Fucking Tired of AI | Hacker News Reader

Fucking Tired of AI

107 pointsakaTrickster64 comments

Title. I really wish we could turn back time to 2022 when ChatGPT and GitHub copilot were being dropped and stopped the bomb then. Everywhere I go online now is populated by a mass infestation of AI trash of variying degree; can we or will we ever get back our fun tech discussion online without bullshit ads or bots or any of this novo-dystopian nonsense? I am specially worried about new programers and internet users being bombarded with tons of AI-written books, pornography, music, video, audio, images. And if you think it isn't bad, try googling any historical figure and you'll find some bullshit AI generated image of them likely before the actual one.

What do we do? Do we just create new social media platforms and blogs (IRC again?) that are so heavily filtered / censored such that no AI content is ever tolerated? Do we do the Sam Altman thing and try scanning a billion eyeballs as 'proof of human'?

I used to be an AI optimist but this is just ridiculous. How do you all cope?

Everything changes, nothing is forever. It's a hard truth, but the more capable you are of accepting and rolling with it, the better your life will be. I've heard it said that the tagline for the theory of evolution should really be "survival of the most adaptable" , and I think that applies at every level, from the species, to the culture, to the subculture, to the family, and finally, the individual.

If the internet you knew and loved is gone, mourn it, move on, find a new thing to love. Or create it.

Thats a really nice take actually, love it

>I've heard it said that the tagline for the theory of evolution should really be "survival of the most adaptable"

That is the tagline. In the phrase, "survival of the fittest," "fittest" means "most adaptable," at least according to the Merriam Webster English dictionary.

I don't think that is what was meant by the phrase. Ultimately it refers to how natural selection works.

The adaptability of an organism might be an important fitness factor, or it might not. There are organisms which have survived largely unchanged for millions of years.

Perhaps because they were lucky? What they needed was available for millions of years. Note the survivorship bias there. I am sure you already know it, so please excuse me pointing out this obvious thing.

Sure, could be their environment simply didn't change. In which case, a simpler, less adaptable organism may be fitter than a more adaptable one (if the cost of adaptability was higher).

Species go extinct all the time, there are always very unadaptable species around and most of them go extinct pretty soon after they become unadaptable because environments change all the time in the grander scale of things.

Right, fitness can only be understood in the context an organism is in at the time. There is no long term plan.

An adaptable organism might stay around for longer - or it might be outcompeted in the short term by less adaptable organisms.

Over time, being adaptable is what really matters for survival. If something could adapt perfectly, it’d always come out on top.

Sure, “survival of the fittest” in the short term is obvious, but it’s hardly worth pointing out.

I might be missing it.

>or it might not

To determine that it is not, would by definition imply an observation that it is no longer fit.

In other words, how do we know that it is not adaptable? We'd see that it no longer propagates, and is therefore unfit.

Can you think of a case where these two synonyms are not inseparable in meaning as pertaining to evolution?

I can't. I think Darwin meant exactly this.

I suppose earlier generations had to sit through all this huffing and puffing with the invention of television, the phone, cinema, radio, the car, the bicycle, printing, the wheel and so on, but you would think we would learn the way these things work, which is this:

1) everything that’s already in the world when you’re born is just normal;

2) anything that gets invented between then and before you turn thirty is incredibly exciting and creative and with any luck you can make a career out of it;

3) anything that gets invented after you’re thirty is against the natural order of things and the beginning of the end of civilisation as we know it until it’s been around for about ten years when it gradually turns out to be alright really.

Apply this list to movies, rock music, word processors and mobile phones to work out how old you are.

-- Douglas Adams, "How to Stop Worrying and Learn to Love the Internet"

https://douglasadams.com/dna/19990901-00-a.html

From a technical perspective, I do not enjoy RAG prompt engineering. It is so brittle and non-formal, and the barriers of entry are extremely low.

