Hacker News Reader

Comment by joe_the_user | Hacker News Reader

I have to mention an article by Allen which I think is the single best critique of the idea fo the "singularity" which I have seen.

Paul Allen: The Singularity Isn't Near

https://www.technologyreview.com/s/425733/paul-allen-the-sin...

View on HN

There is Kurzweil's reponse to this piece linked at the bottom of that article. Since 7 years have passed, there is some opportunity to reflect on the two positions.

Kurzweil makes this comment:

IBM is now working with Nuance (a company I originally founded as Kurzweil Computer Products) to have Watson read tens of thousands of medical articles to create a medical diagnostician.

https://www.technologyreview.com/s/425818/kurzweil-responds-...

Given the history of Watson since these articles were written, I think it lends more weight to Allen's side of the argument.

Good article. I'll also add, progress on AI may also be severely restricted by energy usage. We can build supercomputers which by some measures match the complexity of biological systems, but they consume huge amounts of power compared to the 20W of the brain.

> There’s even exponential growth in the rate of exponential growth.

When you know somebody didn't take calculus or didn't "get" it.

Either that, or he didn't explain what he meant.

(Not Paul Allen, just another interviewee.)

Interviewee? That's a quote from Google Director of Engineering Ray Kurzweil - https://en.wikipedia.org/wiki/Ray_Kurzweil

d/dt e^t = e^t. So while vapid and pop-sci, technically true. Of course, the exponent is constant or we'd be talking about the Ackermanity instead.

Technically true, but the quote makes it sound as if this were not necessarily the case for some kinds of exponential growth, which is clearly false.

Kurzweil really is the hand-waver of all hand-wavers. Exponential progress is the exception rather than the rule, due to hard physical limits.

Reply on news.ycombinator.com