FOS asked of https://youtu.be/z0-IPdNiLkg 40 min – can this be accurate???
Not really, kind of.
The guy makes some clear errors, and he has some things close enough to accurate.
It is already the case that AI is smarter than humans at any game that can be defined.
But, AI is far more energy intensive than humans for now, but that will change with time.
Take a look at this –
This is pure deep reinforcement learning. In a few days, this learned from scratch to beat any human.
This ability is general – not specific to the game of go.
This system can beat any human at any game that has defined parameters – like stock market, futures trading, whatever.
There is no future for humans in zero sum games.
We have to change to positive sum games.
The conclusion of the article above states:
Our results comprehensively demonstrate that a pure reinforcement
learning approach is fully feasible, even in the most challenging of
domains: it is possible to train to superhuman level, without human
examples or guidance, given no knowledge of the domain beyond basic
rules. Furthermore, a pure reinforcement learning approach requires
just a few more hours to train, and achieves much better asymptotic
performance, compared to training on human expert data. Using this
approach, AlphaGo Zero defeated the strongest previous versions of
AlphaGo, which were trained from human data using handcrafted fea-
tures, by a large margin.
Humankind has accumulated Go knowledge from millions of games
played over thousands of years, collectively distilled into patterns, prov-
erbs and books. In the space of a few days, starting tabula rasa, AlphaGo
Zero was able to rediscover much of this Go knowledge, as well as novel
strategies that provide new insights into the oldest of games.
We are indeed in “interesting times”.
Just watched it (Werner Herzog’s movie – Lo and Behold: Reveries of the Connected World).
Kind of gets some of the flavour of what is happening, but misses most of what I consider the interesting stuff.
Security is always a chain, and any chain is only as strong as the weakest link – that is something most security people just don’t get. Human memory is almost always the weakest link.
Rather than security and centralised systems, I prefer multiple independent trust networks.
Rather than markets, efficiency, rules and scarcity based thinking, I prefer fully automated abundance, cooperation, massive distributed redundancy and individual responsibility.
So Herzog paints a tiny sketch of one corner of a grand vista.
It really is vastly more complex.
Hi OM & Deb,
Chappie was entirely CGI – http://www.gizmodo.co.uk/2015/07/how-the-most-realistic-robot-in-cinema-history-was-made/
The idea of an AI starting from a child, and growing, except much faster than a human, is actually how modern AI systems like AlphaGoZero work.
Agree that the movie is unrealistic in many aspects, yet it does hint at otherwise quite realistic aspects.
A world where anyone tolerates a threat to the life or liberty of anyone (human or AGI) that has AGI in it, is not a stable solution.
As I see it, our only chance of survival is to go beyond scarcity and markets to fully automated abundance, respect, diversity and individual freedom. Anything else is unlikely to be pretty, and almost certain to have humans only as zoo animal status.
Deb – I don’t see the internet as evil, I see the lack of responsibility, the lack of universal cooperation in society generally, as unstable and dangerous. Being able to be entirely anonymous releases many of the very necessary constraints of behaviour required for security and liberty generally.
In a really odd twist, the greater the liberty we claim, the greater the responsibility and necessary sets of constraints that come with that. It may seem counter-intuitive to many, but it is really basic systems thinking. Every level of system has necessary boundary conditions. Complex society like ours requires morality – cannot survive without it. We must, each and every one of us, limit the worst excesses of our natures (and we all have them – no exceptions).
We need boundaries, but not too many – only those that really are necessary. The major ones are universal respect for individual life and individual liberty, any individual capable of conceiving of itself, human and non-human, biological and non-biological.