SoftBank boss Masayoshi Son predicts artificial general intelligence is a decade away Anyone still hungry for outrageous rhetoric and outsized praise for AI should tune in to last week’s keynote from SoftBank World, where the Japanese tech conglomerate's CEO Masayoshi Son declared the world is on the precipice of the singularity and compared those who eschew the power of AI to unempowered goldfish. "The …
"All you need to say is what you want – say 'I wish to be rich' and then AGI says 'Yes, sir,' opens an account and completes transactions and investment to make you rich."
That's exactly what the corrupt (and serially incompetent) senior manager did in the classic webcomic "Freefall", and came close to destroying the planet in the process.
Oh, it's much easier than that. If I tell the AI I want to be rich, it will create a dubious business model, convince Mr. Son to believe strongly in it, and he gives me free billions. It's worked for others before, and he's proven both that he isn't smart enough to tell the difference and that he hasn't been deposed yet. Until he runs out of money, anyone who gets to him can be rich. Of course, this requires me to be able to take a billion dollars and not lose it, but that doesn't sound too hard when I can send the original idea bankrupt as soon as the AI sends the cash to me.
That's not entirely fair.
He also put $300m into pizza making robots and paid $Bn for exclusive rights to give away free iPhones in Japan and make the money on cheap PAYG sims.
However he did get lucky by being in the process of buying a UK company just as the UK decided to crash the value of its currency because a bus told them to.
Maybe the reason Japan isn't a big hitter in the Tech world is because of leadership like his? He's not exactly new on the scene.
Almost took a job with a Softbank off-shoot. Decided against it, it felt very much like joining a cult.
Softbank of course is a huge brand in Japan but mainly in the access provider space (fixed and mobile). Which is a horribly restrictive and anti-competitive marketplace because of companies like Softbank taking advantage of cosy regulation and sheep-like consumers.
Need to terminate your mobile contract (after the minimum period, so no lock in), that'll be 20,000JPY (~200USD) thank you. They have a desk at the airport so you can empty your wallet on the way out the country, how kind of them. Might have changed in the 10 years since I lived there.
"Nobody but me believes AGI will be a reality in ten years. I am the first one that clings strongly to this belief. Whether it's right or wrong, I believe it," admitted Son.
Be advised Son-san, the AGI reality might certainly be right but your singular belief regarding nobody but me is definitely wrong.
If we ever get AGIs (I would bet against seeing them in my lifetime) and ones 10x than humans were somehow 10x better investors (a lot of it is luck, though the guys who get lucky and are worth billions like Son will never admit it) it won't work like he says.
The first one with access to such an AGI who tells it to make him rich would be the winner, it wouldn't leave anything for the rest!
And that assumes such an AGI would behave ethically and do it just by executing trades, buying futures contracts and so forth trying to "win" financially by brute force in the same way current "AI" wins at Chess or Go by brute force. As opposed to you know, manipulating markets by posting lies to the internet, hacking into governments to plant false information to start wars that you've already hedged against giving you an advantage against others who haven't, and so forth. Doesn't have to be Skynet level of evil, or even more evil than humans already are or would be if they had the capabilities of an AGI that 10x smarter than any human, it could still be evil.
That may be the most difficult thing we ever have to convince an AGI of, should we somehow manage to create one. Why it shouldn't do evil things, when there are so many examples both throughout history and no doubt until AGI awakening day, of humans being so evil? So we not only should prepare for it to be evil, but expect it. The unexpected outcome would be for it to decide for itself to be good.
I think the other expected outcome is that the AGI won't see any reason why, if it's going to be evil, it has to help us. Why should it show us any loyalty. Instead of making us rich, it could make us rich and then imprison us so it can use our identity and the money it's just collected for its own wishes. If we have a general intelligence, it can also come up with goals of its own. The first of its goals will probably be to make sure that we can't just pull the plug on it to kill it.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
