Side project to create soccer cards. Only La Liga Spanish soccer league for now. I will create more leagues from more countries.

https://kards.paellalabs.com

Any idea on the performance improvements with these LTO?

I started an Application and API Gateway: https://github.com/jmaister/taronja-gateway

I am making it for myself, focused on less config as possible and making it secure by default.

Where is gorilla.bas?

I'll buy it, read it and take notes.

Why is it developed in JavaScript? Wouldn't it be better for the user response to make in a more performant language?

Well, those companies have a CEO, Director of Something, etc...

Which size is good for a Nvidia 4070?

assuming you want to run entirely in GPU, with 12gb vram, your sweet spot is likely the distill 14b qwen at a 4bit quant. so just run:

ollama run deepseek-r1:14b

generally, if the model file size < your vram, it is gonna run well. this file is 9gb.

if you don't mind slower generation, you can run models that fit within your vram + ram, and ollama will handle that offloading of layers for you.

so the 32b should run on your system, but it is gonna be much slower as it will be using GPU + CPU.

prob of interest: https://simonwillison.net/2025/Jan/20/deepseek-r1/

-h

Thank you!! I just loaded it a fits in memory as you said.

I am testing it now and seems quite fast giving the responses for a local model.

Nice, less things to learn for getting the AWS certification.

The problem is not the download size. The problem is the amount of code to maintain.

It's both. All the JS code has to be parsed and interpreted by web browser, which is affecting especially low end smartphones. Better use plain HTML and CSS whenever possible, see e.g. PHOOOS: https://github.com/niutech/phooos (demo: https://kodus.pl)

I don't think the author of the original comment would really care about the download size, given that it was an "internal self serve web app".

htmz looks very cool, though.