Ever thought that Lee Sedol might be the last human to beat Alphago? Good for him, I like the guy…
I have a feeling that the win was kind of given to honour him… I guess no one will ever know…
I must admit that I also couldn’t help having such a slightly suspicious feeling … but then again, would they really do that? I mean, if LS would also suspect so, it might hurt even more, or?
Would a human act in such a way as to avoid totally embarrassing a great opponent? Depending on the people involved, not entirely unlikely. Would a machine do it? Without human intervention, not likely until a machine possesses empathy and understands the context. Even then, that probably would be because humans put it there
you’re right, unless we can get a programmer who somehow deconstructed alpha go’s state - not really. but i got the feeling from that game that alphago had scored itself behind, and was playing a ‘desperate’ game instead of solidly and consistently taking the largest move - because no move led to a win.
and from seeing Lee play that one game, i was convinced that he had found a way to reliably throw it off the track. i was wrong
Is it really a good way to honour a great master – by fixing a game to his favour? I might be naive but I like to think the people involved had a bit more integrity than that.
i don’t know about koreans but in japan, if you respect your opponent, you give your best shot to take him down.
@trohde @temifar Exactly my thoughs but i still couldn’t shake the feeling that the game might have been rigged… But you guys are probably right… In order to honour someone, doing such an unhonorable thing is just wrong…
This might be true considering master 60:0 record. It would be almost impossible for a human to beat alphago from now on.
We should also remember that Alphago has shown us that our current understanding on how Go is properly played isn’t always right (and with “right” I mean “effective”). With suprising moves and deviation from standardized josekis it’s clear that we have something to learn from Alphago, and since we have something to learn, we will get better too.
I’m not saying that we’ll ever be better than Alphago again, just that it might not be unbeatable. And it can’t be, since otherwise it wouldn’t be able to win over itself.
Maybe the AI let the humans have a glimps of hope so that they dont start the inevitable war to soon.
Possible scenarios of a human beating alphago again:
- An even better AI is available to upload straight to our brains (hardware might be insufficient though )
- The human that does it is really a machine but doesnt know and noone can tell the difference anymore.
I would like to see AlphaGo play more matches with long time controls before humans completely give up the contest. However, in the future, perhaps several top pros will have their brains networked to become HPUs (human processing units), enabling them to meet AG on a level playing field.
i have been waiting for the guy named KeJie, he brags a lot , let’s see how this guy deal with Alphago, or don’t even beat AG once. It will be a shameful.