The Terminator Is Not Coming. The Future Will Thank Us | Hacker News Reader

The Terminator Is Not Coming. The Future Will Thank Us

recode.net

17 pointsfchollet11 comments

There is a massive fallacy in this paragraph:

"This doomsday scenario can’t happen. Intelligence is a product of learning. A brain, no matter how big or how fast, does not become intelligent until it learns. In humans, learning is a slow process requiring practice, repetition and study over many years. Unlike humans, intelligent machines can be cloned, transferring the learning from one machine to a new one, eliminating the learning time. However, for an intelligent machine to discover new truths, to learn new skills, to extend knowledge beyond what has been achieved before, it will need to go through the same difficult and slow discovery process humans do. Yes, an intelligent machine could think much faster than a biological brain, but the slower process of discovery still has to occur."

The person writing this fails to understand the difference between engineering and biological evolution. The AI would be engineered to begin with. There would be nothing haphazard or extraneous about its design. The "slow process of discovery" would be bootstrapped, as stated, by "cloning" (copying extremely fast memory compared with humans). Then, the AI would attack new problems with very likely a minimum 100-1 speed advantage over humans, based on electronic versus chemical speeds. Nerve impulses travel at well less than 100 miles per hour, versus the over 150,000 miles per second of electronic impulses.

Consider a first-gen AI engineer, with "only" a 150 IQ human equivalent intelligence, attempting to design a better AI. Such an AI would have no distractions (food, sleep or sex) and would likely have a monomaniacal focus on its job. It would have a perfect, near-instantaneous memory of arbitrary size (we can already do this). I expect it would in fairly short order produce an IQ 300 version of itself, and that that cycle would likely continue.

That is the basis for the idea of the "singularity", and I find it completely believable. What it would mean, I don't know...and neither does anyone else. We have the same relationship to an IQ 1000 AI that a chimp has with us.

That is the concern that Musk, Hawking and others share.

You're confusing "brain" and "intelligence". A brain will be engineered to begin with. Then is has to learn. Once it learns some amount of information, it is intelligent, and this intelligence can be cloned. However, to increase its intelligence, that brain has to start learning again. And this process can be slow, because it is limited by the speed of experimentation.

I agree that the machine will probably be able to learn faster than humans, but it's not at all clear how much faster. Also, we know little about human intelligence. For example, Einstein's brain wasn't much bigger or faster than an average human brain, yet he could make profound discoveries. It's not clear to me that if you take an average brain, and make it faster/bigger, it will suddenly be able to make scientific breakthroughs.

Terminators did not have to be self aware or have desires to do their job, killing. They simply had to follow a set of rules. Intelligence will merely help them follow the rules when circumstances go beyond the originally coded set of conditions.

The real threat of intelligent machines is the people who would use them. Not everyone with ability acts rationally and why would we expect them to use such ability correctly?

"In the 1940s, no one envisioned the smartphone, the Internet or GPS satellites."

Except Arthur C Clarke, and the creator of Dick Tracey? So, no one but comic strip creators and short story writers?

Or are smart phones substantially different from portable communicators? I guess so. Nobody saw the 'fart app' from 1940.

I'm not as worried about completely autonomous AI/robots, as I am about government-controlled "autonomous" robots, or even non-autonomous ones.

For instance, for someone in the Middle East, or any poor country being in war with US, it would very much look like the Terminator war, if the US only sends remote-controlled robots and drones to invade that country.

I would prefer they use drones and robots that have some kind of targeting, even though the targeting is imperfect, than they use cluster bombs on civilian populations.

Especially when they make the unfortunate mistake of making the cluster bomblets the same colour as air-dropped food packages.

That would be true if it were not for the fact that targeting comfortably from your desk makes it much more likely that you do not care about killing people. It is just a game, after all. And you are the good one!

Side thought: what is the purpose of designing a humanoid robot? To make it tolerable to humans.

What is the purpose of designing a terrifying humanoid robot (even if it is covered by fake nice-looking flesh)? None whatsoever!

Whatever shape terminator takes, it won't be the one in the movie ...

A robot becomes much more capable if you can't tell it's a robot!

I put exactly as much stock in predictions like this as I do in those that say that Terminator-alikes are coming.

Zero.

It might be comforting, and it might even be true, but this kind of cheerful optimism is simply in bad taste in my view.

It's either hubristic or it's empty in content, in that we know the promise the future might hold. Technology has been promising us the world for a while now, but what we've gotten has actually been a shift towards dystopia. Yeah, the future might be a paradise where you can virtually travel to see relatives in Guam while also learning to cross-stitch, but Silicon Valley thinking currently presents us with ad-infested crapware.

We don't actually need more cheerleading.

The kind of people that write this kind of article aren't the kind of people that are actually working against the threats these writers want to ignore.

Reply on news.ycombinator.com