Conversation

Exposing the professional scaremongers.
Quote Tweet
The AI Dilemma's 😱 Panic-as-a-Business: - Freaking people out with monstrous AI - Freaking people out with dubious survey stats - Distracting people from the real issues 🚨 New article: bit.ly/41JhjJm
Show this thread
Image
45
303
AI: the trillion dollars industry strikes back… All we want to see is a safe path to AI, you have given us none! Be responsible!
4
9
Actually, I have indicated a path towards controlable & steerable AI (below). But the real reason you haven't been given a path to safe AI, by me or by anyone, is that you haven't been given a path to AI at all.
Not worried about you or others developers Yann, you don't call the shots It's your bosses and others alike that doesn't share your ethical values, tho' share yr narrowness and narcissism about the subject but with tons of money, power, greed and ambition for more that it's scary
2
2
Thankfully, your assumption of moral bankruptcy by my "bosses" are gravely mistaken. If they were as ethically challenged as you say, I wouldn't be working with them.
2
4
Show replies
You reference a paper which depicts a path towards autonomous AI, one which says nothing about safety in the title or the abstract, nor does it have section dedicated to it. You might have written a few words about safety in it (and I’ll read) but this is not a path to safety
1
Show replies
Reality isn’t continuous it’s discrete (planck level, branes, quantum), biological brains do create a discrete version of the world The paper isn’t mentioning “prior programming” aka DNA (biological brains do have prior programming including a basic grammar knowledge) /1
1
1
Imagine AI technology as the alien schematics from Contact. Now imagine the scene when they impose a chair to the device and discovered it fundamentally conflicts. This is what's happening via teaching AI that free-will could be real.
Yann, someone needs to unpack WHY steerability of autonomous AI in your sense relates to "safety" in doomer sense. This far, you just talk another language and spook them even more! They're not ML researchers, they've been taught to fear any speak of capability and autonomy!
1