this post was submitted on 28 Apr 2025
26 points (100.0% liked)
technology
23877 readers
210 users here now
On the road to fully automated luxury gay space communism.
Spreading Linux propaganda since 2020
- Ways to run Microsoft/Adobe and more on Linux
- The Ultimate FOSS Guide For Android
- Great libre software on Windows
- Hey you, the lib still using Chrome. Read this post!
Rules:
- 1. Obviously abide by the sitewide code of conduct. Bigotry will be met with an immediate ban
- 2. This community is about technology. Offtopic is permitted as long as it is kept in the comment sections
- 3. Although this is not /c/libre, FOSS related posting is tolerated, and even welcome in the case of effort posts
- 4. We believe technology should be liberating. As such, avoid promoting proprietary and/or bourgeois technology
- 5. Explanatory posts to correct the potential mistakes a comrade made in a post of their own are allowed, as long as they remain respectful
- 6. No crypto (Bitcoin, NFT, etc.) speculation, unless it is purely informative and not too cringe
- 7. Absolutely no tech bro shit. If you have a good opinion of Silicon Valley billionaires please manifest yourself so we can ban you.
founded 5 years ago
MODERATORS
Already on ollama.
I've found Qwen preferable to DeepSeek for coding so I can't wait to try this out
I've not used Qwen yet, but I have noticed deepseek, specifically r1, is kind of a lazy coder. Lot of 'step 5 draw the rest of the owl' type responses.
Unrelated, but does anyone else's internet speed come to a screeching halt when trying to download models from ollama? I swear I'm being throttled by xfinity.
That might just be LLMs in general. ChatGPT does the same. Copilot is a little more well-tuned, but I really only ever have it do boilerplate.
I've had really good luck with chatgpt 4o, and, to be fair, I have teased some decent responses out of deepseek 3 (iirc). Different ways of expanding on the basic principles of asking it to 'step back and visualize different options before moving forward and fully implementing them with all necessary code, following best practices, etc.' tends to get pretty good results.