the words 'stifle innovation.' My bet is that it will fit nearly every time.

cause. Be aware you're meeting your audience's emotional need (mine, too), not just providing fact-based analysis. Your often sneering tone plays to that need, and the accusation of you being substance-free may be picking up on that vibe.

I was probably wrong about you making a capabilities misprediction. Let me take another stab at it:
Big Tech is generating a lot of fear, anxiety, and resentment, such that even the thought of the AI bubble bursting is comforting, despite the mayhem it would

the bubble, but without acknowledging that AI is more capable than you predicted it would be. The humility gap makes the bubble work less credible, which is a shame. Maybe that's what the "substance-free" dig was getting at?

Definitely not substance-free. A more fair criticism is that, having built a brand around AI skepticism, you seem highly motivated to find evidence to support your skeptical stance and uninterested in falsifying it. IIRC your original position was that AI was not capable. From there you pivoted to

parental rejection of one's sexual identity. Parental love cannot be legislated, and the therapy ban seems a weak proxy that isn't worth the political capital. I am open to being convinced with data showing how conversion therapy bans actually save lives.

Not making light of anything. On the contrary, making much of all the other bad things.
My position is partially based on the hypothesis that, while conversion therapy is part of a constellation of harmful behaviors, a ban on the therapy itself doesn't address the obvious source of harm,

Not trying to pile on, but the inequality dimension didn't make it into the body of the pre-print paper, only the abstract. I've alerted the authors about this error, but the takeaway is that the paper's findings may not match your expectations (overall I found it to be pro-business/automation).

That's childish malice, hurting the cause.

introduced by the authors to interpret the data, and however valid it may be, pointing to it exclusively risks reducing moral repugnance to soft vibes and irrational biases, with no reasonable prudence component.

I agree that out-of-scope topics are not a fair target for criticism. It's the "strongest moral cluster" claim that I'm disputing. Police and judges are right there in the top 20 (before funeral directors), but they're ignored in the discussion. The "care and spiritual authority" category is

And what about private security services, prison guards, soldiers, policy-makers, and other workers performing tasks exercising power over people?

There's a big gap in this report: it seems interested in the ethics of care (clergy, therapists, etc.) but entirely uninterested in power. The data show a high repugnance for automating police officers, but this data point isn't highlighted in the chart.