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

Most AI value will come from broad automation, not from R & D

epoch.ai

Post reply on HN
Print view View original

51–60 of 319 posts

Re: Most AI value will come from broad automation, not from R & D

#51

Post by toogan »

jgilias wrote: post #30
jajko wrote: post #13

Earlier quoted context omitted.

Nobody sane and smart I know is advocating for AI now . All keep saying 'one day', 'soon' etc. Our megacorp very effectively blocked all of this and we don't have anything internal yet, but we are in specific market. And basically all I saw so far was a snake oil from people trying to push it or benefit from it, even here. Trust me bro... well no, show me some hard facts and we can start building trust. Sure assistan…

Advocate or not, it’s here to stay. And if you’re not adjusting to it, you’re doing yourself a disservice. Last week I did the amount of work that would’ve taken me give or take a month. A significant part of it was writing an API client for a system I needed to use. Pretty run of the mill stuff. Doing this ‘by hand’ just takes time. Go look at the docs, type out your data structures, wire things up for the new call,…

> Last week I did the amount of work that would’ve taken me give or take a month. A significant part of it was writing an API client for a system I needed to use. Pretty run of the mill stuff. Doing this ‘by hand’ just takes time. Go look at the docs, type out your data structures, wire things up for the new call, write tests. With “the robot” once the framework was largely in place, you just paste API docs for the endpoint you need, and it’s done in a minute. With tests and everything.

That just points to an inefficiency. Could be tackled in other ways than involving an LLM to produce essentially what's being done elsewhere every day over and over again. A framework automating and hiding all this would be just as effective. Perhaps even cleaner than all that duplication that the LLM created for you.

In other words, that month of busywork that you just saved is inherently unnecessary to do. But progress is not linear in the number of lines of code that you produce. If you think hard about a good architecture and design, coming up with that after 2 weeks of hard thinking, that could be 90% of the work. The remaining 10% are writing all that down. That could take 3 more months, and it taking more than 10% of the time points to the existing inefficiency in tooling / framework / ... But making that more efficient isn't necessarily achieved the best by using an LLM and writing it all out. There are still huge redundancies, which is what made the LLM possible. Once you boil these down in some common frameworks or tools, the LLM will also just produce the same few lines that you'd need to produce to get those 10% done in then just 10% of the total time.

Re: Most AI value will come from broad automation, not from R & D

#52

Post by pjmlp »

surgical_fire wrote: post #49
pjmlp wrote: post #44

Earlier quoted context omitted.

Except this time is quite real for some jobs, maybe programming will take a few more years, but for artistic fields, it is already here. Indie devs now using AI generated assets instead of hiring artists is quite common.

> Except this time is quite real Oh it was very real during crypto hype cycle too. The crypto people were very annoying with it. > Indie devs now using AI generated assets instead of hiring artists is quite common. I would imagine it is fine for some concept art, based on my experimentation with multiple models for Stable Diffusion. For generating actual assets that are consistent? I am highly skeptical. Unless we ar…

Consistent enough for it to be an issue.

And not only indies, when taking voice actors into consideration.

https://www.ign.com/articles/horizon-actor-ashly-burch-says-...

Or some well known studios,

https://www.ign.com/articles/activision-finally-admits-it-us...

Re: Most AI value will come from broad automation, not from R & D

#53

Post by godelski »

This is a bit silly. It is equivalent to saying "most food will come from tractors, not land." There is some truth to this, but you really can't have one without the other. The whole system is interconnected.

The problem with all these jabs at R&D and science is not recognizing that R&D and science produce the very foundations that lead to the ability to do production in the first place. I'm certainly not saying we should dump all our money into R&D, but I do find it weird that others talk as if the ground you stand on doesn't matter. You literally cannot stand without it. The truth is that you need both. I suspect why we shy away from R&D is for 2 big reasons.

First, it has a lot of failure. The hardest thing about doing research is being able to stick with it when 90% of what you do doesn't work. It's fucking hard. But of course it is. If it was easy it would have already been done. So returns on the work can take a lot of time and the visibility of the failure is emotionally draining to those without enough resilience.

Second, we do not accurately capture nor ascribe the value nor of research. People who create ground breaking scientific revolutions that create the capacity for trillion dollar markets never end up with 1% of the result of their work. You don't see Tim Berners-Lee being a billionaire, nor Linus Torvald. You didn't see it in Einstein and cases like Turing are quite common through history. Certainly this is an alignment issue as we as a society should be encouraging such pursuits as their benefits vastly outweigh anything that has been done by Apple or Google. Not to diminish their achievements, they have both done fantastic and incredible things, but do they not sit on the foundations created by Turing, Burners-Lee, Hopper, and others. Or look even today, the work done by LeCun, Sutskever, Karpathy, Fei-Fei Li, Hinton, and yes, even Schmidhuber has created more than a few dozen multi-billion dollar enterprises. Yet as far as I am aware, not a single one of them is a billionaire. I do not believe (I could be TOTALLY wrong) their combined net worth is a billion. Even if it was, that would be a far cry from what many of the mogul we see today. Do we really think Zuckerberg has created more value than these people? Certainly this is entirely dependent upon your definition of value and I think most of use could agree that it would be incredibly naive to believe this exclusively means the money in their pockets. If you really do believe that, I will say that you're part of a problem. Money is a proxy, sometimes we need to stop and think "a proxy for what?"

