Of course, anybody can be an armchair AGI researcher. However, some people genuinely *are* AGI researchers! As of 2020, there were about 70 ongoing AGI projects in the world [https://gcrinstitute.org/papers/055_agi-2020.pdf]. For example, please have a look at (a) https://cambridgeaisocial.org/assets/Slow_down__A_tutorial_on_and_summary_…
Of course there are AGI researchers. But how many of them are so convinced they've found the One True Way that they'd characterize alternate approaches as wrong?
Probably all of them. By definition, if they believed there was a better way than their own approach, then they would pivot towards the better way!
AGI has, in my estimation, a 30+ year postgraduate learning curve. Most AGI researchers, and certainly most AI researchers, are still exploring their own personal learning journeys. I don't believe there's anything wrong with that. Eventually, given sufficient time, the genuinely stronger AGI designs will emerge as superior to the rest. It will just take a few decades!
I'm not sure I fully understand what you mean ("Unfortunate"...?)
Nevertheless, *true* AGI is a cathedral project. However smart you are (and the guys at OpenAI et al are all very, very smart), however much money you have (even billions), it will take many decades (and multiple lifetimes) to design, refine, and implement. There's simply no way of escaping that - if you want true AGI, you have to do the work!
Thus the "lack of concrete successes" to date is only to be expected. To paraphrase the movie Apollo 13, we're on step 8 of 1000.
Perhaps I could have been more precise. For any AGI designer (or any designer, for that matter), they will (of course) believe that their own design is optimal considering all the circumstances (if they didn't, then they would modify it!) But I suspect that many/most will nevertheless still have sufficient humility to admit that they might be wrong to some (either minor or potentially even major) extent. (In fact it's hard to imagine that a truly rational person would not admit that possibility.) But it stands to reason that each individual designer will pursue / advocate for what they personally believe to be right.
Of course, anybody can be an armchair AGI researcher. However, some people genuinely *are* AGI researchers! As of 2020, there were about 70 ongoing AGI projects in the world [https://gcrinstitute.org/papers/055_agi-2020.pdf]. For example, please have a look at (a) https://cambridgeaisocial.org/assets/Slow_down__A_tutorial_on_and_summary_of_the_current_state_of_AI.pdf, and (b) https://bigmother.ai (full disclosure, I am the author of both!)
Of course there are AGI researchers. But how many of them are so convinced they've found the One True Way that they'd characterize alternate approaches as wrong?
Probably all of them. By definition, if they believed there was a better way than their own approach, then they would pivot towards the better way!
AGI has, in my estimation, a 30+ year postgraduate learning curve. Most AGI researchers, and certainly most AI researchers, are still exploring their own personal learning journeys. I don't believe there's anything wrong with that. Eventually, given sufficient time, the genuinely stronger AGI designs will emerge as superior to the rest. It will just take a few decades!
Given the lack of concrete successes, I would hope for more humility than that, but you're much closer to the "ground truth". Unfortunate
I'm not sure I fully understand what you mean ("Unfortunate"...?)
Nevertheless, *true* AGI is a cathedral project. However smart you are (and the guys at OpenAI et al are all very, very smart), however much money you have (even billions), it will take many decades (and multiple lifetimes) to design, refine, and implement. There's simply no way of escaping that - if you want true AGI, you have to do the work!
Thus the "lack of concrete successes" to date is only to be expected. To paraphrase the movie Apollo 13, we're on step 8 of 1000.
If your comment about the lack of humility in the field is accurate, I find that unfortunate
Ah, OK, I understand.
Perhaps I could have been more precise. For any AGI designer (or any designer, for that matter), they will (of course) believe that their own design is optimal considering all the circumstances (if they didn't, then they would modify it!) But I suspect that many/most will nevertheless still have sufficient humility to admit that they might be wrong to some (either minor or potentially even major) extent. (In fact it's hard to imagine that a truly rational person would not admit that possibility.) But it stands to reason that each individual designer will pursue / advocate for what they personally believe to be right.