solomatov

As far as I remember there was report of their run rate. How does it align with this?

>you could choose not to jump on the bandwagon

I think it's not an option. The benefits are just too large for me.

>I do know we will never go back to mainly programming through code again, that’s for sure.

I think, in some niches, i.e. where there's something not well represented in the training set, it still makes sense to write code by hand. But I am not sure that it will continue.

How many toks/s could you get this way?

Which remarks? Could you share a link?

It looks like HuggingFace shows Apache-2.0 but they have AUP. How does it work together?

Thanks for sharing.

Super impressive! Is it possible for you to share your methodology of using LLMs?

I tried a similar approach before, but it didn't work for me. I didn't get a lot of speedup if any from it. IMO, to get productivity you need some kind of YOLO mode (in a sandbox).

IMO, the goal should be to outsource as much work to the model, as possible, while minimizing effort required to understand and review what is did. For example: ask the model to find out why a bug happens, figure out proof of concept for thing X, incrementally optimize something, do a well specified refactoring with some guide, and similar things.

IMO, what people say about creating loops is a very similar thing. You maximize the work done by the model, while minimizing the amount you need to do to control it.

Did they mention the license?

> that he doesn't really understand how inference works from a technical perspective

Could you share what tells about it? I.e. where he was wrong about it?

> progress was slowing

Do you think it's not slowing? Do I miss anything really important?

My understanding is that we have now is incremental improvement on thinking models which appeared more than a year ago. Of course, a breakthrough might happen, but I don't see one yet.

Hopefully we will have more information about these companies when Anthropic IPO filing become public. There's too much speculation without them.

Personally, I thought about it as next gen vscode

So, it's not open source?

> but there is so much evidence to point to this not happening

Could you explain this?

>For coding you always want to go with the best model in the category, not something that would be the best model if we went 1 year back which GLM 5.1 is, and I'm saying that as a big fan of GLM cause I run a translation site where GLM is good enough for the price.

Currently, the difference is substantial, but what happens if capabilities saturate?

How central is it in the discrete geometry? Could anyone with the knowledge in the field reply?

Could you share links about this? When it turned into a different interpretation?

It would have been better if they provided not just weights, but also some frontend where it is usable as is.

>but I have seen the local 122b model do smarter more correct things based on docs than opus

Could you please share more about this

Just curious, the fixes are not about weights but about templates, am I right?

Did anyone try it and Gemma 4? Does it feel that it's better than Gemma 4?

I mean not how to do it, it's not that hard, but how to be productive with it.

Does anyone has any tips for starting with Gastown? I am comfortable with couple of agents running, but not yet comfortable with what Gastown offers.

>I’m not sure what you mean by keep it in your head?

If the project I work on is large enough, it takes me some time to get everything I need to understand for review into the short term memory. If it's small enough, it's less of a problem for me.

But you have to keep it in your head, and remember all stuff at the same time. How is it possible to track, and do reviews one after another? Or are these pretty long running agents?

But how do you cover such amount of multi tasking? Could you give an example? I mean what kind of tasks allow such a parallelization?

> more like supervising 8-15 agents

How do they do it? (My own record is 5 agents, but it is not typical). Do they use gastown or something?

Is there any publication which demonstrates that the improvement is really 10x?