badsectoracula

Not sure if it is competitive but DeepSeek's models are under the mit license, including their most recent one.

The model itself (Qwen 3.8 27B specifically) can handle being interrupted just fine and i often interrupt it to clarify things when i see it finding my instructions vague or to add some extra info.

However there is a bug somewhere between my instructions and the model - i think it is the jinja template, though i'm not sure as i haven't actually looked too deep into it - that in some cases make my last message be sent repeatedly to the model after each tool call, making the model think that i'm getting impatient or something :-P. When i notice it, i tell the model that this is some harness bug and it can ignore the repeated messages, which it does.

It seems to happen when i stop it while it is thinking after a tool call but i'm not certain. I try to stop it while some tool call is in process so it cancels the call as that seems to avoid the issue. Restarting llama-server also seems to make the issue go away so it might be a llama.cpp bug, but TBH i haven't found a clear repro case aside from it happening when i interrupt it while it is thinking. I think it is more common when the context is close to full, or at least i mainly remember it happening when i've been using a session for some time (i.e. not in fresh sessions, which does make reproducing it a bit harder).

In any case, the model itself seems to deal with all the above fine (and accepts the bug explanation for repeated messages), it is the rest of the stack that can show an issue with interrupting it (but even that is relatively rare).

As i wrote in the other comment, i know, i just meant is that the process i use isn't driven by a harness but done manually. The PDF i quoted mentions a harness does that driving so i wrote "There isn't any real harness driving this" to mean that i do that by hand instead of being done automatically.

I can't edit it to clarify, but what i meant is that the process i use isn't driven by a harness but done "by hand", not that there isn't any harness at all. The PDF describes a harness that is actually used to drive the manager/worker role.

Yeah though AFAIK it is discontinued now (or at least i can't find it in the local price aggregator). I think R9700 is a decent alternative, a bit slower (256bit bus vs 384bit bus) but has 32GB of RAM, still isn't too overpriced (at least compared to Nvidia alternatives) and has an opensource stack too (which is why i went with AMD in recent years).

Interesting but (from the PDF):

    Every role but the verifier is the same
    underlying model, called in a fresh context
    and coordinating through a shared workspace on disk:
    
    <ws>/task.md - the problem statement
    
    <ws>/plan.md - the manager’s overarching plan
    
    <ws>/tasks.json - a debug mirror of the task list
    
    <ws>/notes.md - curated ideas / findings / partial proofs
    
    <ws>/solution.py - the current solution

Then they describe how there is a ping-pong between manager and worker with fresh context for the worker, but i wonder, is that part or even the worker really necessary?

My use of Qwen 3.8 27B (which is the only model i use almost since it came out) is really doing something very similar to the above where first a task is described, a `PLAN.md` file is made with the (current) plan and tasks are either in the plan file or in a separate issue tracker the LLM has access to via a CLI tool, a `STATUS.md` that tracks the current status and a `GOTCHAS.md` file for writing down various notes/tidbits/gotchas (TBH this is largely because Qwen 3.8 27B does want to note down stuff anyway so it mainly acts as an attractor for not writing its notes down in the other two places :-P).

There isn't any real harness driving this, i just give it a task, tell it how it'll be handled and off it goes. I guess i'm playing the "role" of the harness there as i tell it to not go and do everything in one go :-P but i never bother with fresh contexts, Pi's "rolling" approach to compaction (i.e. summarizing the older messages to make space while keeping the recent messages as-is) make it almost a non-issue.

I have considered making it a planning tool, though as some tool it can use via the CLI (it is running in a container), not a harness - but at the same time it feels like it can do the same thing with the `PLAN.md`/`STATUS.md`/`GOTCHAS.md` triplet. The only thing the tool would help with is to provide a UI for me to check what it's currently doing, but even then i usually just stop it and ask it directly :-P then tell it to continue.

Perhaps at some point i'll make some custom benchmark with stuff that are vaguely similar to what i work on and try a bunch of different approaches. It is something i want to do anyway to see how and if Qwen4 27B (whenever that comes it) is an improvement over 3.8.

How so? Judging from the examples the model does seem to take advantage of the language (creates and uses reusable elements) and AFAICT the language itself compiles down to "real" Sierra PIC commands so in theory it can be used with a Sierra PIC viewer/engine.

Looks like MOSS was made by some Chinese university as part of an AI education/research program so i guess they don't have to care about funding as much as the other labs :-P

Looks like the people over /r/SD seem to like MOSS (i'm not up to date on the image stuff) but what i notice is that it is Apache2 licensed when every other openweight model seems to have some weird license (from a quick check Qwen-Image even downright forbids any commercial use at all).

It isn't just the ancient sculptors and artists (unless 20 years ago is ancient :-P), when i went to an art school there were nude drawing lessons there too.

Yes, an agent wont tell you if something is boring or funny, at least not unless you ask (and i don't think its response would have much value) but it will tell you if your instructions can be followed or not - which was the main thing the author wanted to test against, at least as far as i understood.

My point wasn't that you can use agents for all feedback, but for the given case of testing your readme file's instructions (the article's title even mentions it is about the readme file) you can certainly use agents for that.

I know there are technical ways to work around IRC's limitations, my point is that their mere existence adds more friction to using IRC so people move on to more frictionless places - and with chat apps being fundamentally social, that friction affects people who don't even mind the frictions themselves but their peers do (though personally i do mind these frictions - i do not want to bother with hosting anything or setting up bouncers or anything of the sort, i just want to pop in a channel, see what people were discussing, talk with whoever is there and/or share some of my own stuff and that's it).

How? From a quick search don't see anything in the spec about files or images. The closest i can think of is attaching metadata (which is a draft), but these are meant to store utf8 strings and i doubt a base64 encoded video is what the writers of the spec had in mind.