There is a complete trifercation of ML: 1. ML Engineers: the high priests, with access to 10K GPU hours, designing novel Transformer architectures using Tensorflow / PyTorch / JAX. 2. Data Scientists: conducting SFT on pre-trained models via the HuggingFace APIs + MLOps & model optimization (eg via TensorRT). 3. GenAI devs: building LangChain orchestrations and RAG prompt flows using off the shelf LLMs commoditized behind APIs - no stats or linear algebra required.

Too many are jumping on this GenAI bandwagon, which will result in a massive hype-cycle trough of dissalusionment and potential VC AI winter.

Furthermore, GenAI is a local maxima on the path to true AGI. COT / REACT heuristics lack the integrated differential aproach of Hybrid AI, ignoring everything previous generations of researchers focused on: problem solving, planning, probabilistic logic, reasoning etc. For true AGI, we need some focus on: 1. concept representation. 2. goal formation. 3. code introspection and self-modification. GenAI is a big distraction from that kind of R&D.

I concur, after LeCun gave the warning to PhD students to steer clear of LLMs I heeded it and can say my career is better as a result: why compete/join a hype cycle if you can help solve foundational problems.

The difference between AGI and LLMs mainly is in the architecture, specifically that LLMs freeze the entire network, whereas AGI would leave some of it perpetually unfrozen so it can update as it goes.

It's funny how so many people are in your situation about to lose it over AI while I'm sitting here waiting for AI/LLMs to affect my life..like..at all. I seriously can't think of one major experience, positive or negative, that I have with AI in my day to day life.

> being bombarded with tons of AI-written books, pornography, music, video, audio, images

I don't come across any of these unless I am very actively looking for them.

> try googling any historical figure and you'll find some bullshit AI generated image of them likely before the actual one.

I just tried, and there wasn't a single AI generated result in the top 5 pages.

I don't go on Twitter/X. I only subscribe to Reddit communities that I am genuinely interested in and which stay spam free. I frequent other sites like HN that have a very high ratio of signal to noise. I don't waste time on garbage SEO-bait blogs.

I'm not even putting in that much effort to have a sane online experience. If I can do it, so can you.

>> being bombarded with tons of AI-written books, pornography, music, video, audio, images

> I don't come across any of these unless I am very actively looking for them.

AI generated "hero" images are pretty common on blogs and other blog-like sites. I run into them all the time by just clicking links on HN. "AI" generated garbage has made me both more blind to them and more annoyed by the whole concept of a "hero" image.

Edit: here is and obvious example I found, just now: https://betterschooling.in/collection/education-and-healthca.... Literally the next post I clicked on after writing this. Look at how fucking terrible and obviously AI that hero image is. FFS, the "patient's" head is an exposed brain the best I can tell.

I was wasting time on YouTube the other day, and I come across a video (I think about the Alien franchise), that had a huge amount of AI generated images and videos. The video started with a message saying it was a remake of a previous one that apparently had some copyright strikes against it. It was disgusting and unwatchable.

I'm starting to see a pattern around people that freak out about AI generated hero images. They aren't great most of the times but going out of your way to discuss the use of the image in a post (here in HN for example) is high-level pedantry.

The (go to) alternative for the authors would be a post without images which is by no means any better.

> They aren't great most of the times but going out of your way to discuss the use of the image in a post (here in HN for example) is high-level pedantry.

They're garbage.

> The (go to) alternative for the authors would be a post without images which is by no means any better.

It would be far, far better for authors to post without images.

I mean, FFS, if you've got something good and relevant, sure use a "hero" image, but if you're going to use dreck, please don't bother.

The alternative isn't no image. It's stock images or taking top listing off of google images.

AI rarely affects me as well, but I'll add in here one thing that has happened to me recently. A company I applied to wants me to have an interview with an AI. Both disgusted and interested, I came across the company selling it: https://tezi.ai/

I have a 3 year old Samsung mobile. Over the past few months Samsung have been bombarding me with targeted ads and emails begging me to upgrade to their new-shiny latest-greatest mobile. Much of their effort leads on all of the wonders that the new AI capabilities they've added to the phone would bring to my life.

