3 ms·
How striking is it that everybody has so much to say about the dangers of AI. What if we look at it this way: there are already real-known dangers coming at us
by nuancebydefault 4y ago
How striking is it that everybody has so much to say about the dangers of AI.
What if we look at it this way: there are already real-known dangers coming at us: climate change and a change of instability of the world order.
What if we only have one shot to solve these known dangers and what if an exponential growing corpus of smartness could teach us how?
We need to grab that chance, the stakes are too high.
Pausing is not an option. If 99 per cent pauses, the 1 percent that does not obey will have a tremendous advantage.
- biscottigelato 4y agoClimate change and instability of world order doesn't end humanity. They are barely considered an extinction event. AI Alignment is consistently voted the top potential cause to extinct the human race. They are not even close. https://riskfrontiers.com/insights/ranking-of-potential-causes-of-human-extinction/ https://riskfrontiers.com/insights/ranking-of-potential-caus...
- nuancebydefault 4y agoAI is unstoppable, it is a race. On the other hand, problems related to climate change are solvable, but only using science and technology, there is no way around that. Putting the heating a few degrees lower or only use windmills and solar panels will only slow down, not mitagate the risks, it is too little too late. We need more sophisticated solutions. Do you see any other way than more intelligence? I mean better models, faster accurate simulations and prototypes. On the world order front: machines might come to know us better than we know ourselves and provide us with a solution. AI could be arriving just-in-time.
- kromem 4y agoEXACTLY! The relative surface area AI represents in the existential threats pile is notable but only a slice of the pie. But there's not much in the pile for mitigation of all those other human-driven threats other than pulling a deus ex machina out of our rear ends. So while we should definitely discuss the inherent risks, we should also be discussing the risk of opportunity costs in delaying greater scalable intelligence being added to the mix. This letter seems a bit like myopic fear mongering by people still more caught up in 70s thought experiments about paperclips than the realities facing us in the present day.