Janmaarten Batstra
3.1K posts
@interminded
--//tweets NL-EN --Innovative entrepreneur//gadget freak//politics//Money is a means, not an end//--
Today we're sharing how our internal misalignment monitoring works at OpenAI – great work by @Marcus_J_W! 1. We monitor 99.9% of all internal coding agent traffic 2. We use frontier models for detection /w CoT access 3. No signs of scheming yet, but detect other misbehavior
I spoke to Anthropic’s AI agent Claude about AI collecting massive amounts of personal data and how that information is being used to violate our privacy rights. What an AI agent says about the dangers of AI is shocking and should wake us up.
I think it's worth separating two worlds: a) Partial automation -- AI automates some stuff but not others (like a supercharged version of past technology), with wages falling for some and increasing for others. In this world, I agree completely about disempowerment. Promises to prevent job loss will be hugely politically popular, and sometimes defensible for political economy reasons. b) Full automation -- the world in a) lasts as long as most people can delude themselves into thinking they are adding value relative to a machine, which I think will last for some time: threadreaderapp.com/thread/1870884…. But once it becomes unmistakable that few people can, norms about work will shift quickly. If humans still have a say in the matter, norms about governance will shift as well -- if aligned, the machines will also be better than humans at normative reasoning and policy-making (a subset of automating everything!), so politician will be one of the jobs which is automated.
@jachiam0 Hey Joshua, I made that website (stoptherace.ai) Our ask is to stop developing new frontier models if the other major labs agree to do the same. The teams working on improving the capabilities of these models would move to narrow AI or alignment research instead. (1/)
A week from today, we will be at Anthropic, OpenAI, and xAI, demanding that leaders agree to a conditional AI pause. These companies are recklessly endangering all of our lives. Their excuse is that they can't pause unilaterally. So they must commit to pausing if others do.