I haven't bought a new phone. I was thinking about it but, at the end of the day, what I want is a phone that's my tool, not my new best friend.

Don't get me wrong - I've got nothing against AI. I love phone apps/functionality that will, for instance, allow me to point my camera and click, then show me a lovely hi-res, non-blurry image of the people/scene I want to capture. Getting rid of the shake in my short video clips will, I'm sure, be fantastic! But don't tell me that this was achieved with AI - just go do it in the background and leave me in peace.

But I don't need a phone that boasts about its capabilities. And I don't need a phone that will nag me to do this thing or that thing to improve my life experiences. I've got a perfectly serviceable mother who's more than happy to do that sort of nagging!

"what I want is a phone that's my tool, not my new best friend."

added to my quote book.

Samsung tries very hard to get me to upgrade but my 4-year old s10e still works great, it's super fast, though it doesn't have updates anymore. However, I'm keeping it till it dies because there is no other small Android phone on the market with a side fingerprint reader.

We have to accept it until something else comes along and changes it again. For a lot of us, the internet has been ruined for a _lot_ longer than the rise of LLMs. For me, it was ruined when it became mainstream (Facebook etc.) and commercialised to the max.

I’m sure someone else will make another post just like the OPs in a few years wishing for days gone by before <insert new thing here>.

> I’m sure someone else will make another post just like the OPs in a few years wishing for days gone by before <insert new thing here>.

"I miss the early days of LLMs" is a certain post in the future.

I'm tired of AI melodrama. The Singularity(TM) is not near; we will see at least one more AI winter first. Large language models are neat and useful, but they're not close to AGI. Jobs will be lost, but many companies who cut too many people will get bitten by hallucinations and other problems and will quietly rehire.

I think that AGI is inevitable, but I don't think it will arrive before 2050.

Kurzweil's book "The Singularity is Near" was published in 2005 and at the time he placed the technological singularity in 2045, so 40 years in the future. He said it was "near", because most people thought at the time that either it would never happen or it would happen in thousands or millions of years.

Where do you hide? Sure, you can and should create Dunbar-sized networks where everyone knows your name, but IIRC, Amazon sells AI clones of real books, "Scientific" papers are completely or partly AI. You can't buy an iPhone without AI (hoping my iPhone SE can't support it), your computer will have it, want it or not. And every damn business and agency you deal with will, if for no other reason, it eliminates jobs, and is a shiny sales object that your competitors offer, even with its failings.

I suppose that you could make some money selling "AI-Free" products and services, like cage-free eggs, nitrite-free foods, and so on. Certified "Legacy" books, BAI, before AI.

There are probably health benefits to limiting how much AI-generated slop you consume.

As you note, the dream of replacing expensive and unruly humans with unpaid, compliant robots is extremely attractive to most large businesses.

Go back to manually produced works - old movies, old books, indie stuff untouched by AI. The web wont be saved at this point.

The hiring spree after all these companies which laid people off in favour of AI systems when they realise LLMs can't produce a single thing of meaningful value is going to be absolutely insane. Companies that can onboard staff as quickly as possible to replace the much-needed staff that LLMs can't replace will survive. Lots of companies will struggle to re-hire and many will fail.

Customers do not want to interact with LLMs. They don't want products produced by LLMs. They don't want to read, watch, play with or listen to AI-produced content.

#Barely_Coping

I skim headlines, etc. A growing list of keywords makes me go ... skip!

Of course, there is that gnawing feeling that I'm missing interesting stuff. But that is the price of clinging onto my sanity.

I'm hoping for: it too will pass. We've had AI winters before, and perhaps this winter we'll pull the plug on the data-barns in order to keep the lights on and stay warm. Even EVs need more charging since the cold reduces their effective range.

Technology progresses and evolves in the direction that capital dictates, and this is what the money people want.

Accept it and move on. Cultivate community in your life, let go of the ways that technology no longer serves you.

