News for hackers. Stuff that matters.
Re:1 TW of solar to be deployed annually by 2030 by vkou on Wednesday May 18, 2022 @07:14PM (#31428783)
Nothing in life is exponential, but there are a lot of people who confuse the first half of a sigmoid curve for it.
Re:1 TW of solar to be deployed annually by 2030 by epistasis on Wednesday May 18, 2022 @07:19PM (#31428820)
Of course, but we are without a doubt experiencing exponential technology change. All models are approximations, and an exponential approximation is far better than whatever is going into EIA and IEA advice to policy makers.

We have not yet hit the linear part of the sigmoid, and are not close. But the good modes out there take that sigmoid shape into account.

(Also, why don't Singularity advocates know about sigmoids?)

Re:1 TW of solar to be deployed annually by 2030 by kelseyfrog on Wednesday May 18, 2022 @08:01PM (#31429107)
Since the early parts of sigmoid and exponential curves are indistinguishable, the singularity advocates opt for the more interesting of the two. The question people are most interested in "what's the saturation level?" of sigmoids has such a large prediction interval that it's next to useless to estimate from data.
Re:1 TW of solar to be deployed annually by 2030 by graycat on Wednesday May 18, 2022 @09:08PM (#31429532)
Right, extrapolation from the present and past where the growth has been exponential is not promising. But if somehow know the saturation level, upper asymptote, then what are doing is more like interpolation than extrapolation. Then the sigmoid curve is the solution to the differential equation I posted here a few minutes ago here at

https://news.ycombinator.com/item?id=31429475

and there is just the one free parameter k. Sometimes in practice have a shot at estimating k.

Re:1 TW of solar to be deployed annually by 2030 by flir on Wednesday May 18, 2022 @08:52PM (#31429419)
> (Also, why don't Singularity advocates know about sigmoids?)

They think the curve goes so far up the y axis before it flattens off that it might as well be exponential to us monkeys stuck at the bottom of the slope.

Re:1 TW of solar to be deployed annually by 2030 by ben_w on Thursday May 19, 2022 @12:35PM (#31436836)
While I think this, I have definitely seen at least one attempt to argue that progress follows a doubling time itself regularly halves and will actually reach infinite “real soon now”.

I also see, as a common trope, people talking about “the knee of the exponential”; as they clearly don’t know how the exponential function works, I have no idea what they think.

Re:1 TW of solar to be deployed annually by 2030 by flir on Saturday May 21, 2022 @01:08PM (#31459675)
Oh well that's just nuts. Can't help those guys.
Re:1 TW of solar to be deployed annually by 2030 by graycat on Wednesday May 18, 2022 @09:02PM (#31429475)
Yup: For time t, upper asymptote b, arbitrary constant k, value at time t y(t),

d/dt y(t) = k y(t) (b - y(t))

Solve that and get a sigmoid, lazy S, curve. Yup, the solution, easy closed form, first calculus, is in terms of exponentials.

Could use that to model, say, the growth of immunity to Covid-19. Once it was used to project the growth of FedEx -- pleased the BoD and saved the company.

Addition: Ah, will save you a little first calculus. In TeX,

$$ y(t) = { y(0) b e^{bkt} \over y(0) \big ( e^{bkt} - 1 \big ) + b} $$

Re:1 TW of solar to be deployed annually by 2030 by esturk on Wednesday May 18, 2022 @10:19PM (#31429964)
How about the expansion of the universe?
Re:1 TW of solar to be deployed annually by 2030 by ben_w on Thursday May 19, 2022 @12:28PM (#31436745)
Definitely not exponential. Based on available evidence, it was much faster in the inflationary epoch, then slowed down, and is now speeding up again.