But how do you prompt this smaller model to give back information? I've tried that in the past but didn't go well. What I do is to send written handoff files between models to pass context around, but only had good results with big vision models as well.

So many possibilities for how you could glue it all together. However, when I send Gemma 4 12B in llama.cpp an image with no accompanying text, it assumes I want a description and gives me one.

I just tried with an audio file, and it transcribed the lyrics as I hoped. Then it made a bunch of suggestions about what do next, which I wasn't after. I could probably fix that by sending some text to narrow the scope.

This is a common issue, and most developers I worked with are not aware of it until they see the performance issues.

Most people are not aware of how Dapper maps types under the hood; once you know, you start being careful about it.

Nothing to do with LLMs, just plain old learning through mistakes.

Avalonia version of my terminal Redis client https://github.com/cosmez/RedisMan

> I am never going back.

For some reason, it sounds like someone being held against their will, which is exactly how I feel about it too

I don't think diversity in a software ecosystem is a bad idea. Nothing is stopping Hugging Face from changing their terms and screwing us over

There are many pull requests trying to implement this feature, and they don't even care to reply. This is the only reason I'm still using llama.cpp serve instead of this.

wouldnt it be more practical to make a PR for llamacpp to replicate what Ollama does well instead?

wow, i had the same experience. reaching my 40s and decided to drop all the cool kids stuff. i used to have emacs/scheme evangelist phases in my life, now i do all my coding using enterprise languages and tools.

i have never been so productive.

How is it going to be worse? He's lowering taxes in the border and trying to convince Central American immigrants to stay in Mexico.

He supports Maduro.

I've been working for the same company for 12 years, same role, same position.

Why did i stay? i was the only dev for like 8 years, being responsible for your work and not being able to blame someone for legacy code is an eye opener.

it makes you love your craft even more, i joined to start a port, turns out this porting process never ends (I like improving my own code).

right, actually one of the topics were i stopped on my first readings.

https://mitpress.mit.edu/sites/default/files/sicp/full-text/...

Contextualizing this with Hoare's work on pre and post conditions helped me to understand this chapter. Predication is fundamental to advanced computer science.