This market resolves yes if, before 2053, no human-level AI (as in non-brain-emulation AI) has been created yet, a human brain is scanned and functionally emulated in another medium, such as silicon, and the emulation appears to successfully behave like a human brain.
1,000
1.00I've previously published on WBE:
• https://sciendo.com/downloadpdf/journals/jagi/4/3/article-p130.pdf
• https://informatica.si/index.php/informatica/article/view/1874
so I believe my opinion on this topic carries some weight. I used to be a lot more optimistic about pre-AGI WBE, but I've become a lot more pessimistic on it over time.
This piece by Steve Byrnes does a nice job articulating many of the reasons why WBE in particular may be hard:
https://www.alignmentforum.org/posts/PTkd8nazvH9HQpwP8/building-brain-inspired-agi-is-infinitely-easier-than
Basically, since the brain was evolved instead of intentionally designed, the functional parts of brains (e.g., neurons) likely contain many important "side channels" for their proper I/O behavior, and we'd have a really hard time pinning all these down. In order for WBE to come first, we'd also need to validate all this behavior before anyone else was able to use the knowledge necessary to validate these parts to build some other type of brain-inspired AGI (i.e., we don't have the higher-level understanding of how the brain works to build an AGI based on this general understanding, yet we somehow validate all the parts of WBE to make sure it works; ditto for we can't build a more alien-like AGI using neuronal-circuit-like architectures, but we can do WBE).
My all-things-considered probability on getting WBE before any other type of AGI is around 3%. Most of my probability mass here is on longer timelines, though. The odds I'd put on this question – getting WBE first and getting it by 2053? Probably ~1%.
@HarlanStewart how does it resolve if we get WBE first and then we also get other forms of AGI after (but before 2053)?
@DanielEth actually yes I suppose with how I was thinking of this, and I think with the way I've worded it, it would resolve yes in that case.
As in, if we emulated a human brain tomorrow, I would resolve yes immediately rather than waiting until 2053.
Let me know if you think that this was too confusing or would have changed your original bet and I will reimburse any loss from selling shares
@HarlanStewart I think this makes sense. Okay, here's my model now:
• ~1% chance we get WBE first and get it by 2053
• ~60% (?) chance we get AGI by 2053, all things considered
• so the chance I'd put on this market resolving yes is ~1%/40% = ~2.5%