ai-christianson

12 ms·

181.
by ai-christianson 2y ago
I think it's more CCP that's the problem not Deepseek itself. They don't have any choice with that stuff, do they?
182.
by ai-christianson 2y ago
Right if I'm doing a typing test, I can do 150 on a good day with good posture and being warmed up. Typing tests are really picky about accuracy --if you have a single typo, it'll kill your WPM and they typically won't let yo
183.
by ai-christianson 2y ago
Arch on a Thinkpad is pretty good. Ironically, arch has been a better "stable" distro for me than Debian, which I used for years previously. The problem with the stable distros, when used as a desktop, is you inevitably get into a
184.
by ai-christianson 2y ago
Personally I don't get sick of it. There's a lot of hype around Deepseek specifically rn, but to run SOTA or near SOTA models locally is a huge deal, even if it's slow.
185.
by ai-christianson 2y ago
It's it a good rush to want to run open models on your own computer? This isn't like Bitcoin mining where there is some direct monetary reward for hoarding compute.
186.
by ai-christianson 2y ago
Couple of notes: - 96gb is not in the default numbers. This is a super common amount (4x24gb) - The custom params input has bad UX where if I select the text and type 200000, it insists on putting a 0 in front (0200000)
187.
by ai-christianson 2y ago
If you want something like composer but want to stay fully FOSS/CLI oriented, check out https://github.com/ai-christianson/RA.Aid
188.
by ai-christianson 2y ago
Just realized RSS+LLM might be a really nice combo. Use RSS to get the full take then use a local LLM to filter out the noise and customize the feed to one's personal preferences.
189.
by ai-christianson 2y ago
Seems like an ideal compression method for LoRa/Meshtastic-style communication. An LLM wouldn't run on an ESP32, but there are several that could run on a raspberry pi. It's not just natural language that could be compressed
190.
by ai-christianson 2y ago
Would you consider giving these crawlers access if they paid you?
191.
by ai-christianson 2y ago
I think they're real. The model is performing better than claude-3-5-sonnet-20241022 on the claude leaderboard: https://aider.chat/docs/leaderboards/
192.
Show HN: I put aider in a ReAct loop and it works well(github.com)
1 points by ai-christianson 2y ago|0 comments
193.
by ai-christianson 2y ago
It was a response to the approach of carefully planning and specifying every aspect of the software up-front. The problem with that "waterfall" approach is that, in the real world, we often work in fast moving/evolving proble
194.
by ai-christianson 2y ago
> I certainly see a point when people switch to XFCE from Gnome or KDE. But what about if coming from the other side of the spectrum? I think you basically answered it right there. It sits between. For the lxde people it would be more of
195.
by ai-christianson 2y ago
There's a lot of nuance in how AI is used vs. the fact that it is used at all. Which specific AI tool is used matters a lot as well. For example, LLMs can generate a ton of slop, whether in writing form, code, or anything else. But
196.
by ai-christianson 2y ago
Is this worth going through the effort of hosting?
197.
by ai-christianson 2y ago
Tried it out. The ideas seemed pretty good. The downfall was the names. Has anyone else noticed that? The AI (in general, not just this product) seems really bad at coming up with names for anything.
198.
by ai-christianson 2y ago
Are you finding building MCP integrations to be worth it? We've been using agents (e.g. langchain), which are pretty good at bringing in context and taking actions. Tool results become part of the context --it just works.