People are looking for ways not to burn through their premium subs when in many cases all you have to do is move down to 5.4-mini codex and it will probably solve your issue while barely touching your 5 hour or weekly limits.
Yes, if your trust model allows you to use API providers or the big 3, you 100% should. They have better util than anything you self host, so they can be more efficient. On top of that, they're shoveling cash into the fire to try to capture marketshare, so they're offering inference for well below break-even costs.
The main reasons to use local models are:
1. Self-sovereignty & control
2. Data security
3. Offline availability
If none of those apply to you, then you should just use OpenAI or Anthropic.
Because there isn't a causative link in any literature available anywhere? (If there is pls cite). That's like saying "caffeine affects your brain you should stop taking caffeine"
I've read that New Zealand study and it's identifying an effect in a small group of heavy users who started using cannabis very young (a full third as teens) and continue to use it heavily (four times a week) three decades later. That's not a huge surprise.
They also state "cognitive functioning among midlife recreational cannabis users was similar to representative cohort norms". It's clearly shown in Figure 1 - midlife recreational cannabis users actually got smarter than people who quit cannabis according to their data.
There may well be some confounding factor in there. The study was done in New Zealand where cannabis was illegal for the majority of participants, and usage was self-reported so there's a basic issue there as well. One of the meta-analysis citing this (Crisafulli, 2026) and finding no effect criticizes the study design.
can't imagine a world in which I'd download a little known distro to put on my home network and use as a server. also, doesn't fedora already have something like this already?
Game’s truly gone. I remember when all we did was try to find the most obscure indie band of a Linux distro, form emotional attachments and then argue their merits.
Of course you can choose a general purpose Linux dist. But you'll still have to go through the extra work of partitioning the disk, installing the system, selecting packages, possibly follow some other guide to install Docker Engine, Docker Compose, Docker Buildx, and doing that thing with the docker group. At that point you'll pretty much have what Lightwhale provides out of the box. Except you'll be left with a system that isn't immutable, and where system and data is entangled on one filesystem. And if that's your thing, you should do definitely that. But...
If you just want to run your containers, then Lightwhale is best in class.
I wouldn't trust it. When I do check its work, I often find factual or corectness errors. No way it's going to be the last step of defense against its own mistakes. I mean for me. Other people seem to have more luck. I'm probably still holding it wrong.
That's my point - it's great as a tool to talk something through or rubber duck it, but as soon as you just let it loose to slop out thousands of lines a day and never read them all you're really doing is filling your base with thousands of lines of technical debt.