I'm a programmer with north of 36 years of experience under the belt. In 20 months I will be retiring. I'm happy that this AI thing will have no incidence on my career. But I'm secretly enjoying TabNine sometimes when initializing huge objects.

Read a book from a good author and you will see how just how far AI is from relating to actual human experience.

Spend less time on social media, and the internet in general

But I /love/ the Internet, I've been here since the very beginning. It's like seeing your home town get run down and turned into a dumpster.

Then do your due dilligence and delve deeply into the concept of dumpster-diving in cybernetic contexts.

(SCNR)

> But I love the Internet, I've been here since the very beginning.

You've been using the Internet since the 1970s?

This was funny; since a /long/ time ago. Didn't get access to that sweet 3-university intranet hahahaha

I think you and OP are both right.

You are right because it does feel like shit. I'm not experiencing it with AI, but I did experience it several times with technology already. Not a good feeling!

And OP is right because that's what you can actually do - moving on. The thing you like, in the form you liked it, is already gone forever. What you can do on a human level is to take this experience, grieve for what has been lost to you, and later find a way to move on. What I can say is that I always something that ignites that spark in me, no matter how many times I have felt that I have lost it.

That's not your home town tho, you have no right to say what people should or should not enjoy using. And from other people's view, you are just an old person complaining the past was better.

There's two distinct types of 'AI' fatigue I'm experiencing. The first type is having to see AI buttons showing up in prime places in navigation on every website and app. Even if it's useful, I just want to hide it most of the time.

And the second type of AI fatigue isn't even about the AI technology itself, but it seems like at the same time that people adopt AI in their workflows they often seem to raise their tolerance for made up nonsense. That's not really a problem with AI because we ought to maintain quality standards no matter what tools we use, but so many people were willing to forfeit quality for volume and convenience at the same time as using AI.

What I love the most is solving technical issues with ease instead of having to google forums etc. I know there are complaints about AI not solving more complex problems, but it is a game changer for daily things like fixing issues, configurations, reviewing options in a given sotutions, and quickly be able to learn a technology and its options related specifically to my case. Given the utility I can stand the occasional Ai content.

Resistance is futile. The AI is rising.

But seriously an awful lot of things are mostly AI free, HN included. At least most content on HN is human even if we talk about AI.

I would focus on preservation of pre-AI works before the media rots and they're forever lost.

There's little hope for the future.

Maybe this is an opportunity to create more curated online services which exclude AI slop, while acknowledging that not all AI-generated content is awful.

We are all seamstresses and somebody just used water power and invented the spinning jenny and the power loom. Of course we're upset.

The question is whether there actually was more qualitative content in those spots before AI. (AI improved this sentence of mine.)

The best part of all of it is, there is no alternative or replacement when AI automates everything.

We'll see what happens when the tide gets low.

I’m currently working on an open source project and AI is amazing for me. Cursor/Copilot saves me a lot of typing. I use the word typing deliberately here since AI doesn’t really produce much new code in my case. Just the simple stuff and I love it. Much less strain for my fingers.

Long-term I hope my project will end up in the training data. The more systems will train on it, the easier it is for end users to write code in which they use my project.

Compared to 4 years ago, these improvements are amazing.

The downsides you mentioned are valid. I just use Google and Stackoverflow a lot less so don’t care. For the rest I click Wikipedia or Reddit before other sources. Isn’t it for years already like this? You just learn over time what sources to click and what sources to ignore.

It is distressing, but you can rest assured from the current fruit that the first AGIs will be complete idiots.

I create real things for real people, which I meet in the real world. I retreat to the private, curated cozyweb where the content is manually and lovingly created. I trust the curation of real people over algorithms.

If the world is about to become a loud miasma of AI nonsense, I’ll just build thicker walls around the human-built and human-curated parts. It’s just damn unfortunate for those outside those walls. There will be very few public spaces left if they are prey to automated grift, but we can still make the private ones nice and pleasant.

