re: Ray Kurzweil. Google's Director of Engineering predicts Singularity by 2029

quote:

---

"The singularity" is when machines are so advanced they can improve their technology at an exponential phase that the interval between changes is not discernible. Ushering in essentially AI, the inability to differentiate reality from non-conscious stimulation.

It goes like this:

You build a computer in 1985. To double the power f that computer take 5 years until a more powerful machine happens. Then you have a more powerful machine 2.5 years later.

Then you apply new computational power to making new machines. Then these machines make new machines to make better machines.

Essentially to the point that the chip or processesor or android or computer you just made immediately channels that power to make a NEW improvement. So incremental tech advances no longer take years, nor months, nor day, but HOURS. Then seconds. Then the limit approaches zero and it is constant exponential growth.

Like using IBM Watson to make a better Watson. Then that makes a better Watson. Then that one then the next then the next. So advanced they can assess and make and produce improvements instantly.

Frightening stuff when you grasp the concept

---

All of this.

The machine then consuming the human, let's not use the friendly word "merging," is a very real possibility either for energy, resource, or efficiency purposes.

quote:

---

And an eventual reality

---

Bingo.

quote:

---

AI research is essentially an un-regulated field of study.

---

Yeah this is currently a problem, so I see what you're saying. But I am seeing more and more people in academia who want to put an end to this. It's only a matter of time before someone does something so "uncanny valley" that the government steps in to regulate.

quote:

---

But seriously 12 years is a laughable timeline. Another poster pointed me to the complexity break, which is basically the idea that not all technological progress is exponential since the more information becomes known about a subject or field the harder it becomes to innovate further within that field, and I think we are approaching that break with computation. Hell I think Moore's law failed for the first time ever a couple years ago and has been progressively slowing down.

Once we can have a conversation with an Alexa type thing (actual conversation, pas Turing test), master quantum computing and start getting really good at mapping the brain, I'll start getting nervous.

---

Yeah, Kurzweil is an extremist on this subject and always has been. There are a lot of other very well respected scientists, engineers, etc who agree with what you just posted. It's definitely a debate, and no one really knows the answer either way currently. Although Kurzweil is far more extreme in his views than most of the famous names surrounding this issue (Musk, Gates, etc.).

I personally share the not nearly as optimistic as Kurzweil view that we don't know if the exponential growth can continue, and if it does and ASI is achieved, it's fricking scary. I think it's likely the greatest risk to humanity and our future as a species.

Waitbutwhy had a great article on AI a few years ago:

LINK

quote:

---

Waitbutwhy had a great article on AI a few years ago:

---

I literally posted this on page 2. Pay attention

quote:

---

Yeah this is currently a problem, so I see what you're saying. But I am seeing more and more people in academia who want to put an end to this. It's only a matter of time before someone does something so "uncanny valley" that the government steps in to regulate.

---

They'll never be regulated. It impedes progress with us as a species. Too much money is on the line with R&D, and someone will create it behind closed doors regardless, unintentionally or not. The Singularity is inevitable.

I mean couldn't we regardless sense its processing power? I don't know how it could hide it, and as previously stated that once it reaches human levels of intelligence, it will reach thousandfold levels of human intelligence within merely weeks.

quote:

---

Too much money is on the line with R&D, and someone will create it behind closed doors regardless, unintentionally or not.

---

This scares me - the fact that it will most likely, from things I have read, happen behind closed doors, and the scariest part of all, is that there is decent chance that when AGI/ASI emerge, we might not even realize it until it's too late. People a hell of a lot smarter than me have put forth the fear that when AGI does emerge, it will be smart enough to conceal how smart it is until it has had the chance to replicate itself in order to protect itself from being deleted, since being deleted will be the #1 danger to AGI/ASI. And not in a "fear of dying" kind of way, but in a more practical "I can't complete my programmed goal, whatever that is, if I am deleted" find of way.

quote:

---

That's already a reality, the stock market is completely run by AI.

---

That's kind of my point - indirect, but advanced, research in making smarter and smarter programs is already taking place in places like Wall Street. And think about that - Wall Street sees advancing the intelligence of machines as a way to make a metric shite-ton of money. That should scare the ever-living frick out of everybody. They have more money to spend in order to advance machine intelligence than almost anybody.

quote:

---

It's a very narrow version of intelligence, not really a threat to replace humans in anyway outside of it's narrow scope.

---

That's true, as far as we know, right now. But again, its the speed of advancement, with no oversight, that is the danger. What is narrow today might not be in a year or two. The idea that narrow AI can't inadvertently progress to AGI is the type of thinking people in the field have warned against. It leads to complacency.