Ah shucks. Guess it's time for bed.

Show Post

And, probably most importantly- what if the tools work perfectly, but the monstrous folks in charge decide that they need all messages to be 30% more patriotic and 10% more fascist? We've seen that they'll just shove that out live and untested. Is this awful oligarchy worth entrenching?

Show Post

What if the summarising AI misreads a joke as serious or just fully inverts the meaning of what you said and their human-written response is completely misguided? What if they read your message in full and then their AI prompted email went wrong and you think they ignored it?

Show Post

Like, if someone replies to you with a seeming non sequitur, is that because that's how their brain works, or because their email summarising AI hit a snag and now their email writing AI is functionally trying to find the seahorse emoji? If they're not using AI, those aren't possible problems.

Show Post

The way I see it is this: if someone uses genAI to handle any part of communication, I fundamentally cannot trust that communication anymore. I don't know whether my words are being read, or whether the other words are being written, or even read. But I know they will blame the machine for mistakes.

Show Post