

I can’t really differentiate among any of the models cause if one doesn’t work I’ll just toggle to a different model on duckduckgo. They tend to work very well for my purposes, which is locating sources of info that a regular search doesn’t turn up (or the results would take a lot of time to parse through).
Their shortcomings aren’t a big deal for this because if one returns an incorrect result or hallucinates I can just correct it or hop models. And the ability to do fuzzy searches using plain language descriptions engines have never been good at, so in that case the LLMs are straight up better. But yeah otherwise a lot of their role for me is filling in for how bad search engines have gotten.
I only code a little bit to solve my own problems so I’ve left it to people who’ve actually investigated to have an opinion on that angle. But last week I needed to add a quick function to one of my things so I tried one of the models, and it knocked it out immediately with no need for corrections. Less fun than solving the problem myself but woroed in a pinch since I was busy.
I need to look into local models again now that they’re getting good and lightweight. But I’m not sure I’d be able to get much utility since my sole use case right now is search and I’m running an old thinkpad. One of these days I’ll tinker with it.








Yeah lol that’s why I hedged my language a bit, some people find it unsavory
Assuming the bidet does its job and you aren’t reusing towels before washing them, it’s fine.