flux3125

And now he's part of the confusing comment section lol

Max 60 bucks a month. More than that and I'd just move to local qwen 35b or some other cheaper model on openrouter.

This reminds me of Coconut Simulator

Sorry, I forgot to add "on Linux" at the end. Still, that's a nice one!

I'm curious if there's a program that uses a simple detection model for UX components to locate clickable areas. This would allow for global navigation similar to VimiumC

It's called Time Series Forecasting

They'll not only do it, but they'll also wrap it in a huge fat rounded border

It's also never deleted anything it shouldn't or done anything crazy or unrelated to what I asked

Until it does. A simple curl request to a compromised website could inject a malicious prompt into it.

Don't listen to him

That's the idea

You're absolutely right!

At least it doesn't know my age

Oh wait

Point 2 is so true, I definitely find myself spending more time reading code vs writing it. LLMs can teach you a lot, but it's never the same as actually sitting down and doing it yourself.

I was (positively) surprised that this didn't involve any AI.

Code quality and IDEs aren't going anywhere, especially in complex enterprise systems. AI has improved a lot, but we're still far from a "forget about code" world.

They are okay for vibe coding throw-away projects without spending your Anthrophic/OAI tokens

You ARE the training data

I stopped using Ollama a couple of months ago. Not out of frustration, but because llama.cpp has improved a lot recently with router mode, hot-swapping, a modern and simple web UI, MCP support and lots of other improvements.

so llama-swap can load

Just in case you haven't seen it yet, llama.cpp now has a router mode that lets you hot-swap models. I've switched over from llama-swap and have been happy with it.

In my experience if you're coding or doing something that requires precision, quantizing the kv cache is definitely not worth it.

If you're just chatting or doing less precise things it's 1000% worth it going down to Q8 or sometimes even Q4

but with people throwing money behind it, there’s a plan to extract value from the user base from day one.

They'll start injecting ads in your commit messages, forcing you to subscribe to a premium plan.

Finally Claude Code can now control Claude Code

dunno why you'd need such a silly thing though

I'm not sure I follow, what alternative to CUDA on Linux offers similar performance?

I remember someone on HN once saying they treat LLM agents like human coworkers, security-wise, and that stuck with me.

You don't give your GH keys, email credentials and ssh keys to a coworker. They have their own accounts with scoped permissions. Need them to read an email? Forward it. Need them to work on a repo? Add them as a contributor and enforce the same branch policies you would for any human.

There are still risks, but they're similar to delegating work to humans, so it's up to you how much access and trust to give.

Interesting. If you look at the sources it cited, there are a few links about "Sacred Songs and Solos" (likely from related/side content on the page), my guess is it didn't read the main article and instead anchored on those and hallucinated

to be fair, llama.cpp has gotten much easier to use lately with llama-server -hf <model name>. That said, the need to compile it yourself is still a pretty big barrier for most people.

That’s not what it means. "-it" just indicates the model is instruction-tuned, i.e. trained to follow prompts and behave like an assistant. It doesn’t imply anything about whether thinking tokens like <think>....</think> were included or excluded during training. Thats a separate design choice and varies by model.

You're not doing anything wrong, that's expected

I bet the paper was vibe written too

An error occurred. Try again.

But seriously, OP should somehow change this message to something like "Too many people are chatting right now, please try again in a moment."

(that would be even more appealing to recruiters)

I completely removed nanobot after I found that. Luckily, I only used it a few times and inside a docker container. litellm 1.82.6 was the latest version I could find installed, not sure if it was affected.