Live data from Hacker News

I Am Tired of AI

ontestautomation.com

781–790 of 1001 posts

Re: I Am Tired of AI

#781

Earlier quoted context omitted.

LLMs have been improving exponentially for a few years. let's at least wait until exponential improvements slow down to make a judgement about their potential

In some domains (math and code), progress is still very fast. In others it has slowed or arguably stopped. We see little progress in "soft" skills like creative writing. EQBench is a benchmark that tests LLM ability to write stories, narratives, and poems. The winning models are mostly tiny Gemma finetunes with single-digit parameter counts. Huge foundation models with hundreds of billions of parameters (Claude 3 Opu…

>But once a model is large enough to generate coherent sentences, persuasiveness kinda...stops. All large models are about equally persuasive. No runaway scaling laws are evident here.

Isn't that kind of obvious? Even human speakers and writers have problems changing people's minds, let alone reliably.

Re: I Am Tired of AI

#782
post #530
post #339

Earlier quoted context omitted.

> If an elderly but distinguished scientist says that something is possible, he is almost certainly right - Arthur C. Clarke Geoffrey Hinton is a 76 year old Turing Award* winner. What more do you want? *Corrected by kranner

When he said this was he imagining an "elderly but distinguished scientist" who is riding an insanely inflated bubble of hype and a bajillion dollars of VC backing that incentivize him to make these claims?

What are you talking about? How would Hinton be incentivized by money?

Re: I Am Tired of AI

#783

Earlier quoted context omitted.

Google didn't change it, it embodied it. The problem isn't AI, it's the pervasive culture of PR and advertising which appeared in the 50s and eventually consumed its host. Western industrial culture was based on substance - getting real shit done. There was always a lot of scammery around it, but the bedrock goal was to make physical things happen - build things, invent things, deliver things, innovate. PR and ad cul…

Just yesterday I watched a video blaming China for "oversupplying" electric cars at low prices and how that was hurting car manufacturers elsewhere e.g Germany. These manufacturers were trying to secretly lobby for tariffs to be placed on Chinese cars, while publicly denying this because THEIR largest and growing market was China. They had some "expert" talk about how China is known to oversupply and is doing this to…

Maybe we should stop subsidies to buy electric cars and let China subsidize on the production side instead, while lowering taxes on the production of our own cars.

Re: I Am Tired of AI

#784

Earlier quoted context omitted.

>But superhuman AI seems now only few years away Seems unreasonable. You are afraid because marketing gurus like Altman made you believe that a frog that can make bigger leap than before will be able to fly.

Flying is a good analogy. Superman couldn't fly, but at some point when you can jump so far there isn't much of a difference

There is an enormous difference. Flying allows you to stop, change direction, make corrections, and target with a large degree of accuracy. Jumping leaves you at the mercy of your initial calculations. If you jumped in a way that you’ll land inside a volcano, all you can do in your last moments is watch and wait for your demise.

Re: I Am Tired of AI

#785

Earlier quoted context omitted.

>The most depressing thing for me is the feeling that I simply cannot trust anything that has been written in the past 2 years or so and up until the day that I die Do you think AI has changed that in any way? I remember the sea of excrement overtaking genuine human written content on the Internet around mid 2010s. It is around that time when Google stopped pretending they are a search company and focused on their pr…

Google didn't change it, it embodied it. The problem isn't AI, it's the pervasive culture of PR and advertising which appeared in the 50s and eventually consumed its host. Western industrial culture was based on substance - getting real shit done. There was always a lot of scammery around it, but the bedrock goal was to make physical things happen - build things, invent things, deliver things, innovate. PR and ad cul…

Thanks for writing your comment, I think it’s a public service.

Cultural commentary that makes complex long term trends simple to understand isn’t often this clear or concise. What really makes it powerful though is that it manages to stay in a relatively detached observer-mode without becoming an angry rant. And so rather than provoking (understandable!) anger in others, hopefully it’s inviting more reflection.

People who haven’t thought about it this way might take a harder look at what they are doing and who they really want to be. People that are already thinking along these lines will probably benefit from a reminder that they aren’t crazy.

