Benvenuto nel Poliverso

Jeff Atwood

1 settimana fa • •

Jeff Atwood
1 settimana fa • •

The LLM / GAI people are hitting exponential difficulty walls with massively diminishing returns. I don’t care how many GPUs and “training data” you throw at the problem, you can’t brute force your way out of this… but you can certainly waste billions trying. My self driving car bet with Carmack is the canary in the coal mine. When, and only when, we have fully autonomous SAE level 5 self driving cars, we thus by definition have true, full general purpose artificial Intelligence: blog.codinghorror.com/the-2030…

The 2030 Self-Driving Car Bet

It’s my honor to announce that John Carmack and I have initiated a friendly bet of $10,000* to the 501(c)(3) charity of the winner’s choice: By January 1st, 2030, completely autonomous self-driving cars meeting SAE J3016 level 5 will be commercially…

^{Jeff Atwood (Coding Horror)}

Oblomov reshared this.

in reply to Jeff Atwood

p

in reply to Jeff Atwood • 1 settimana fa • •

AI? thermodynamics says no
@codinghorror

@Jeff Atwood

in reply to Jeff Atwood

Don Marti

in reply to Jeff Atwood • 1 settimana fa • •

Do you win if the self-driving cars are available in all 10 cities, but refuse some pickup points and/or destinations within those cities where human cab drivers are allowed and willing to go? (It seems like a lot of the work would in covering the weirdest streets with the most different kinds of users)

in reply to Don Marti

Jeff Atwood

in reply to Don Marti • 1 settimana fa • •

@dmarti not sure; read the SAE definitions closely.

@Don Marti

in reply to Jeff Atwood

Richard Degenne

in reply to Jeff Atwood • 1 settimana fa • •

Don't get me wrong, I'm a huge fan of Carmack (do you know if he is on Mastodon by any chance?), but SAE-5 passenger cars in 5 years sounds absolutely delulu.

in reply to Jeff Atwood

Softwarewolf

in reply to Jeff Atwood • 1 settimana fa • •

Hadn't read this blog post before, but I still agree with your take on AR vs. VR in there. Too bad all the millions that have been dumped into VR, cryptocurrency, and LLMs aren't instead being used on something useful like battery technology.

in reply to Softwarewolf

Jeff Atwood

in reply to Softwarewolf • 1 settimana fa • •

@faoluin exactly. I hate being right about all this, I just more people would listen and focus on what actually matters.

@Softwarewolf

in reply to Jeff Atwood

Shafik Yaghmour

in reply to Jeff Atwood • 1 settimana fa • •

I am more interested in the second order effects of things like this huge push to build data centers. Which is incidentally raising power prices.

So when these AI start ups start folding and the capacity is just sitting there. Do we end up w/ like a decade of oversupply? Where get data center capacity is so cheap you have to give it away but power is so expensive they become millstones around someones neck.

You wonder if they will end up undermining what is now a profitable sector for many. It is not like we have not done this before.

in reply to Shafik Yaghmour

Jeff Atwood

in reply to Shafik Yaghmour • 1 settimana fa • •

@shafik I think having a lot of compute is a net benefit for a lot of other related problems that are useful. There will be a fair amount of overbuilding, which means prices will go down, and that's also kind of a good thing.

@Shafik Yaghmour

in reply to Jeff Atwood

Matt Palmer

in reply to Jeff Atwood • 1 settimana fa • •

I don't think even full self-driving is evidence of AGI. Self-driving is a massively complicated problem, but driving doesn't require coming up with any really *new* ideas, which is an important component, in my opinion, of anything that wants to claim the ability to compete with humans cognitively.

⇧