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

Self-driving as a case study for AGI

karpathy.github.io

Post reply on HN
Print view View original

151–160 of 199 posts

Re: Self-driving as a case study for AGI

#151

Post by mechagodzilla »

gitfan86 wrote: post #112
steveBK123 wrote: post #101

Seeing the guy who couldn't deliver FSD decide to make analogies from FSD to AGI does actually give me confidence we are decades away. Yes, yes, I know there's like 1-2 companies that have highly modified vehicles that are pretty good, in a limited geofenced area, in good weather, at low speed, driving conservatively, local roads, most of the time. This is not "FSD". They've been making very impressive incremental im…

AGI has become a philosophical term in the way you are using it. Which is fine to discuss philosophy, but to the point of the article, AI enabled automation is beginning to have significant impact on the economy due to the new functionality.

I think for us to say it's "beginning to have a significant impact on the economy", there needs to be an actual measurable increase in productivity due to AI-enabled automation, which so far has been elusive. It might be impacting employment in some specific jobs, but the mix of jobs in the economy is always constantly changing.

Re: Self-driving as a case study for AGI

#153

Post by gitfan86 »

mechagodzilla wrote: post #151
gitfan86 wrote: post #112

Earlier quoted context omitted.

AGI has become a philosophical term in the way you are using it. Which is fine to discuss philosophy, but to the point of the article, AI enabled automation is beginning to have significant impact on the economy due to the new functionality.

I think for us to say it's "beginning to have a significant impact on the economy", there needs to be an actual measurable increase in productivity due to AI-enabled automation, which so far has been elusive. It might be impacting employment in some specific jobs, but the mix of jobs in the economy is always constantly changing.

Just because something is difficult to measure does not mean it doesn't exist. People who have automated their jobs are unlikely to report that, since they like having a paycheck. But that doesn't mean it isn't happening.

Re: Self-driving as a case study for AGI

#155

Post by minwcnt5 »

mlazos wrote: post #95
motoxpro wrote: post #70

Earlier quoted context omitted.

How is he over selling it when he says explicitly that waymo exists in two cities? It's solved in that it works and they have to scale horizonatally.

lol why do you think they picked those cities? No snow or ice or much inclement weather other than rain. It is not just scaling horizontally.

I mean a year or two ago plenty of people said the tech was a dead end because it didn't handle rain and fog. They will add snow and ice capability as they expand.

Re: Self-driving as a case study for AGI

#156

Post by minwcnt5 »

1oooqooq wrote: post #122
sgt101 wrote: post #91

Earlier quoted context omitted.

I think you may be underselling :) I agree, they have remote drivers waiting to take over, but I think that the current SOA is delivering a very high % of autonomy. Is this viable? Dunno... Will it creep up? Dunno...

you have no idea how much they drive. they avoid talking about it on every interview. Ars explicitly called then out on this and they called saying this is their main profit sucking issue right now "having one overseer per car". to me that means they are, for all practical purposes, driving. no matter how much they try to spin it.

No, they have talked about it in interviews. They've repeatedly said they don't have remote drivers. They have "remote assistance" that the car can call to ask questions if it's confused. You can tell when this is happening in the car, because it plays a message like "our team is working to get you moving". It's rare, maybe once every few rides. There is no way there's one overseer per car.

Re: Self-driving as a case study for AGI

#157

Post by minwcnt5 »

ActorNightly wrote: post #42
CuriouslyC wrote: post #10

I don't things like "full self driving" are meaningful (and probably also AGI), because in reality it isn't a binary thing, rather it's a spectrum of power based on error rate and problem space coverage. Waymo self driving works within a defined subset of the problem space, we can stick a goalpost in the sand in term of the known problem space and error rates and say that represents "full self driving" but the realit…

The issue with FSD systems as they are implemented today is that they aren't AI as much as just complex control algorithms. You can only go so far with mapping sequences of world snapshots to control actions. I do think that once we start to investigate ML/AI structure in the direction of figuring out the correct solution rather than trying to just find functions for control algorithms based on input->output mappings…

This is not true. The most advanced systems in the industry today like Waymo, Cruise, Baidu, Pony.AI, and Tesla, are all primarily AI.

Re: Self-driving as a case study for AGI

#158

Post by HarHarVeryFunny »

staticman2 wrote: post #115
a-dub wrote: post #26

"AGI: An autonomous system that surpasses human capabilities in the majority of economically valuable work." -- what an obscenely depressing reduction of a fascinating field of inquiry. who the hell snuck in and redefined the science of thinking machines to this sad and reductive get rich quick crap?

I think that definition is useful because it is measurable. It sidesteps the endless "It's just a text prediction engine/ I dunno ChatGPT seems pretty smart to me!" discussions. It also sidesteps the "It did well on a test designed to measure human intelligence it must be smarter than humans"/ "no, the test of human intelligence wasn't designed to measure machine intelligence and tells us very little" discussion. It…

> It reduces it to "Can I fire 50% of my workforce? Then it must be AGI."

Well, no, that's job automation, and if it's job-specific then it's narrow AI at best (assuming this is a job requiring intelligence being automated, not just a weaving loom being invented), in other words specifically not AGI.

It's really pretty absurd that we've now got companies like OpenAI, Meta, Google (DeepMind) stating that their goal is to build AGI without actually defining what they mean. I guess it let's them declare success whenever they like .. Seems like OpenAI ("GPT-5 and AGI will be here soon!") is gearing up to declare GPT-5, or at least GPT-N as AGI, which is pretty sad.

Re: Self-driving as a case study for AGI

#159

Post by 29athrowaway »

iknowstuff wrote: post #86
29athrowaway wrote: post #33

As a driver you not only see cars. You apply theory of mind to each car, assigning a personality to each moving car, even if you can't see the driver. Let's see you see a car full of bumps, marks and broken lights. You might think: this car has crashed before, I am going to avoid it. Or a car with racing parts and decals and tinted windows, you know that car likes to accelerate faster than usual and may be unsafe to…

Hardly an absolute necessity. But also, if FSD 12 is truly end to end, nothing is preventing it from learning that as well.

Is it optional? if you see a car zig zagging, or a car is going in the wrong direction, or there's a driving school sign, etc. As a human you can tell something is going on.

Re: Self-driving as a case study for AGI

#160

Post by electrondood »

distortionfield wrote: post #147
randomdata wrote: post #136

Earlier quoted context omitted.

Software developers on HN are saying the same thing. I don't see much evidence that AI has changed the game such that it would cause the work to dry up, though. In fact, I am surprised it hasn't created more work exploring the possibilities of new AI APIs. But when you can get 5% returns just by sticking your money in the bank, why would you bother investing in software, where 99¢ is considered an exorbitant cost by…

I think software devs on here are conflating AI with a hidden recession and other economic factors causing a slump in tech. I still can’t find a single thing that AI could reasonably replace about my job even if I was a junior developer. As a senior engineer using LLM tooling, it offers some benefits, but it’s still nowhere near “job stealing” capabilities.

There's a new fallacy, where AI is dismissed if it "can't replace" a human completely. AI is more like an augmentation tool, that allows fewer people to create work faster by effectively outsourcing the grunt work.

Better tools like Copilot and ChatGPT reduce the amount of time it takes to deliver a feature, reducing the need to horizontally scale developers.

Why hire 10 when 3 could do the same work with AI tooling?

I think while tech is obviously contracting post-COVID from overhiring, it's also true that you just don't need as many people anymore.

Post reply on HN