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

301–310 of 319 posts

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

#301

Post by bluefirebrand »

stego-tech wrote: post #193
bluefirebrand wrote: post #140

Earlier quoted context omitted.

> Work(ed) for a company super big on a force of agents. > RIFed this year Yeah that's why they are big on the force of agents, so they can justify trimming their workforce Anyone who doesn't see this is kind of a sucker

Oh, 100%, and I knew that going in. Still, I'm the type of worker (sucker? sap?) that will go down with the ship if it helps the rest of the folks get into lifeboats. I know I can take care of myself when stuff hits the fan, and I'm never one to half-ass my output because of my own cynicism. I'm hoping that pays off for me someday. Thus far, it's just been a lot of burnout/layoff cycles.

> Still, I'm the type of worker (sucker? sap?) that will go down with the ship if it helps the rest of the folks get into lifeboats

> engineers to approvals only for ~20 tickets a week averaging ~40hrs of work, freeing up said engineers for actually valuable work instead of handholding customers through routine tasks

Not trying to be too judgy here but to me it sounds more like you are the one helping sink the ship than help everyone to life boats

Something to think about. Agreeing to bring AI into our workflows is actually digging out the foundation beneath our own feet (and our coworkers feet)

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

#302

Post by hyperturtle »

teleforce wrote: post #300

I'm really surprised out of current nearly 300 comments and counting, nobody really is mentioning constraint programming or CP, a seemingly forgotten deterministic sibling of the stochastic data-driven AI [1],[2],[3]. "Out of these 14 tasks, we guessed that only 6 require abstract reasoning alone to perform. Strikingly, we classified only one of the top five most important tasks for medical scientists as relying sole…

I've always wondered if it would work out well for an LLM to evaluate a constraint program/script when it encountered a problem involving constraints / logic. like or-tools with python or evaluating a minizinc program.

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

#303

Post by techpineapple »

paulddraper wrote: post #234
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.

This is what the Amish/Luddites would say.

I don't think Amish/Luddites would make quite the nuanced point I'm trying to make, but also, I don't have any particular beef with the Amish or the Luddites, in fact the older I get maybe the more I side with them, but again, that's aside from the particular point I'm making now.

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

#304

Post by teleforce »

hyperturtle wrote: post #302
teleforce wrote: post #300

I'm really surprised out of current nearly 300 comments and counting, nobody really is mentioning constraint programming or CP, a seemingly forgotten deterministic sibling of the stochastic data-driven AI [1],[2],[3]. "Out of these 14 tasks, we guessed that only 6 require abstract reasoning alone to perform. Strikingly, we classified only one of the top five most important tasks for medical scientists as relying sole…

I've always wondered if it would work out well for an LLM to evaluate a constraint program/script when it encountered a problem involving constraints / logic. like or-tools with python or evaluating a minizinc program.

Yes, that's very plausible and I've thought similarly but not exactly the same approach for the diagramming limitations of LLM [1].

Fun facts, the presenter John Hooker was asked about how to determine the suitability of a particular heuristic solver for specific problems in his presentation. He casually answered that if he knows the solutions to that he will probably win a Nobel Prize. But perhaps AI/LLM can help in a way to recommend the solver based on the type of applications or problems.

If I'm not mistaken there's also Donald Knuth (TAOCP) asking questions after the JH's presentation, how often you see that?

[1] Diagrams AI can, and cannot, generate:

https://news.ycombinator.com/item?id=43424916

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

#305

Post by fakedang »

robertlagrant wrote: post #296
asdf6969 wrote: post #238

Earlier quoted context omitted.

The most likely future is that we start standardizing infrastructure and building in a different way that’s easy for robots to use. Kind of like how cars would suck if we didn’t have roads and Benjamin Franklin couldn’t have predicted power plants and wires to every home. Then people who can’t afford the new stuff will keep paying plumbers but there’s a slow path to get rid of them. The old stuff will always be aroun…

> Think like how the Catholic Church never disappeared but the rest of the world moved on around it What does this mean? The world didn't start Catholic and gradually reduce its Catholicism over time. There are apparently about 1.4bn Catholics out there[0]. [0] https://en.wikipedia.org/wiki/Catholic_Church

The Catholic church once decided which kings and princes could marry which queens and princesses, which alliances would be allowed and which wouldn't. Now it's nothing but a shadow of its former self.