Re: I Am Tired of AI

#786
post #676

I'm tired of LLMs. Enough billions of dollars have been spent on LLMs that a reasonably good picture of what they can and can't do has emerged. They're really good at some things, terrible at others, and prone to doing something totally wrong some fraction of the time. That last limits their usefulness. They can't safely be in charge of anything important. If someone doesn't soon figure out how to get a confidence me…

I'll keep buying (and paying premium) for dumber things. Cars are a prime example, I want it to be dumb as fuck, offline, letting me decide what to do. At least next 2 decades, and thats achievable. After that I couldnt care less, I'll probably be a bad driver at that point anyway so switch may make sense. I want dumb beautiful mechanival wristwatch. I am not ocd-riddled insecure man trying to subconsiously immitate…

What dumb cars are you looking at?

Re: I Am Tired of AI

#787
post #751

Earlier quoted context omitted.

Thank you, you've stated this all very clearly. I've been thinking about this in terms of "doing work", where you care about the results, and "performing work", where you care about how you are evaluated. I know someone who works in a lab, and pointed out that some of the equipment being used was out of spec and under-serviced to the point that it was essentially a random number generator. Caring about this is "doing…

"Doing work" vs. "performing work": the epitome of this is consulting. Companies pay huge sums of money to consultants that often spend most of their time "performing work", doing beautiful slides even if the content and reasoning is superficial or even dubious, creating reports that are just marketing bullshit, framing the current mission in a way that makes it possible to capture additional projects and bill the cl…

For better or for worse, you’ve described the exact output some of those client companies want, so they can show off the shiny slides and have the status symbol of expensive consultants.

Re: I Am Tired of AI

#788

I'm not tired, I'm afraid. First, I'm afraid of technological unemployment. In the past, automation meant that workers could move into non-automated jobs, if they were skilled enough. But superhuman AI seems now only few years away. It will be our last invention, it will mean total automation. There will be hardly any, if any, jobs left only a human can do. Many countries will likely move away from a job-based market…

I agree with most of your fears. There is one silver lining, I think, about superintelligence: we always thought of intelligent machines as cold calculators, maybe based on some type of logic symbolic AI. What we got instead are language machines that are made of the totality of human experience. These artificial intelligences know the world through our eyes. They are trained to understand our thinking and our feelin…

> made of the totality of human experience

They are made of a fraction of human reports. Specifically what humans wrote and has been made available on the web. The human experience is much larger than text available through a computer.

Re: I Am Tired of AI

#789

Earlier quoted context omitted.

genocides and murder are very human ...

this is so annoying. i think if you took a random person and gave them the option to commit a genocide, here a machine gun, a large trench and a body of women, children, etc... they would literally be incapable of doing it. even the foot soldiers who carry out genocides can only do it once they "dehumanize" their victims. genocide is very UN-human because its an idea that exists in offices and places separated from t…

> you are so wrong and ignorant that it makes my eyes bleed when i read this comment

This jab was uncalled for. The rest of your argument, agree or disagree, didn’t need that and was only weakened by that sentence. Remember to “Be kind. Don't be snarky. Converse curiously; don't cross-examine. Edit out swipes.”

https://news.ycombinator.com/newsguidelines.html

Re: I Am Tired of AI

#790
post #769

Earlier quoted context omitted.

"They're really good at some things, terrible at others, and prone to doing something totally wrong some fraction of the time." I agree 100% with this sentiment, but, it also is a decent description of individual humans. This is what processes and control systems/controls are for. These are evolving at a slower pace than the LLMs themselves at the moment so we're looking to the LLM to be its own control. I don't thin…

> I agree 100% with this sentiment, but, it also is a decent description of individual humans. But you can understand individual humans and learn which are trustworthy for what. If I want a specific piece of information, I have people in my life that I know I can consult to get an answer that will most likely be correct and that person will be able to give me an accurate assessment of their certainty and they know ho…

It defeats the purpose of LLM as personal expert on arbitrary topics. But the ability to do even a mediocre job with easy unstructured-data tasks at scale is incredibly valuable. Businesses like my employer pay hundreds of professionals to run business process outsourcing sites where thousands of contractors repeatedly answer questions like "does this support contact contain a complaint about X issue?" And there are months-long lead teams to develop training about new types of questions, or to hire and allocate headcount for new workloads. We frequently conclude it's not worth it.
Post reply on HN