[AussieWog93]

I just don't understand how someone could think we're going back to a pre-LLM world. Have you not used one since Opus 4.5 came out?

Especially as somebody that's no longer in the software profession and so doesn't have 40 hours a week to just sit there and crank out code, agentic AI is a technological leap on the level of search engines, bulletin boards or compilers.

You can literally just open Claude Code, say:

- "Add these domains with these mailboxes to an SMTP server running on my homelab, forward all incoming and outgoing via mxroute"

- "Port this unmaintained app from 2017 to modern Android, and then test it on an emulator and sign it with my real key for distribution via Google Play"

- "get my GitHub Actions runner on Linux, Windows, macOS and Android, ensuring that it supports all available architectures from the last decade and a half, then check your outputs in a loop until it works"

and it just spits it out while you do the dishes.

Someone the other day talked about "AI psychosis psychosis" and it absolutely stuck with me. The people who deny this is a big deal are in la-la land.

We could literally hit a brick wall in terms of model intelligence now and it would still be a game changer.

[AnonymousPlanet]

> - "Add these domains with these mailboxes to an SMTP server running on my homelab, forward all incoming and outgoing via mxroute"

I laughed reading this. Have you actually tried it? I recently tried to get Claude (Opus) to give me a setup for a simple SMTP relay with Postfix. I had to constantly fight it not to include completely unnecessary components and configurations like Dovecot or just override my requirements. Even Claude is nothing but a semi-competent tutorial clicker.

To someone who never did anything like this themselves the result might seem fantastic. They might not battle Claude on details like I did because "hey, it works, doesn't it".

Don't get me wrong, I agree in general that LLMs are a significant milestone. But I would like to add, especially because of your tone, that perception of how much of a leap it is, is very likely a matter of individual competence. So your enthusiasm might say more about yourself than about the tool.

[rsync]

"To someone who never did anything like this themselves the result might seem fantastic ..."

I've been "doing this" for over 30 years now and I consider these LLM results to be fantastic.

If one has a deep, conceptual understanding of tools, systems and protocols and if one has the ability to carefully provide careful, concise specifications, these LLM tools are like being granted superpowers.

In fact ...

My next project upon which to bring LLM tools to bear is a fully local, solar-powered, self-hosted LLM toolchain.

Not only do I intend to never work without these tools again, I intend to do so without the ongoing environmental externalities that they produce.

[oblio]

> If one has a deep, conceptual understanding of tools, systems and protocols and if one has the ability to carefully provide careful, concise specifications

Most people already don't have the ability to do this and their use of LLMs atrophies the little they had. "AI natives" will be completely doomed.

[jazzyjackson]

Umm, hate to break it to you, but by the time you load up an open weights model, the environmental externality has already happened

[rsync]

This is a fair point - and well taken.

However, I never asked for LLMs nor did my consumer nor professional behavior send market signals that I would use them. They were thrust upon me apropos of nothing.

Now that I find them useful, the least I can do is stop making the problem worse.

[jazzyjackson]

As a rule of thumb the energy use to produce a silicon wafer is roughly equal to the energy it will use in its lifetime. For GPU and VRAM this is considerable. To equip yourself with the capability to run an LLM locally is making an upfront investment in the production of extremely toxic and resource intensive products.

Why buy a personal EV with 600 mile range that sits idle 95% of the time when you can take a bus (aka the cloud) that maximizes the resource utilization?

[goatlover]

But you do plan on continuing to work, not have the AI bootstrap itself into a machine god and do all possible work for humanity instead.

[dehsge]

It’s not going to go back to pre LLM times, it’s also not going to go to a generative ai utopia where we can turn off our brains and a model will understand every nuance of human existence without some from of defined guidance.

Just like block chain still exists, it just didn’t revolutionize decentralization. It created a new tool that can be implemented when you want the benefits of what the blockchain can do, same thing will happen with language models.

[preg_match]

I think the biggest realization to me of this whole thing is that most people don't want to turn off their brains. I think we, very naively, thought that some hypothetical life of zero-thought pure-leisure was not only achievable, but desirable.

And it's just... not. People want to be challenged, they want to read human material, they want to listen to human music. Sure those like crappy Spotify fake artists might make some coin for Spotify. But are people going out to those non-existent tours, or are they instead listening to Confessions II and praying nightly that Madonna tours again?

And, Madonna, if you're reading this: please go on tour. Please.

[phatskat]

Right? It feels akin to people who oppose social welfare and things like UBI because “people just don’t want to work.” The reality seems more like “people want to work, they mostly want to have a say in what that work is.” How many people working service jobs didn’t have access to higher ed, or couldn’t afford a guitar, or accidentally wound up taking care of a family before they could research algae, etc etc.

Personally, I’d be knitting and crocheting things for other people with most of my day if I didn’t have to worry about money, or I’d help my neighbor with the website she wants for her business venture, or I’d teach kids how to code. These are things I’d like to do, that I can’t readily do while having the overhead of a full time job unless I run myself ragged and tbh, I’d rather not.

[neetle]

I think the sentiment comes from people that just don’t find pleasure from working/mastery. The incentives have changed so much with llm-enabled development, but not which ones people respond to.

I miss the code mines

[phatskat]

I’m still in the mines but I don’t know for how long. Part of my performance evaluation for the year will include a section on using AI, so I’ll be biting the bullet and using copilot for glorified autocomplete like I did when it was still in beta.

I’m also perusing for shops that aren’t necessarily “anti-ai”, but where lack of using it isn’t seen as a negative

[sph]

People say they want to get challenged.

> most people don't want to turn off their brains

Biologically, life is an effort-minimizing machine. Our brains have a whole system of habit-forming to minimize the enormous amount of energy it takes to think.

