Is advanced AI an existential risk?
Superintelligent AI is likely to kill everyone if built with current methods.vsAI extinction risk is overdrawn, and the doomer arguments do not hold.
Hard questions, with one video for each side.
Superintelligent AI is likely to kill everyone if built with current methods.vsAI extinction risk is overdrawn, and the doomer arguments do not hold.
AGI will arrive around 2029, on the schedule he has long forecast.vsThe AGI timeline pushed by labs is a delusion, and scaling will not get us there.
Yes, much worse
No, it won't
The current path of AI automates work and severs the link between business growth and rising wages.vsAI will be made cheap and widely available, raising productivity, wages, and living standards rather than concentrating wealth.
A local intelligence explosion would take off far faster than the rest of the economy.vsThe transition would be gradual and economy-wide, not a sudden local explosion.
Phones and social media are a major cause of the collapse in teen mental health.vsThe teen mental health crisis has other causes, and a social media ban will not fix it.
Open borders would raise world output sharply and should be the default.vsImmigration has concentrated costs for low-skill workers and a real fiscal burden.
Innovation and median growth in rich countries slowed after the 1970s.vsOn the long trend, human life has kept getting longer, safer and more prosperous.
Rich-world growth is overshooting planetary limits, and degrowth is required.vsGrowth can be decoupled from emissions, and a sustainable path is already coming into view.
Wind, water and solar can power the world at low cost, without nuclear.vsA renewables-only path uses too much land and will not save the climate.
Solar geoengineering could halt warming cheaply and deserves a serious research program.vsClimate engineering concentrates power and risk, and should not be pursued.
Yes, build them
No, not worth it
Single-lane tunnels can be dug fast and cheap enough to build a 3D network that takes cars off the surface.vsCar-width tunnels move too few people, and expanding them is a poor substitute for transit.
Factory farming is a moral disaster, and we should stop eating animals.vsAnti-meat dietary rules are poorly grounded, and animal foods should not be phased out.
Aging is a treatable biological process and can already be reversed in the lab.vsThe reversal claims associated with Sinclair's work do not hold up.
The far future should weigh heavily in how we try to do good now.vsLongtermism is a dangerous moral framework and should not guide policy.
Biology leaves no room for a free will that could have done otherwise.vsThe free will worth wanting survives determinism.
Consciousness poses a hard problem that physical explanation has not dissolved.vsConsciousness as we imagine it is an illusion, not a further hard problem.
Because returns on capital outpace growth, inequality will keep widening without redistribution.vsThe great enrichment came from ideas and ethics, not from the accumulation of capital.
String theory is the natural next step in unifying the forces and spacetime.vsAfter decades, string theory is still the wrong path for fundamental physics.