Their strongest case:

• Nobody has ever been good at predicting which technology turns out dangerous. Preemptive bans have a poor record.
• The cost of restriction is certain. The harm it prevents is a guess. Slow something down and you know exactly what you gave up — the treatment that didn’t get developed, the tool that never shipped. The disaster you avoided might never have happened at all.
• The harms are speculative, hypothetical and sits somewhere in the future, if it even happens at all.
• Whoever controls the model controls the answer. A company setting its own rules is fair enough — you can go elsewhere. But when three companies set nearly identical rules, “go elsewhere” stops meaning anything. The answer isn’t fewer companies. It’s more.
• Nobody made email mandatory. It just won. That’s not force — that’s a lot of people quietly agreeing.
• Adults get to make choices other people think are unwise. That’s what the word means.

The strongest case against it:

• You can recall a car. You can’t recall a file. Once a model is downloaded, every copy is permanent — including the ones on machines you’ll never know about.
• The person harmed usually didn’t opt in. An AI generated image of you was never your choice.
• “Let people decide” assumes the one bearing the cost is the one deciding the rules.
• And: your attention is already being optimized against. How free is a choice made inside a system tuned to your reward circuitry?

Now that you read both sides — then tell me what I got wrong?