Article en

Corrigibility should not be the only target

Source: LessWrong · Ajouté Hier

Many people believe that corrigibility is sufficiently promising to be the primary alignment method people should work on. I outline here several reasons I think that’s not the case. First, it’s hard to imagine AI companies pivoting radically enough to produce AIs trained only for corrigibility. Even with near-superhuman AI-assisted research, presumably developers will know the time pressure and not be willing to try something radically different when their opponents may not. So the realistic choices we make are in worlds where corrigibility/obedience competes with object level values like in

#target

Ouvrir l'original Découvrir plus

Foire aux questions

Qu'est-ce que Corrigibility should not be the only target ?

Corrigibility should not be the only target is a article discovery curated from LessWrong. Many people believe that corrigibility is sufficiently promising to be the primary alignment method people should work on. I outline here several reasons I think that’s not the cas

Comment puis-je y accéder ou l'essayer ?

You can experience this content directly on its original host (LessWrong) by clicking the "Open Original" button.

D'où vient ce contenu ?

Curated from LessWrong by the OnSuggest team to surface the most interesting and serendipitous content on the web.

Toutes les découvertes Culture et idées →

Article · Culture et idées

This Next-Level Nonstick Cookware Set Is on Sale for Hundreds of Dollars Off Ahead of Black Friday

I personally can’t imagine my kitchen without nonstick cookware. But ever since I started learning about the PFAS and PTFEs (aka “forever chemicals”) that are sometimes used for nonstick coatings, I’ve been more intent on picking pots and pans that aren’t made with them. Forever chemicals can have adverse effects on our health, so for the sake of both my health and peace of mind, I’d rather not take a risk. READ MORE...

Kitchn | Inspiring cooks, nourishing homes Voir les détails →