Live data from Hacker News

I Am Tired of AI

ontestautomation.com

761–770 of 1001 posts

Re: I Am Tired of AI

#761
I'm actually excited about AI. With a dose of realism. But I benefit from LLMs on a daily basis now. There are a lot of challenges with LLMs but they are useful tools and we haven't really seen much yet. It's only been two years since chat gpt was released. And mostly we're still consuming this stuff via chat UIs, which strikes me as sub optimal and is something I hope will change soon.

The increases in context size are helping a lot. The step improvement in reasoning abilities and quality of answers is amazing to watch. I'm currently using chat gpt o1 preview a lot for programming stuff. It's not perfect but I can use a lot of what it generates and this is saving me a lot of time lately. It still gets stuff wrong and there's a lot of stuff it doesn't know.

I also am mildly addicted to perplexity.ai. Just a wonderful tool and I seem to be getting in the habit of asking it about anything that pops into my mind. Sometimes it's even work related.

I get that people are annoyed with all the hyperbolic stuff in the media on this topic. But at the same time, the trends here are pretty amazing. I'm running the 3B parameter llama 3.2 model on a freaking laptop now. A nice two year old M1 with only 16GB. It's not going to replace bigger models for me. But I can see a few use cases for running it locally.

My view is very simple. I'm a software developer. I grew up a few decades ago before there was any internet. I had no clue what a computer even was until I was in high school. Things like Knight Rider, Star Trek, Buck Rogers, Star Wars etc. all featured forms of AIs that are now more or less becoming science fact. C3PO is pretty dumb compared to chat gpt actually. You could build something better and more useful these days. That would mostly an art and crafts project at this point. No special skills required. Just use an LLM to generate the code you need. Nice project for some high school kids.

Which brings me to my main point. We're the old generation. Part of being old is getting replaced by younger people. Young people are growing up with this stuff. They'll use it to their advantage and they are not going to be held back by old fashioned notions about the way the things should work according to us old people. The thing with Luddites is that they exist in any generation. And then they grow tired, old, and then they die off. I have no ambition to become irrelevant like that.

I'm planning to keep up with young people as long as I can. I'll have to give that up at some point but not just yet. And right now that includes being clued in as much as I can about LLMs and all the developer plumbing I need to use them. This stuff is shockingly easy. Just ask your favorite LLM to help you get started.

Re: I Am Tired of AI

#762

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…

> So AI isn't really a tool for creating substance. It's a tool for automating impression management.

I was speaking to a design lecturer the other evening. His fascinating insight was that:

1. The best designers get so much fulfilment out of practicing the craft of design.

2. With the advent of low cost “impression making”, the role has changed to one of “review a lot of mediocre outputs and pick the least crap one”

3. This is robbing people of the pleasure and reward associated with craftsmanship.

I have noted this is applicable so many other crafts, and it’s really sad!

Edited afterthought… Is craftsmanship being replaced by “clickandreviewmanship”?

Re: I Am Tired of AI

#763
post #687

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

They have been improving a lot, but that improvement is already plateauing and all the fundamental problems have not disappeared. AI needs another architectural breakthrough to keep up the pace of advancement.

>but that improvement is already plateauing

Based on what ? The gap between the release of GPT-3 and 4 is still much bigger than the time that has elapsed since 4 was already released so really, Based on what ?

Re: I Am Tired of AI

#764

Earlier quoted context omitted.

China's economy would simply crash if they ever went to war with the US. They know this. Everyone knows this, except maybe you? China has nothing to gain by going to "hot" war with the US.

I think it more likely that China will sabotage our electrical grid and data centers.

For what purpose, so that we can't buy more stuff from them? Do they really hate our business that much? China really has nothing to gain from crippling the US.

Re: I Am Tired of AI

#765

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. It's not so much that I think people have used AI, but that I know they have with a high degree of certainty, and this certainty is converging to 100%, simply because there is no way it will not. If you write regularly and you're not using AI, you simply…

+1, and to put more simply, AI as we know it today makes zero guarantees about its accuracy. That's pretty insane for a "tool" to make no guarantees about being correct in any way for any purpose.

A spellchecker makes guarantees about accuracy. So does a calculator. Broad, sweeping guarantees.

Imagine if we built a tool that could automatically do all the electrical work in a new home from the breaker box to every outlet, and it could do it in 2 hours. However, what if that tool made no guarantees about its ability to meet electrical code? Would people use it anyway? Of course they would. Many dangerous errors would slip through inspection and many more house fires would ensue as a result.

Re: I Am Tired of AI

#766

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. It's not so much that I think people have used AI, but that I know they have with a high degree of certainty, and this certainty is converging to 100%, simply because there is no way it will not. If you write regularly and you're not using AI, you simply…

> It's not so much that I think people have used AI, but that I know they have with a high degree of certainty, and this certainty is converging to 100%, simply because there is no way it will not. If you write regularly and you're not using AI, you simply cannot keep up with the competition. I am writing regularly and I will never use AI. In fact I am working on a 400+ pages book right now and it does not contain a…

Also currently working on a book (shameless plug: buy my book!) and feel no pull or need to involve AI. This book is my mine. My faults. My shortcomings. My overuse of commas. My wonky phrasing. It has to have those things, because I am those things (for better or worse!).

Re: I Am Tired of AI

#767

Earlier quoted context omitted.

>If you write regularly and you're not using AI, you simply cannot keep up with the competition. You're out. A very HN-centric view of the world. From my perch in journalism and publishing, elite writers absolutely loathe AI and almost uniformly agree it sucks. So to my mind the most 'competitive' spheres in writing do not use AI at all.

i regularly (at least once a week) spot a typo or grammatical issue in a major news story. I see it in the NYTimes on occasion. I see it in local news ALL THE TIME. I swear an LLM would write better than have the idiots that are cranking out articles. I agree with you that having elite writing skills will be useful for a long time. But the bar for proof reading seems to be quite low on average in the industry. I thin…

Heh, when I see a spelling error in a news article.. I oddly feel like I can trust it more because it came from a human being. It's like a nugget of gold.

Re: I Am Tired of AI

#768
Current generative AI is a set of moderately useful/interesting technology that has been artificially blown up into something way bigger.

If you've been paying any attention for the past two decades, you'll have noticed that capitalism has had a series of hype cycles. Post COVID, Western economies are on their knees, productivity is faltering and the numbers aren't checking out anymore. Gen AI is the latest hype cycle, and it has been excellent for generating hype with clueless VCs and extracting money from them, and generating artificial economic activity. I truly think we are in deep shit when this bubble pops, it seems to be the only thing propping up our economies and staving off a wider bear market.

I've heard some say that this is all just the beginning and AGI is 2 years away because... Moore's law and that somehow applies to LLM benchmarks. Putting aside that this completely nonsensical idea, LLM performance is quite clearly not on any kind of exponential curve by now.

Re: I Am Tired of AI

#769

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…

"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 how to accurately confirm their knowledge and they’ll let me know later if it turns out they were wrong or the information changed and

None of that is true with LLMs. I never know if I can trust the output, unless I’m already an expert on the subject. Which kind of defeats the purpose. Which isn’t to say they’re never helpful, but in my experience they waste my time more often than they save it, and at an environmental/energy cost I don’t personally find acceptable.

Post reply on HN