Re: Most AI value will come from broad automation, not from R & D

#54

Post by BriggyDwiggs42 »

surgical_fire wrote: post #49
pjmlp wrote: post #44

Earlier quoted context omitted.

Except this time is quite real for some jobs, maybe programming will take a few more years, but for artistic fields, it is already here. Indie devs now using AI generated assets instead of hiring artists is quite common.

> Except this time is quite real Oh it was very real during crypto hype cycle too. The crypto people were very annoying with it. > Indie devs now using AI generated assets instead of hiring artists is quite common. I would imagine it is fine for some concept art, based on my experimentation with multiple models for Stable Diffusion. For generating actual assets that are consistent? I am highly skeptical. Unless we ar…

I was very involved in the crypto scene, and really the only real utility was monero for buying the funny chemicals.

Re: Most AI value will come from broad automation, not from R & D

#55

Post by godelski »

mirekrusin wrote: post #48

So they are arguing which one will get more benefits first - r&d or general automation? What is the point in arguing around it? Disconnect between progress being done (ie. alphafold) and trying to infer answer from some historic stats on r&d investment, ratios, their past estimated impact etc. is... just weird. It's also funny that the whole ai itself together with constant breakthroughs is r&d.

I hate these people that try to argue we don't need R&D. It's extremely short sighted and equivalent to pretending that the ground you stand on is worthless. Without the ground, you literally have nothing to stand on... Talk about pulling yourself up by your bootstraps...

Re: Most AI value will come from broad automation, not from R & D

#56

Post by toogan »

beau_g wrote: post #46
yarekt wrote: post #40

Earlier quoted context omitted.

Regulation is violence? sorry for putting my socialist hat on, but free market is efficient, so efficient that without regulation it’ll optimise away human happiness and find a way to turn tears into profits. Regulation is basically saying that you can make money in ways that benefits the humanity also, at least in theory. Lobbying and corrupt regulators muddy the waters

Of course, laws/regulations are enforced by the party in a country that has a monopoly on violence, and use the threat of violence to enforce (either imprisonment or a monetary fine, monetary fines are a derivative of kidnapping as money takes time to accumulate). Of course I'm not arguing against these functions in general, they should be used in ways that prevent an even worse act of violence (ex. a corporation was…

Threat of violence is not violence.

A policeman standing on a public square threatening to incarcerate anybody who is violent results in no violence actually happening at that square. Take away that regulation (in form of the policeman) and watch the actual violence start.

Re: Most AI value will come from broad automation, not from R & D

#57

Post by BriggyDwiggs42 »

gom_jabbar wrote: post #24
AndrewKemendo wrote: post #9

This is exactly correct technically and perfectly on the existing “paved path” of technology determinism. My company is actively in the process of demonstrating the “Learning machine” which is a Weiner style cybernetic system of systems It’s unquestionable at this point that machines will displace all human labor where labor efficiency is the key factor for investment/use. Start with transfer learning from existing h…

> Almost nobody is working on what to do after One interesting possibility is that this is not something humans need to work on at all - because, as Nick Land famously put it, "Tomorrow can take care of itself." [0] [0] https://retrochronic.com/#meltdown

Why’s he write like if I tried to mimick anti-oedipus

Re: Most AI value will come from broad automation, not from R & D

#58

Post by toogan »

pjmlp wrote: post #41
techpineapple wrote: post #2

Does anyone else find techno-optimism really depressing? I guess one for the more common reason that it’s displacing humanity with technology, but the second reason is just you can’t get excited about a hype that is unlikely to manifest. It feels very ungrounded from tangible benefits to society.

I am as optimistic as a factory worker about to replaced by robots. Hence why I can't stand such techno-optmism, apparently most folks live in an ideal world free of economics.

Has the last 70 years of productivity increases led to a reduction in weekly work hours? No.

Some jobs will be automated away. Good thing. Braindead stuff that a machine can do should be done by a machine. Doesn't mean we'll all soon be just picking our noses. There will be other work to be done, and if unregulated capitalism has its say then it can easily lead to even more worker exploitation.

Re: Most AI value will come from broad automation, not from R & D

#60

Post by onemoresoop »

elzbardico wrote: post #14
Swizec wrote: post #12

Earlier quoted context omitted.

> Broad automation of tasks can be great for society if and only if the product of that automation is not treated as some kind of private property. Star Trek or Bladerunner? That’s a choice we can make.

No. Most of the time WE can't make this choice. It is imposed on us by the elites.

We still have a lot of power to decide what we want, the problem is that this power is not well leveraged yet.
Post reply on HN