AI - the biggest decision

[wtg]

Do we let it train itself?

https://www.theguardian.com/technology/ng-interactive/2025/dec/02/jared-kaplan-artificial-intelligence-train-itself

[ShiroKuro]

Thanks for posting that!

Self-training AI is a bit scary to me. We (or rather, AI trainers) already don’t fully understand how it does what it does. And quality control, checking for veracity etc. is already an issue. How will we be able to evaluate AI that’s self-trained?

[jon-nyc]

Unfortunately this is a situation where the broad question is not ‘do we aggressively pursue AI’, it’s do we aggressively pursue AI too, or just let China do it’.

The later likely results in a global technological panopticon - a perfection of power which renders its exercise unnecessary, to paraphrase Foucault. I wish it were otherwise.