phainopepla2

Sorry, I meant non-locally.

I'm assuming privacy is not a concern since you mentioned using Deepseek already. The cost of V4 Flash for small tasks is so minuscule as to be almost free, and you don't have to deal with a churning laptop (or even buying a high-end laptop, for someone who doesn't already have one).

I guess what I'm really asking is, what's the advantage of using these small local models if privacy isn't a concern?

Why not just use DS V4 Flash for the small stuff? Very fast and extremely cheap.

Have any major open weight models been "open data"? Wouldn't that entail distributing vast amounts of copyrighted data?

That's because Qwen's flagship models are not, in fact, open weight. Qwen3.7 Max, Qwen3.7 Plus and others are closed weight.

You can use Qwen3.6 35B A3B (for example) on Openrouter with a US-based ZDR provider, because it's one of their open weight models

Insurance companies spend a lot of time on bill review and disputes over coding. They will develop an adversarial AI process to counteract this (assuming they haven't already). Race to the bottom and we all win

He allowed allowed the destruction of Gaza to start, and did nothing to stop it even when it was clear what was happening and it was within his power to do so. Of course, Trump would have been no better on that front. But that brings us back to...

The lesser of two evils is still evil

This is very true. My parents were organic farmers back before there was any legal or official designation, and keeping synthetic pesticides out of the soil and watershed was a bigger motivation for them than fear of eating conventionally grown food.

I think as organic food became more mainstream it also became more about an individual's health and less about larger environmental and social concerns, but those concerns are at the heart of the original organic movement.

Ah, missed that. Well, I think the point remains that it's an abominable and misleading graph

And to top it off both Y axes have the exact same label. The graph has no indication as to which line belongs to which axis. Of course, we can easily infer which is which from the surrounding article, but still... terrible data visualization

What's forcing people to keep eating terrible diets and getting fat? What's forcing <bad habit> that results in <bad outcome>. The answer is usually human weakness or someone seeking to make a profit by encouraging the bad habit, or a combination of the two.

As individuals, there's nothing we can do other than exert willpower, and we should. But as a society we should also be asking questions about the supply side of these bad habits.

Where can I read more about this? I've been pilled on the co-sleeping thing, but hadn't heard this about "back is best"

Meth is already incredibly cheap. Meth addicts would definitely still engage in this behavior even if meth were free, because they still need money for rent, gas, food, whatever.

My apartment was broken into once and I knew it was a meth head because they color coordinated my closet before they left. Turned out to be someone I knew (on meth)

Wood construction is not typically considered woodworking, although there is often a lot of overlap. But the skills needed to make furniture are pretty different from the skills needed to make decks, fences, etc.

> Is there an actual quantitative check that says "AI or not AI"? I'm genuinely curious.

There are plenty of them that will give you a number, Pangram is a commonly-used one. Of course, whether they actually work well is a different matter. In my experience they have a huge false positive rates. I haven't tested the inverse.

I think pay per account (like the Metafilter model) would be better. If you charge $10 for lifetime access and ban slop-posters, it quickly becomes expensive to be a slop merchant

Then they should use an appropriately scoped API, as OP suggested.

I am experiencing this using Opencode. Caching works fine via Deepseek API but not so good via Openrouter

In the case of something like a world war, which is the type of scenario we're talking about here, I think people would begrudgingly accept that bananas and coffee are unavailable or very expensive.

I'm highly skeptical of DeepSWE. It rates GPT-5.4-mini as three times better than deepseek-v4-pro, but every time I use GPT-5.4-mini I find that it completely sucks at following directions.

If these were in the internal evals then the output would be much better. The 4.8 pelicans are pretty meh

And how are we meant to look at Mythos? Do you have access?

> humanic

No, thank you

> we should just be labeling it accurately to avoid becoming confused about what we are doing.

I think you are doing quite the opposite, and your overexpansion of the term obfuscates things rather than clarifies them. As another user wrote, there is a perfectly good word that covers all your points: ideology.

And that way you don't get the side effect of claiming that cultural food preferences are religion, since they also can't be scientifically validated.

This is a strange definition of religion, to basically mean anything that isn't science. Are all aesthetics and ethics a matter of religion?

It's also the case that respiratory viruses predating covid often cause temporary loss of smell, it just wasn't much in the public consciousness before covid

I have also been woken at night by the smell of smoke from a wildfire. Could be only possible during the lighter phases of sleep I suppose

Even those of us with a good sense of smell often can't smell ourselves, so I don't think you're at too much of a disadvantage there

Sometimes gatekeeping is a good thing. I don't mind being gatekept from some areas of life, not everything is for me. Mass tourism absolutely has made some places less pleasant to visit, and more importantly, less pleasant to live in.

> There's always a limit

That's true, but the limit isn't fixed, it depends on their ability to keep things under control. AI is poised to increase that ability if we allow it to develop entirely under the control of private interests.