@bradrcarson@ahall_research Thanks for that detailed insight! I feel this takes me back to Arrow's impossibility theorem. For me, the biggest potential upside of AI is lowering voters' costs to politically engage, increasing transparency on inflated bills or regulations. Good problems to think about!
@ahall_research Open weights models do not solve steering by themselves. Aside from a capability gap towards frontier models, their bias could be baked in via the training process. Fine-tuning comes at quite a meaningful initial cost - although I've read about portable tuning lately. We'll see.
@ahall_research Currently writing on that topic as well, inspired by you. For the first issue we need to be very aware that many of our less strong preferences are constructed upon elicitation. Since that AI deals with our rational ignorance that makes steering a lot easier. 1/2
We are now in a position where a tiny proportion of the population uses Fable or soon GPT-5.6, while everyone else's experience of AI is 8-30b-model level - Google's AI Overviews, Meta AI, ChatGPT free tier, maybe MS Copilot at best. People outside of tech must be completely baffled how this is supposed to take their job, and annoyed that hundreds of billions are being poured into it.
@D_Langenmayr@faznet@NadineRiedel2 Wie gesagt, ich gehe voll mit, dass der Aufwand für die Steuererklärung gesenkt werden muss. Ich sehe nur die gleichen Probleme wie im Artikel mit der politschen Umsetzbarkeit.
@WernerderChamp@D_Langenmayr@faznet@NadineRiedel2 Alles richtig, aber man muss sich natürlich klar machen, dass einige dieser komplexen Regeln zumindest mit dem Motiv der horizontalen Gerechtigkeit erschaffen wurden.
@EconTalker I am not sure that holds up for football as coaches are not usually judged on these metrics, but wins, titles etc. Take Germany's last game. If they hadn't been denied that 2:1 in overtime (or won the penalty shoot out), their coach would not face possible dismissal now.
That Wall Street Journal article about GLM catching up with Mythos (which is not true & the reporting doesn’t back up) is another one of those “everyone will ask me about it at every conference or meeting” articles. Big impact on the policy zeitgeist, even if not fully accurate.
There are so many bad videogames for childreen abusing gambling mechanics as @ahall_research pointed out a while back. I suggested just building your own with AI, so I did for my 4-year-old. You can test it here:
https://t.co/WHVJhXMKct