AI is word-association. True AGI is a way off. And even when it arrives, still no more a game changer today than bow over spear, or the atom bomb. Remember how fast such leads dissipate.
Our best move is to do the work, pay for research, not vote-buying by politicians, and remain realistic.
This is a fascinating strategy topic. Alas, as usual, I have more questions than answers.
Will the real breakthru in ai (agi) constitute a different category of technology, one more similar to a person than an iPhone? Will it have a will, an independent intent, an ability to frame its own ethical views, etc that are quite beyond the control or intent of its creator? And how might such true autonomy affect the open/closed strategy discussion in ways that iphone-like tech chattel does not?
Short of that robust version of ai, are there different flavors of bounded AI that might better suit different cultures who might pursue the tech -- such that, for instance, Germany would want an ai with different privacy guardrails than Brasil might? If so, might it not be culturally impossible for Brasil to find value in licensing Germany's product or vice versa?
Open does not necessarily imply symmetrical openness. An interesting scenario for consideration is prisoners dilemma when the other prisoner is known to prefer cheating. The expected payoff from collaborating then falls to zero even for the normally-willing party. It only takes one bad actor to precipitate a race to the bottom. So an interesting strategic question is what can one do to hold the collaboration together when significant players are well-known to intend to cheat on the collaboration? Classic cartel theory comes into play, it seems.
AI is word-association. True AGI is a way off. And even when it arrives, still no more a game changer today than bow over spear, or the atom bomb. Remember how fast such leads dissipate.
Our best move is to do the work, pay for research, not vote-buying by politicians, and remain realistic.
This is a fascinating strategy topic. Alas, as usual, I have more questions than answers.
Will the real breakthru in ai (agi) constitute a different category of technology, one more similar to a person than an iPhone? Will it have a will, an independent intent, an ability to frame its own ethical views, etc that are quite beyond the control or intent of its creator? And how might such true autonomy affect the open/closed strategy discussion in ways that iphone-like tech chattel does not?
Short of that robust version of ai, are there different flavors of bounded AI that might better suit different cultures who might pursue the tech -- such that, for instance, Germany would want an ai with different privacy guardrails than Brasil might? If so, might it not be culturally impossible for Brasil to find value in licensing Germany's product or vice versa?
Open does not necessarily imply symmetrical openness. An interesting scenario for consideration is prisoners dilemma when the other prisoner is known to prefer cheating. The expected payoff from collaborating then falls to zero even for the normally-willing party. It only takes one bad actor to precipitate a race to the bottom. So an interesting strategic question is what can one do to hold the collaboration together when significant players are well-known to intend to cheat on the collaboration? Classic cartel theory comes into play, it seems.