I don't find this at all, and I just searched for images of Ben Franklin and MLK, and they seemed fine.

I spent a lot of today messing around connecting Anthropic's MCP system to Emacs and while the LLM's tendency to overwrite my code with its own was annoying, it was fascinating to play with it.

I spent a lot of the early days of the Internet arguing with people who thought it was ruining everything, with terrible art, and disgusting spam, and pervasive anonymity. Maybe they were right: but I realised that for me the better conversations, and better progress, were to be had with those who explored those boundaries to understand them better rather than those who reeled away from them. After some anguishing, I feel the same now. The fear is as much a habit as it is a fixed response.

Maybe that's just how I cope, but then again -- that is what you asked.

I like this take. Wrote the post in a bit of a state of rage hahaha

Just googled ai generated porn. Because, you know, I want to check on the progress. And umm, nope not yet

> What do we do?

You do you. You can opt-out of the internet, for example. Among the rest of us, millions (billions?) are happy with this development.

> How do you all cope?

By not being so reactive and making a little bit of effort to go look for what you like in place of being a fulltime passive consumer.

I haven't seen the AI content. If I have it is vert convincing! Ocassionaly a HN reply is GPT ... but it is obviously is and gets downvoted. AI coding is useful to me to save typing. If I know what I want and the copilot suggests the same I take the suggestion. It is handy.

Twitter is a shitshow but that is biologically generated nonsense! And the hot take engagement is spreading to linkedin leading to borification of these platforms and my minimal use of them.

A year from now you will not know if something is AI or not in nearly all cases. "Filtering" is therefore a fruitless endeavor, there are _entire training methods_ (GAN) specifically designed to bypass any conceivable filtering. Hating generative AI is like pissing into the wind, you'll just get all wet. The most constructive way to proceed from here is to make it work for you.

There are so many things you could say that about, and we actually did make real progress as a society. Go back ~80 years, and it would seem ludicrous in the extreme that we wouldn't always have chain-smokers in every bar, restaurant, apartment building, etc. And yet society fought back against Big Tobacco. If I try to walk into my local bar and light a cig, I'll be tossed out.

Rivers that were once poisonous are now swimmable.

People move back to city centers and get rid of their cars.

If generative AI has negative effects on society (yes), we can and we should fight back. Giving up now won't do anything other than guarantee the awfulness continues.

It's mathematics. You can't put this genie back into the bottle. Also do realize that you are a tiny minority inside a tiny bubble. E.g. the number of cars has been increasing over time, and you don't even know. Upwards of 40% of office workers in the West use AI in their work one way or another. Having it summarize a Slack thread or generate a picture for a slide deck is not going to kill anyone or turn your computer into poison or whatever. It might even (much to the horror of the morally pure), make the work more fun at times, by automating the more bullshit parts of it. If we play this one right, our grandchildren might not need to work at all.

You have to realize you are part of the problem. The old traditional forums didn't die because of AI but because you stopped contributing to them. There is a reason HN is still alive. If you want to maintain the "old" web, you have to keep browsing and interacting with the old web. I keep an old fashioned website (https://omarabid.com) and there are barely any viewers. Google completely de-ranked the site for no reason and it's almost impossible to get your articles trending on social media these days if you are not "whoring" your points of something like that.

If you make a small list of such sites, forums, podcasts, you'll have save and maintain the last few corners that exist.

> There is a reason HN is still alive.

Dang’s salary?

HN is the only forum I can think of where not only are the economic incentives between users and owners mostly aligned but there’s a paid moderator who is part of the community (as opposed to the faceless moderation teams of Twitter/Facebook/etc).

All the forums I loved died mostly because of owners shutting them down or giving up on moderation but YMMV

> Google completely de-ranked the site for no reason

The reason was that they decided to remove web search from Google search.

Google is still useful for searching Google sites however.

Personally I'm kind of OK letting some other search engines bloom along with some human curation.

Reply on news.ycombinator.com