If AI experts agreed, you wouldn't need a newsletter to sort it out. These five questions divide the field more than any others, and your view on each shapes how worried you end up.
1. How soon will AI match humans at most tasks?
In the 2023 survey of 2,778 AI researchers, the median estimate for a 50% chance of machines beating humans at every task was 2047. That was 13 years earlier than the same survey found just one year before. Some lab leaders expect it within a few years; many academics think it's decades away or further.
Why it matters: nearly every other risk gets more urgent if this arrives sooner.
2. Could advanced AI slip out of human control?
Turing Award winners Geoffrey Hinton and Yoshua Bengio have warned this is a serious possibility. Fellow Turing winner Yann LeCun has called those fears overblown. The same survey found a median 5% chance of extinction-level outcomes, which some read as alarming and others as a guess with little basis.
Why it matters: this is the difference between AI as a powerful tool and AI as an actor in its own right.
3. Are openly released models safer or more dangerous?
One side says releasing model weights openly lets anyone, including bad actors, remove safeguards. The other says openness lets more people find flaws and keeps power from concentrating in a few companies.
Why it matters: it decides whether the most capable AI is controlled by a few labs or available to everyone.
4. How much does AI really help someone build a weapon?
The International AI Safety Report 2026 found leading models now match experts on some biology knowledge tests. The open question is whether knowledge is the bottleneck, or whether the physical lab work is still the hard part.
Why it matters: if knowledge is the bottleneck, the risk is already here. If not, there's more time.
5. Will AI take jobs faster than new ones appear?
Optimists point to past technologies that created more jobs than they destroyed. Pessimists argue AI is different because it can learn new tasks as fast as humans can retrain.
Why it matters: this is the risk most likely to touch the most people first.
Our take
When experts disagree this much, the honest position is uncertainty, not certainty in either direction. We'll track the evidence on each of these and tell you when it shifts.
Get every issue in your inbox. It's free, and you can unsubscribe anytime.