I find it hard to believe that any scientific breakthrough could prevent everyday people from turning more anti-AI over the next few years.

And I don't think this depends on AI pushing up unemployment. I think it depends on the gap between when breakthroughs pay off for companies and everyone else.

For example, a fusion or cancer breakthrough probably wouldn't improve the average american's life within the next five years, as infrastructure takes time to build and drugs have to go through trials. But the AI lab and partner companies that made the discovery would feel it immediately, in their market caps and in massive windfalls.

Something similar already happens with pharma. People hate pharma companies, but still use their products.
The same will probably happen to AI companies, except it will happen across many industries at once. This will take shape in the next presidential elections.

The more successful AI companies become, the more pushback they'll get, in the form of regulation or taxation. And I don't think AI companies should really resist that. They will just make themselves more unlikeable. Both USG and labs should push hard on distributing the benefits, through something like equity based dividends for every citizen. Labs should also prioritize innovations that pay off for people quickly, and ones that shrink the pay-off gap.