Whatever people say, give them a way to think less, and they will necessarily prefer it unless they literally are a Buddhist monk, seeking to be aware over every action and thought.

[Starman_Jones]

Our brains also have evolved rewards for performing life-continuing behaviors, and then evolved to rewarding proxy behaviors, instead of the actual life-continuing behavior. This is perfectly in line with what you are saying, but now we are rewarded for spotting enemy snipers in CODBLOPS, rather than spotting lions/snakes/crocodiles.

We have gotten very good at hacking this reward function. It has reached a point where a thinking-minimization engine will pay a significant amount of resources to, say, Paradox Studios to do what is effectively the opposite of not thinking.

[preg_match]

They want to be challenged in such a way that they perceive to be enriching to their life. The reason people use AI at work and don't blink, but then won't listen to AI artists on Spotify, is that they see one as enriching, and the other as menial.

Optimizing the menial is the end-goal, optimizing fun and humanity and expression is not desirable. It's just not, and we're seeing that over and over again.

[fabianholzer]

> I think we, very naively, thought that some hypothetical life of zero-thought pure-leisure was not only achievable, but desirable.

May I recommend Aldous Huxley's "Brave New World" to you?

[dtj1123]

May I recommend "Look to Windward" to you?

[maxerickson]

How did you get that impression?

[blizarre]

Yeah and no... I have been a developer for 20 years and most of my time isn't spent writing code. Most of my time is spent monitoring production, being in meetings, doing design work, doing code/spec reviews, creating tickets for infra/IT to deploy stuff...

So yeah I love AI and it really helps my personal projects. But my actual work? A little, I can work on some nice-to-have PR in the background during my meetings. I might be 20% more effective coding? Mostly because I do more of these low-hanging fruit PR like bug fixes and refactoring. The hard stuff still require thinking. But that's not even half of my time. It's fun, but not world-changing.

[toasty228]

> I just don't understand how someone could think we're going back to a pre-LLM world.

That's not the problem, the problem is that for every hour of dev saved society will lose 10 hours on fake data, fake bug reports, automated scams, needlessly convoluted email, etc.

[59percentmore]

Because having a slave do whatever you say can also enable some pretty remarkable workflows, but that doesn't stop it from being too morally and economically expensive to become ubiquitous AND sustainable. LLM usage certainly doesn't involve sapient beings and so doesn't carry the same moral reprehensibility as slavery, but in real terms, it's probably even more of a financial quandary.

[treis]

Nothing contradicts AI psychosis psychosis by comparing a computer program to slavery

[goatlover]

Yes it's a useful tool. What does that have to do with the singularity, recursive ASI, fully automated economies or whatever is being hyped as the radical transformation of society?

[oblio]

1. Everything you said is more about revolutionizing IT though. IT is more easily testable than other things in real life.

2. The gap between the singularity and really strong automation is huge. It's basically that between 1 million and infinity, or even better, that joke about the difference between 1 million and 1 billion being roughly 1 billion.

[saghm]

#1 is what's so crazy to me that people have trouble understanding. No one is saying that we're going back to the pre-LLM world; it's just that the world isn't honestly that different than before. Most people's lives are pretty much the same. Their power bills are higher, but that's not all that different from when gas starts costing more because of a new oil war. The government has some new ways to spy and target weapons, but that's pretty abstract to most people (and it's not like they were having trouble doing those things before). Some new tech companies pop up and have a bunch of money, and some existing ones have less, but nobody outside of the tech industry who isn't already a huge outlier in terms of how much they pay attention to stock prices really cares.

If this is the singularity, consider me extremely underwhelmed. My job my have changed, but to the rest of the world, it's just more of the same.

[skydhash]

The issue is not capabilities (and never was). The issue is reliability.

And that’s assuming you really need to do these type of tasks using an LLM. As GP said, and as you’ve proven, most people are trying to use LLM as an hammer for everything that looks like a nail, without verifying first if it’s a screw or something else.

[YeGoblynQueenne]

>> Have you not used one since Opus 4.5 came out?

Every time there's a new model This Changes Everything. Every single time.

Every time someone points out a limitation of LLMs it's always the same refrain: "Have you tried the latest? Because they're so much better now!". Except this happens with every single model generation. Every time the new models are so good that it's a seismic shift, a step change, a paradigm shift. It's like the finest curve asymptoting ever upwards, a few billions of training cost at a time.

But- to where? AI keeps getting so groundbreakingly better and we're still stuck in the same old world, except now you can...

... port your unmaintained app from 2017 to modern Android? Why was your app unmaintained? What is the societal and technological upheaval that will come from automating such low-stakes tasks? For such huge costs?

[Schlagbohrer]

But it is now possible for a very non technical user to vibe code Grand Theft Auto II for fun. In two years they'll be able to vibe code GTA V for fun.

Likewise with many other things - people who don't know how to code can produce large custom programs which do what they needed. They're probably not amazingly well written nor maintainable by a human with limited time on their hands but that also doesn't matter at all for nearly all use cases.

[YeGoblynQueenne]

That should completely destroy the software industry and the software development profession, but this hasn't happened yet far as I can tell. Maybe we're not quite there yet?

I've only found this guy trying to code GTA 6 before it's released (not II):

https://mashable.com/tech/ai-startup-founder-attempts-to-vib...

Is that the one you meant? I wonder why he didn't try to code GTA 7. I mean, why not?

[UncleMeat]

LLMs are very good at programming. This is clear. Software engineering will look very different in 2030 than it did in 2020. There are also opportunities now for people to use code to solve problems that weren't previously economical because hiring a programmer was expensive.

Will AI function the same way in all domains? That is the current narrative.

[zombot]

> and sign it with my real key

That's the kind of irresponsible foolishness that makes it hard for me to take you seriously.