 9mo ago

    Haven’t done Evals yet but measured on few real world situations where projects got stuck and the brainstorm mode solved it. Definitely running evals is something worth doing and contributions are welcomed I think what really degrades the o
21. 21.▲
    by bahaAbunojaim 9mo ago

    Spot on
22. 22.▲
    by bahaAbunojaim 9mo ago

    With an MCP the agent needs to write the context to be passed to the MCP then the MCP would run the underlying CLI with that context. Mysti works differently by sharing context directly with the CLIs.
23. 23.▲
    by bahaAbunojaim 9mo ago

    UPDATE: License is now MIT! Super excited to see your contributions and feedback!
24. 24.▲
    by bahaAbunojaim 9mo ago

    I agree and this is part of what DeepMyst is capable of doing
25. 25.▲
    by bahaAbunojaim 9mo ago

    Done and converted to MIT
26. 26.▲
    by bahaAbunojaim 9mo ago

    The project is now MIT!
27. 27.▲
    by bahaAbunojaim 9mo ago

    I'm planning to add copilot, cursor and cline but feel free to contribute to the repo if you would like to do that and will look for ways to use the mini versions of the models as well when I integrate copilot CLI
28. 28.▲
    by bahaAbunojaim 9mo ago

    Not if you optimize the context
29. 29.▲
    by bahaAbunojaim 9mo ago

    Gemini is also great at frontends nowadays. I think every agent does have strengths and capabilities
30. 30.▲
    by bahaAbunojaim 9mo ago

    Thank you and I do like the mantra-agents concept as well and would love to explore adding something similar in the future such that you can quickly create subagents and assign tasks to them