jbellis

11 ms·

1. 181.▲
   by jbellis 1y ago

   for me the problem here is an org that debates sprint length for months
2. 182.▲
   by jbellis 1y ago

   Love to see it! Wonder how to reconcile the description of almost-negligible admin overhead with this description of a similar effort that warns, "We wanted to keep costs extremely low, so we had parent volunteers do all admin for the
3. 183.▲
   by jbellis 1y ago

   Used in Dropwizard Metrics, among other places. https://metrics.dropwizard.io/4.2.0/
4. 184.▲
   by jbellis 1y ago

   worktree + pr support is coming soon, in the meantime you gotta do it manually
5. 185.▲
   by jbellis 1y ago

   you're right about the context limits i pinged what i think is the right ghuntley on linkedin, rizzler looks like the next feature i'm building for brokk :)
6. 186.▲
   by jbellis 1y ago

   no, but I've tested on intellij (~5M loc, takes forever to import b/c of delombok, do not recommend)
7. 187.▲
   by jbellis 1y ago

   (creator here) idk, everyone else seems to want to take the 40 year old IDE paradigm we're all used to (really! that's how old Turbo Pascal 3 is!) and graft AI onto it. I think we need a fundamentally different design to truly tak
8. 188.▲
   by jbellis 1y ago

   Oh, cool. I want this but not quite badly enough to work all the corner cases out. Were you able to figure out how to sync pages visited on mobile?
9. 189.▲
   by jbellis 1y ago

   That officially makes you an early adopter, thanks! Yes, it's basically just parsing for declarations. (If you doubleclick on any context in the Workspace it will show you exactly what's inside.) You have to import the dependencie
10. 190.▲
    by jbellis 1y ago

    Hi all, Brokk creator here, happy to answer any questions! I made an intro video with a live demo here: https://www.youtube.com/watch?v=Pw92v-uN5xI
11. 191.▲
    by jbellis 1y ago

    Yeah, that seems possible, but a dumb preprocessing step won't help and a smart one will add significant latency. Feels a bit halting-problem-ish: can you tell if a problem is too hard for model A without being smarter than model A you
12. 192.▲
    by jbellis 1y ago

    Sweet! I'd love to hear how it goes, hmu on https://discord.gg/QjhQDK8kAj
13. 193.▲
    by jbellis 1y ago

    Good to see experiences from people rolling out AI code assistance at scale. For me the part that resonates the most is the ambition unlock. Using Brokk to build Brokk (a new kind of code assistant focused on supervising AI rather than aut
14. 194.▲
    by jbellis 1y ago

    I created an OSS ai coding platform as well: https://brokk.ai But it's a different take, Brokk is built to let humans supervise AI more effectively rather than optimizing for humans reading and writing code by hand. So it&#
15. 195.▲
    by jbellis 1y ago

    What I'm seeing with Brokk ( https://brokk.ai ) is that models are not really interchangeable for code authoring. Even with frontier models like GP2.5 and Sonnet 3.7, Sonnet is significantly better about following instruction
16. 196.▲
    by jbellis 1y ago

    I'm kind of disappointed in this article, Splade is a cool way to improve results of a TF/IDF index with minimally invasive changes and this obscures that more than it clarifies. > Next, my SPLADE implementation in Elasticsearc
17. 197.▲
    by jbellis 1y ago

    I'm not sure that's correct -- the first party APIs are priced per query but BD is per 1k results . Not immediately obvious what they count as a "result" tho.
18. 198.▲
    by jbellis 1y ago

    Can you ask whoever owns dashboards to make it so I can troubleshoot quota exceeded errors like this? https://x.com/spyced/status/1917635135840858157
19. 199.▲
    by jbellis 1y ago

    Sort of. The benchmarks showing Flash 2.5 doing really well are benchmarking its thinking mode, which is 4x more expensive than Mercury here
20. 200.▲
    by jbellis 1y ago

    Unfortunately a 4o mini level of intelligence just isn't enough to make this work, no matter how many iterations you let it try.
21. 201.▲
    by jbellis 1y ago

    What is the price on Mercury Mini?
22. 202.▲
    by jbellis 1y ago

    does it actually make a difference to do M rounds of N vs one round of M*N?
23. 203.▲
    by jbellis 2y ago

    I think I saw o3 hallucinate more in a single day of heavy use then I saw o3-mini do in a month. That said: so far, I'm putting up with it because o3 is smart .
24. 204.▲
    by jbellis 2y ago

    All the AR glasses are stuck at 1080p still. :-/
25. 205.▲
    by jbellis 2y ago

    The ones that require base stations like the BB are not very portable.
26. 206.▲
    by jbellis 2y ago

    How about more sample code for the streaming transcription api? I gave o1pro the docs for both the real-time endpoint and the stt API but we couldn't get it working (from Java, but any language would help).
27. 207.▲
    by jbellis 2y ago

    Partly it means you can tell it to do X and it will figure out that implies Y and Z without you having to spell it out And partly it can actually execute more at the same time without starting to make mistakes
28. 208.▲
    by jbellis 2y ago

    Remarkably capable is a good description. Shameless plug: One of the reasons I wrote my AI coding assistant is to make it easier to get problems into o1pro. https://github.com/jbellis/brokk
29. 209.▲
    by jbellis 2y ago

    If you're in the habit of breaking down problems to Sonnet-sized pieces you won't see a benefit. The win is that o1pro lets you stop breaking down one level up from what you're used to. It may also have a larger usable con
30. 210.▲
    by jbellis 2y ago

    A better writeup on reverse engineering CC: https://github.com/Yuyz0112/claude-code-reverse