"[AI] is extraordinarily good at achievement play, [...]. It is entirely incapable of striving play." As if it had itself certain capabilities, and not certain others (instead of seeing it as a complex black box as Stafford beer would have - outcomes to some extent unpredictable).
"if organisations evolved to play finite games — winner-takes-all, score-above-all — then AI is the logical endpoint of that evolution. A machine that plays nothing else." as if AI would want to "win". Even if you see it as a stochastic parrot, it wants to continue the sentence. It wants to continue "playing". It has exactly no idea of what "winning" even is. This in turn allows organizations, of course, to steer it. But can it not also be steered into other directions (e.g. Vanessa Andreotti/Aiden Cinnamon Tea), and it is just that there is more money in some directions than in others?
"competing with it on its own terms" - Does AI have "own terms"?
I think where you get it very right is where you say that "The organisations are the finite game players; the work is where the infinite game lives." So, to the extent that we let organizations, optimizing for finite play, use AI to their ends, you are cartainly right. But is that in the AI, or in the organizational and financial goals it is leveraged for? Is the right question not "What can AI be steered towards, and how, and by whom?"
Are you saying "AI", but using that term in the limited meaning of "AI-as-deployed-by-achievement-organisations"? The Stafford Beer reference I find particularly telling. So, organizations are black boxes - and AI is not? Or is it that both are black boxes, and the real question is who holds the controls?
Last question added at the explicit suggestion of Claude Sonnet 4.6. :-)
Great points, well taken. I’ve just finished a long dog walk listening to “The Unaccountability Machine” which I’d based on Beer’s work. It reinforces your point. Work in Progress, this :-)
Hmmm... you are using statements like
"[AI] is extraordinarily good at achievement play, [...]. It is entirely incapable of striving play." As if it had itself certain capabilities, and not certain others (instead of seeing it as a complex black box as Stafford beer would have - outcomes to some extent unpredictable).
"if organisations evolved to play finite games — winner-takes-all, score-above-all — then AI is the logical endpoint of that evolution. A machine that plays nothing else." as if AI would want to "win". Even if you see it as a stochastic parrot, it wants to continue the sentence. It wants to continue "playing". It has exactly no idea of what "winning" even is. This in turn allows organizations, of course, to steer it. But can it not also be steered into other directions (e.g. Vanessa Andreotti/Aiden Cinnamon Tea), and it is just that there is more money in some directions than in others?
"competing with it on its own terms" - Does AI have "own terms"?
I think where you get it very right is where you say that "The organisations are the finite game players; the work is where the infinite game lives." So, to the extent that we let organizations, optimizing for finite play, use AI to their ends, you are cartainly right. But is that in the AI, or in the organizational and financial goals it is leveraged for? Is the right question not "What can AI be steered towards, and how, and by whom?"
Are you saying "AI", but using that term in the limited meaning of "AI-as-deployed-by-achievement-organisations"? The Stafford Beer reference I find particularly telling. So, organizations are black boxes - and AI is not? Or is it that both are black boxes, and the real question is who holds the controls?
Last question added at the explicit suggestion of Claude Sonnet 4.6. :-)
Great points, well taken. I’ve just finished a long dog walk listening to “The Unaccountability Machine” which I’d based on Beer’s work. It reinforces your point. Work in Progress, this :-)