cobaltblue

Fan Hui is 2p, so very skilled, but the ranking system goes up to 9p. To give a sense of how large a gap that is, there is and has only ever been one Westerner to achieve that rank, Michael Redmond. The article states they plan to face off against Lee Sedol 9p, and if they beat him in a no-handicap game, that will be as impressive as Deep Blue against Kasparov.

You'll want to watch this year's Computer Go UEC Cup[0] in March, with Zen and CrazyStone being the typical victors. (CrazyStone in particular has sustained a 6d rating on KGS, a popular Go server, but lately has been at 5d. In the past it has beaten a 9p with a 4 stone handicap, which is impressive, but that handicap is huge and it hasn't yet won with 3 stones.) Of interest this year is that Facebook is competing, and AFAIK they seem to take a similar approach as Google by training the AI using deep learning techniques and then strengthening it further with MCTS. In their public disclosures they claim to beat Pachi pretty often, which puts their bot around 4d-6d, it'll be interesting to see how it fairs against Zen and CrazyStone in the Cup and if it wins against a 9p.

[0] http://jsb.cs.uec.ac.jp/~igo/eng/

[1] https://news.ycombinator.com/item?id=3819346

Work time should generally only be the time between clock-in and clock-out. For professionals that are freed from clock-in-clock-out, I think they would still clump around that 40 hour figure if you made them clock-in-clock-out for the data rather than for the pay.

For instance the other comment brings up the uncertainty around giving an Oracle AI the task of finding what benevolence means to humans, to then plug into the final AI hoping it will be Friendly. You don't know how to precisely define 'benevolent' so let the Oracle AI do it (or help you do it), though you think you can at least program the Oracle AI to do a good job in finding answers to vague, complex, fragile problems. (Why? Past success in AI tech?) Is what the Oracle AI outputs actually good? How do you know? What did the Oracle AI have to do to reach its answer, or gain a last bit of certainty in it? (e.g. Did it reason it needed more data and so built or hastened the arrival of brain-scanning technology, started scanning brains from volunteers, the very recently deceased, cryonics patients, or just without its creators knowledge through black market deals, and at some point did it simulate trillions of human minds under all sorts of gruesome scenarios to test responses? Do you even morally care about sentience on silicon? Would you care more if you were an upload?) You've got a lot of problems just building an Oracle AI (or anything powerful enough that can help you solve the hardest problems). Even supposing the output it gives is good and the costs (practical and ethical) were acceptable, will a seed AI allowed to recursively self-improve be guaranteed to preserve this Friendliness property in all subsequent improvements?

With the idea of using "AI tech", you've basically drawn a line in the outcome space that on one side says something like "let's just program the thing" and the other says "hold on, let's just program the thing but also use all these other relevant computer tools we've developed over the decades to help program the thing". It's the start of an approach, but does it really prune all that much? What do these relevant tools actually buy you in safety, if a safety-guarantee-tool is not among them? Another line which may be what you were heading towards, could be on one side say something like "we might not be able to solve all the necessary problems with our current intelligence, so let's also spend time looking for ways to augment human intelligence -- brain-computer interfaces, brain emulations, tweaks to our genetic code, that sort of thing" and the other side "no, we're capable, let's try our best right now". Using ems based on friendly-ish human minds that become better than human could very well be a better approach to solving the superintelligent problems correctly and efficiently than doing so on raw human levels, on the other hand it could turn very bad if one of those ems instead just solves the problem of recursively improving themselves and their goal and value system is insufficiently friendly. Or many other failure modes. (Though an interesting "most likely scenario" given certain assumptions and only looking at a span of 2 years which by itself doesn't look too terrible is Robin Hanson's Age of Em idea.) Just like approaches involving molecular nanotech to help us, there's reasons why it could be seen to help the best outcome along, but it also opens the door to a lot of other risks that aren't necessarily there if you took the other side, so again how much is really pruned?

What I take as your general idea of "maybe we need and should employ additional help and knowledge that we've yet to gain to even really start" isn't bad, though, much better than "Benevolence is easy, a King and his subjects all know he is a great ruler when everyone is smiling frequently", it's just lacking in detail to constitute a plan. Throwing in the positive feedback loop idea (which might even manifest as impressive accelerating returns, who knows) to me seems only relevant to figuring out time scales, I don't think it says anything about the fraction of good vs. bad outcomes apart from whether short or long time scales matter.

You give a director a big budget from their past successes, not necessarily because the underlying thing is somehow easier or harder to make money off of due to copyright. That's icing, not cake.

(Em mine.) I have a problem, my rig's not very powerful. Where can I vote to get a better rig? Or does my rig do the voting weighted by its power? (http://lesswrong.com/lw/jb/applause_lights/)

An alien species would have to be observing this planet at very frequent intervals to be able to catch any of these recent developments -- because if millions of years ago they determined nothing interesting was happening, and decided to only look in every five hundred thousand years or so, it's very likely the last time they looked our species wasn't even around.

Now you take the development of an AI that is roughly human level intelligent with a given amount of resources. If you give it the same amount of resources again, it can "breed", copy itself perfectly onto them, and now we have a "population" of two AIs whose combined intelligence should roughly match that of two twin humans. Maybe it doesn't even copy perfectly, maybe it switches some things about to see if it can create something smarter without modifying its own code directly. In any case its breeding is only restricted by its ability to breed and desire to breed in the first place and the compute and power resources it needs to run a new copy, and these can always get lower than initially. What does the long-term look like? Even if the initial resources for the first one are astronomical (e.g. all computing power in use on the planet as of 2016), I would still bet you that left to breed without any other restriction it would take the AI family far less time than 100 thousand years to reach 7 billion instances. So you're looking at many, many more AIs than non-AIs, plus each with human level intelligence that isn't distributed normally (assuming soft eugenics doesn't become widespread in non-AI populations) but is mainly all the same with perhaps increases here and there (assuming a soft takeoff), each coordinating with the same goals in mind. This situation without further specification can be either extremely good for the non-AIs or extremely bad for the non-AIs. The question of whether this will happen (and whether it happens in that exact form, personally I think hard takeoff from a single AI is likely) in the next 100 years or the next hundred thousand years (perhaps a near-extinction-level event occurs forcing our species to basically start over but with an even harder battle to survive since many low-hanging-fruit resources have been depleted) might help you determine whether to worry now about doing actions that make the extremely-good more likely than the extremely-bad, but if one thinks human-level AI can't happen at all in any amount of time, that there's something special about our squishy brains or something inherently limiting about our so-far-general intelligence such that we can't solve the engineering problem of creating another intelligence directly and must always go through breeding, the argument needs to take place at a different level.

The paperclip maximizer argument (not sure if Bostrom repeats it in his book) is yet another level of argument but it's meant for those who already buy the premise of human level AI and self-improving AI but who also think increased intelligence will naturally converge to a human-loving benevolent AI that sees how self-evidently obvious it is to love all life or whatever, that we don't have to worry about all the messy details around Friendliness because they'll just fall out of the entity naturally. No one is actually worried about Clippy tiling the solar system with himself, its probability is epsilon. But a lot of people just see this one argument out of context and infer the arguer believes it to be a relevant possibility to worry about and write off the whole group as crazy...