🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

yup, and imo similar applies for ai

Companies are going to need lots of data to get good at fuzzier, taste-based tasks. One way to do that: get good at being *you*. By training on the data you (and many other users) create by completing a task, they could get better at those tasks too.

Will AGI crown trillionaires and make the rest of us obsolete? 📚🧠 Luke Drago & Rudolf Laine unpack “The Intelligence Curse”. AI that zeros out labor, bends democracy, and rewrites the social contract as we know it. Find out how we fight back. Full episode👇 ------ TIMESTAMPS

fine i guess i might as well deanonymize myself time.com/7289692/when-a…

For a sketch of longer term consequences - and a way out - see here: time.com/7289692/when-a…

I'm finally reading the quite excellent "intelligence curse" paper by Luke Drago & Rudolf Laine They argue that "the key economic effect of AI is that it makes capital a more and more general substitute for labor" From this, all manner of inequalities threaten to follow:

I am becoming increasingly convinced that the key to safe AI is to delegate as much of the goal/value-setting to human operators as possible and make the AI maximally corrigible + good at instruction-following (correctly interpreting the user's intent) and explaining its thinking

More ideas here: intelligence-curse.ai/breaking/!