Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The AI Misinformation Epidemic

approximatelycorrect.com

Post reply on HN
Print view View original

281–286 of 286 posts

Re: The AI Misinformation Epidemic

#281

Post by knieveltech »

kenjackson wrote: post #242
knieveltech wrote: post #240

Earlier quoted context omitted.

Which is kind of the point. ROI is not a strong predictor of societal improvement.

It may be, but there isn't a real good metric for societal improvement. And it seems like progress would always have ebbs and flows as different technologies have new opportunities that are squeezed -- and then other technologies look plumper, so they are squeezed, and so on. And that's fine. I don't think anyone is complaining today about timepieces not keeping even more accurate time. Clocks today (in whatever devi…

I think there are several quality metrics for gauging societal improvement: - healthcare access and outcome statistics - poverty & homelessness statistics - education levels - access to clean water in sufficient quantity - access to healthful foods in sufficient quantity - levels of environmental contaminants - delta in income of top and bottom economic tiers

And so on. Toss in a few bullet points intended to flag totalitarian/authoritarian/fascist tendencies, worker exploitation, and the like and I think that while likely not a comprehensive blueprint that's more than enough to get started. Especially considering how poorly so many industrialized, first world, "democratic" countries in the world do on so many of these basics.

That said I agree there's no ROI on improving household items like clocks. The problem is there's also no obvious ROI on resolving homelessness.

Re: The AI Misinformation Epidemic

#282

Post by graycat »

ericd wrote: post #278
graycat wrote: post #219

Earlier quoted context omitted.

For charging overnight, my concern is that (A) if drive very much during the day and (B) the house has only a 100 A circuit breaker box, then the charging time will still be too long, that is, that overnight is not long enough. The back of the envelope arithmetic I did on overnight charging at 100 A boiled down to driving a relatively large golf cart not very far or fast. I could be wrong. Yes, a Tesla is a lot bigge…

Well, I think 100A at 120V should theoretically be enough - 12kw. The biggest Tesla battery is 100 kwh, so assuming no charging losses (bad assumption, I know), you should be full in 8ish hours from empty. In reality, you won't want to use your whole rated electrical capacity on one thing, but on the flip side, you can go higher voltage, or upgrade your breaker, assuming your wiring isn't too marginal. Tesla sells a…

You make a point: Of course, my house here in NYS is drawing 240 V from the electric utility. So, IIRC the 100 A of the main breaker is for 100 A at the 240 V, not just the 120 V. So, just have an electrician run a 240 V line from the breaker box to the garage for the car, say, the same way as there is a 240 V line to the kitchen electic stove and the electric clothes dryer.

But, the 100 KWh is 123 HP for an hour or 61 HP for 2 hours or 30 HP for 4 hours. That's not a lota juice.

Apparently gasoline, burned, releases

33.4 KWH

So, the Tesla battery has the energy of burning 3 gallons of gasoline.

If a gasoline powered car is 33% efficient, then the 100 KWH of a Tesla is about like a 10 gallon tank of gasoline.

Hmm -- my old Chevy S-10 Blazer has a 20 gallon tank for gasoline, and I can fill it in a few minutes.

But if only want to zip, the fantastic acceleration, around Manhattan a little during the week and take a 2 hour round trip to the Hamptons on weekends, then it appears that a Tesla will work!

IIRC, the time I used my Firebird to blast overnight from my place in Maryland to my fiance in Indiana, the whole trip was about 9 hours. At the speeds I was driving, I might have been using 100 HP. So, the Tesla battery might have lasted a little over an hour. On that trip, even with gasoline, I had to stop a few times. A Tesla would have been a bummer, make me late to get to my fiance, in particular, truncate my manhood -- OUCH!

Car was a Firebird 400 4 speed with, sadly, a 3.23 rear axle ratio. The first gear was 2.52, and fourth was 1.0. Bummer. Now, wisely, much wider ratios are common and more gears, e.g., in the Corvette automatic 8 gears. So, can try to compete with a Tesla for standing start acceleration but on a highway in one of the higher gears cruise along at maybe 1200 RPM and get maybe 28 MPG. Gasoline powered cars have been improving, too!

It just hit me: Maybe a Tesla is using the electric motors also for braking and, then, also charging the battery! Nice.

Trains and big boats commonly have a Diesel engine driving a generator and the generator driving electric motors -- "Look, Ma, no or a perfect transmission!". Could also have a battery in there! Use the battery around town, and charge the battery with the engine.

Re: The AI Misinformation Epidemic

#283

Post by shahbaby »

eli_gottlieb wrote: post #230
akyu wrote: post #172

Earlier quoted context omitted.

>composable, hierarchical generative models and probabilistic inference. Deep neural networks are basically just huge continuous circuit approximators. These two things are not mutually exclusive.

What sort of neural network implements human-style learning and inference?

numenta's HTM

Re: The AI Misinformation Epidemic

#284

Post by kordless »

jodrellblank wrote: post #52
kordless wrote: post #4

Earlier quoted context omitted.

I'm from the camp the singularity already happened! ;)

I'm from the camp the singularity is a continuous process. When you travel away from the Sun, the light from it gets dimmer and dimmer and dimmer, until it becomes single photons appearing with more and more delays but on average energy of 'dimmer' light. Fly towards the Sun and there isn't a sudden cutoff point where it goes from dark to bright, or bright to incandescent, it's a continual increase. Same with life, l…

> becoming a specific point where everyone burns up and dies

Some of us seek this in the attainment of Nirvana.

Re: The AI Misinformation Epidemic

#285

Post by tim333 »

>see the recent Maureen Dowd piece on Elon Musk, Demis Hassabis, and AI Armaggedon in Vanity Fair for a masterclass in low-quality, opportunistic journalism

I finally read through most of the 8000+ words of it and it's not that bad - mostly a bunch of interview quotes from Musk, Kurzweil et al and some of it's a bit sceptical eg:

>When I mentioned to Andrew Ng that I was going to be talking to Kurzweil, he rolled his eyes. “Whenever I read Kurzweil’s Singularity, my eyes just naturally do that,” he said.

Re: The AI Misinformation Epidemic

#286

Post by AJW451 »

ExactoKnight wrote: post #269
knieveltech wrote: post #240

Earlier quoted context omitted.

Which is kind of the point. ROI is not a strong predictor of societal improvement.

What needs to happen is for VC and other sources of investment to stop chasing easy wins and realize huge money can be made by moving more out into the world of atoms.

How do you know moving into the world of atoms will lead to huge economic wins? Its easy for you to say someone should chase less risky investments when its not your money being invested.

Would you know a few VCs who have lost several million or more on an investment. Do it sometime and then express your optimism for the future.

Post reply on HN