I think your interpretation at the end is correct.
OK.
Your conclusion still leaves me confused. In any given choice considered by itself, the scale of value also doesn't make sense, you need to compare with something outside of that choice to say that the difference between the available options is insignificant. In a decision problem, you zoom in, not give up. So what is the difference between today and tomorrow insignificant in comparison with? Any personal-level change is much smaller, likely also much smaller than predictable differences in the world if I had bothered to find them. So to make that judgment, it seems necessary to involve something like the (low) expected impact of your actions on progress, and at that point I lose track of the (hypothetical) argument.
The conclusion is that the effects of progress are small compared with anything that has an appreciable effect on the future (for someone with aggregative, time-insensitive values). If we break down an action as the sum of two changes---one parallel to progress, one orthogonal to it---the effects of the orthogonal part are typically going to be much larger than the effects of the parallel part.
Originally I ended with a discussion of x-risk reduction, but it became unwieldy and I didn't want to put time in. Perhaps I should end with a link to some discussion of future-shaping elsewhere.
Over the last few months I've started blogging about effective altruism more broadly, rather than focusing on AI risk. I'm still focusing on abstract considerations and methodological issues, but I hope it is of interest to others here. Going forward I intend to cross-post more often to LW, but I thought I would post the backlog here anyway. With luck, I'll also have the opportunity to post more than bi-weekly.
I welcome thoughts, criticisms, etc.