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

#306

Post by KoolKat23 »

Absolutely agree with this article, also there is an astronomical amount of low hanging fruit out there. Areas where the cost/reward calculation have not made sense in the past.

One only has to think of the ubiquity of excel VBA still, and that'd probably still be regarded as fancy for most.

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

#307

Post by rqtwteye »

milesrout wrote: post #286
rqtwteye wrote: post #255

Earlier quoted context omitted.

True. I have a serious DIY porn addiction due to Youtube.

But do you actually do any DIY? And is it more than you would have done without watching hours of it on YouTube? I have at times watched a lot of videos of people making stuff or doing DIY. I suspect like 99% of the audience of those videos, they are just entertainment.

Yes, I do a lot. Everything I feel confident doing and have the time for.

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

#308

Post by babyent »

teleforce wrote: post #300

I'm really surprised out of current nearly 300 comments and counting, nobody really is mentioning constraint programming or CP, a seemingly forgotten deterministic sibling of the stochastic data-driven AI [1],[2],[3]. "Out of these 14 tasks, we guessed that only 6 require abstract reasoning alone to perform. Strikingly, we classified only one of the top five most important tasks for medical scientists as relying sole…

I’m using constraint programming in my platform.

Any experts want to help me?

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

#309

Post by stego-tech »

bluefirebrand wrote: post #301
stego-tech wrote: post #193

Earlier quoted context omitted.

Oh, 100%, and I knew that going in. Still, I'm the type of worker (sucker? sap?) that will go down with the ship if it helps the rest of the folks get into lifeboats. I know I can take care of myself when stuff hits the fan, and I'm never one to half-ass my output because of my own cynicism. I'm hoping that pays off for me someday. Thus far, it's just been a lot of burnout/layoff cycles.

> Still, I'm the type of worker (sucker? sap?) that will go down with the ship if it helps the rest of the folks get into lifeboats > engineers to approvals only for ~20 tickets a week averaging ~40hrs of work, freeing up said engineers for actually valuable work instead of handholding customers through routine tasks Not trying to be too judgy here but to me it sounds more like you are the one helping sink the ship t…

The situation was way more complex than that - we were on borrowed time anyway, and this was our attempt to pivot away from a single on-prem private cloud group into the team overseeing both the enterprise pipeline and multi-cloud work. It would’ve freed us from busywork babysitting other teams’ own stuff while we were gradually chipped away via a war of attrition so we could actually build useful stuff to justify our long term existence and make us indispensable, a strategy that had buy-in at the time.

Ah well. At least seeing me get RIFed hopefully sent up signal flares with the rest of the team to GTFO while they can, and I had a lot of good accomplishments towards the end in particular. Take the good where I can find it, I suppose.

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

#310

Post by yellowapple »

> To be clear, we agree that AIs will eventually outperform humans at nearly all economically valuable activities, and this will lead to a significant acceleration in economic growth.

The (literal) trillion-dollar question, though, is who benefits from this significant acceleration in economic growth? Every major technological revolution, from the Agricultural through the Industrial and into the Digital Revolution, has produced further and further stratification of society. Will this one be the exception?

I can see four ways this can go:

1. We repeat our mistakes, with power concentrating into ever-fewer hands while everyone else enjoys less and less autonomy (at best, in exchange for the bare minimum token quality-of-life improvements to prevent revolts)

2. We learn from our mistakes and achieve the fully-automated luxury gay space communist utopia that techno-optimism has promised us

3. We riot and conduct an anti-automation purge, Butlerian Jihad style

4. We discover that AI wasn't all it's hyped up to be and any AI-driven technological revolution is still a long ways off

I'm hoping for #2. I'm expecting #1 or #4. Society is trending toward #1 or #3.

Post reply on HN