Environment is an interesting example because you go from complete poverty (no environmental impact) to middle income (rampant growth, environment not a priority, think Brazil/Indonesia and their rainforests, or manifest destiny USA and their forests) so impact worsens, then at high income concerns about environment become more of a priority so you get  environmental protections.

Unless the goal is to prevent people rising out of poverty entirely (it shouldn’t be) the best outcome comes from faster development

I think it serves 2 purposes:

1. Most people want to feel like they are good, kind. Preventing harm to something much smaller/weaker than themselves reinforces this. Even better if it requires very little effort.
2. Social signal. I personally immediately trust people more if they take their spiders outside rather than kill them. I think they're more likely to have good intentions in whatever else they do. I think many people feel the same way and are vaguely aware that carrying themselves like this sends a useful signal to others.

You're preaching to the choir here on the EA forum but I think most people outside this community will intuit the slippery slope that this takes you down:

0.1% x 5 million lives saved is the same EV as 0.0000001% chance of 5 trillion lives saved

Somewhere between those two this becomes a Pascal's Mugging that we seem to generally agree is a bad reason to do something.

Where's the line?

You're skeptical that concerns for animal welfare track with socioeconomic development? The animal welfare movement has arisen and mainly operates in rich countries

Progress studies and longtermism sound good in theory and then in practice they don't seem to have produced anything beyond theory, which is not that helpful.
The randomista movement that produced the RCTs GiveWell based its recommendations on was a response to the longstanding failures of development economics to actually make an impact on development.

The meat-eater problem is an issue with people going from not being able to afford meat (dreadfully poor) to being able to (only slightly less poor).

I’m talking about development beyond those very low levels

Global development is an animal welfare issue. The wealthier a country is the more free time and resources the population has to entertain the idea that animal torture is bad.

If you or I were living in a favela in Brazil struggling to get by we probably wouldn't have animal welfare on our radars as a political concern. We'd have bigger problems. Give us a comfy middle-class life and maybe we'd have room to care.

Having a strong precedent for this in countries like the UK and trying to nudge foreign standards with welfare-based import controls both help, but development is critical.

Seems to be true. Had assumed that a few hundred k years would not be enough for this. Changed my mind

As a datapoint against “what we evolved to eat is what we should eat”: our bodies also aren’t evolved to eat cooked food. But cooking food (meat, milk at least) seems to be better for us than raw.

1. The strongest arguments in those areas are by analogy. Analogising is much easier with fellow humans or even animals close to us than worms. “If you were X, you probably wouldn’t want to be discriminated against” or “they probably suffer like me, we should avoid that”. This starts to break down around the level of shrimp and then is completely broken by the time you get microscopic.
2. Nematodes being morally significant is far more disruptive and absurd than any of these ideas. Accepting that the Old Testament can be ignored on yet another moral issue is pretty easy. Accepting that human welfare is a rounding error compared to microscopic worms is society-upending.