Two points:

1. Why stop at insects, why not write this same article about demodex mites, earthworms or krill?

2. I think there’s a big reason why the concerns of insects and smaller animals are dismissed that you haven’t touched on, which is that any consideration of these animals leads to absurd conclusions, like that every moral pursuit of humanity up to now is actually meaningless compared to improving the lives of insects. Most people can see that this is not a fruitful avenue of thinking.

I think you’re underestimating the average person by suggesting that the only reason they’re not interested in insect welfare is entrenched social norms. Whereas there were reasonable alternatives to slavery, and there are reasonable alternatives to factory farming, I think the average person can intuit that there’s no reasonable alternative to just politely ignoring the suffering of the quintillions of insects, worms and mites on the planet.

Consequentialists should be strong longtermists

Disagree on the basis of cluelessness. 

Uncertainty about how to reliably affect the longterm future is much worse than uncertainty over our effects on the near-term.

I find the Hilary Greaves argument that neartermist interventions are just as unpredictable as longtermist interventions unconvincing because you could apply the same reason to treating a sick person (maybe they'll go on to cause disaster), or getting out of bed in the morning (maybe I'll go on to cause disaster). This paralysis is not tenable.

Vote power should scale with karma

Risks groupthink (this is hard to prove). Also bad optics to outsiders (looks liable to groupthink, cultish).
The benefit that it makes sock-puppeteering harder means I'm a little ambivalent

There's a distinction in theory but in practice the vague definition of "sentience" is so tied to moral relevance I don't think you can argue for one without also arguing for the other.

The question "does a worm feel pain" isn't really asking "does the worm have nociceptors and some degree of integration of those nociceptive signals that causes learning and behavioural changes". It's really asking, at the core, "does a worm 'feel pain' in a way that's morally important"

Uncertainty about the net utility of wild animals is also true of human life. It’s an open question whether the average human life is net negative or net positive.

Would you therefore also say that propagating human lives on other planets is “extremely bad”?

You could ask the same question about worms, mites or nematodes.

I think the reductio ad absurdum that if any of these things matter even slightly, then all human moral concerns become completely irrelevant, means we need a high bar for believing this

That’s a verbose way of saying: “looks like it feels pain, probably feels pain”. Invoking Bayes’ Theorem gives the argument a false depth.

Being unnecessarily verbose comes across very negatively in EA communication, important to avoid it

“it at least somewhat increases the risk of animal life being propagated on more planets. This seems extremely bad, since we have no idea how to ensure that those animals will live good lives.”

Do you assume that wild animal life is net negative?

If given a magic button that instantaneously wiped out all wild animals, ignoring the consequences for humans of doing this, would you press it?