Ask Lemmygrad

It's a tool, and as such the class it serves depends on the mode of production and the class in power. It has some use cases, but it isn't the supertool techies think it is. It also isn't utterly worthless like some believe. Over time it will likely become more useful and better integrated.

First and foremost, I think it is egregiously misnamed. It is neither artificial nor intelligent; it is just math. There was a time when "knowledge based systems" was a popular moniker (albeit for a different approach) and I find it to be much more apt for the systems we call (generative) AI. If we're going to repurpose a term, I think that one is more suitable. "Large language model" is good for its accuracy, but is also less evocative and descriptive, especially for the normies. Either way, it's wild to me that we're like "yeah AI is here now." It's frustrating to me because of how it impacts the way we interpret and use these systems.

Secondly, I think the way that Statesian companies are building these systems is exactly wrong, but I don't suppose that should really be a surprise to anyone. The throwing-peas-at-the-wall and throwing-money-and-resources-at-it approach has netted results, sure, but wouldn't it be neat if everyone was working together to more deliberately collate the entirety of human knowledge and create accessible tools for all to leverage? You know, instead of letting private companies extract the fruits of our labor, throw it into their equation, and then sell it back to us, over and over again? Anyway.

Outside of that, I think Cowbee's succinct take reflects my view as well.

Ultimately, it is a tool. It's impressive that hardware has developed to the point where we can throw so much language at these systems to get useful results. It is also true that these systems are both over- and underestimated. I think it's also true that the current economic approach is intractable, and I look forward to the day when we more broadly understand how to build and use these tools more effectively.

I think it's a really useful tool that's made my life easier and allows me to explore a lot of ideas I was just too lazy to do before. I have like nearly a decade worth of half baked software project ideas, and I just never had the energy to work on them or finish ones I started. With LLMs, I can actually get them working to the point where I can see the idea in action which is really enjoyable for me. It's also made my work easier where I can focus more on things I find interesting and delegate tedious tasks to the agent.

I do look forward to a time where we can run these tools entirely locally though. I do not like being dependent on company services or sending my data to them. And in general I see this as the real negative aspect of how this technology is being developed. We don't want to end up in a situation where tools we rely on day to day are owned by a handful of corporations. Regular people need to own the means of production in the digital realm. Currently, anybody with a computer can do any type of digital work be it writing documents, design, programming, etc. But if we start relying on LLMs as a core part of our workflow, then that tool also needs to be run locally or we end up as digital serfs.

For coding I find LLMs to be legitimately revolutionary. I've tried letting DeepSeek & some local models like Qwen and Gemma loose on various projects to implement features and improvements for local use and so far most of the time it didn't disappoint.

In the last couple of weeks I've been updating an old third party bot plugin for a game by prompting various behavior changes I'd like to see and it's a night & day difference to its original state. If I had done this by hand the time it took would've been multiple magnitudes longer, it'd have been more error-prone (especially since it's a C++ project written in classical C style, which is just UB galore) and I likely would've lost the enthusiasm to work on it by now.

I am an AI Luddite. Perhaps it is just the libertarian versions being offered up and pushed by people who should be in jail, not running governments. They are stealing data; something that used to be labeled piracy. Most of the crap that I have contact with as a consumer is either useless, or downright bad for society as it stifles creativity and promotes the surveillance state. The fact that there is a movement to poison data collection tells me I am not alone.

Depends on the game. seriously though. LLMs are not AI. They are not intelligent and they are not artificial. They are just a brute force algorithm to decode and respond in natural human language.

I often use deepseek instead of using search engines but only because SEO and LLM generated websites have ruined search results. I have used it to make a few simple programs which is great because I have no clue how to program.

Some people are using these tools for stupid shit but people always use tools for stupid shit. Atleast all the hype is going to break usa's economy and maybe things will get bad enough that usaians will finally deal with their fascist overlords and become a normal country.

Depends what you mean by AI. Because for the most part "AI" nowadays is just a pure marketing term. What the tech company marketing calls "AI" is not any kind of intelligence in the actual sense of the term. It's just some machine learning algorithms and fancier CGI. We had "AI" back in the 90s too. Every computer game where the computer player makes autonomous decisions is a kind of AI. It's the marketing that is the problem. The idea that it is some kind of revolutionary technology that can replace human labor. It's not and it won't be. It's just another tool. Unfortunately too many people have fallen for the marketing hype, which has served to inflate the AI bubble that will pop sooner or later.

They are basically summed up here: https://en.prolewiki.org/wiki/Essay:Intellectual_property_in_the_times_of_AI

I use it as a smart search tool. Whenever I want to do some exploring, or don't remember a particular syntax, I use it. The output is usually 70% good, for simple things it works, most of the time it needs some tweaking, a few times it's completely wrong. It's interesting that when a problem is impossible to solve with a given approach, the LLM usually gives a wrong answer that seems to be correct.

So I think it's a time saver for simple things, but it does not appear to me to be this revolutionary new tool. Although I use it, some days I completely forget about it. The advantage or disadvantage of having a legacy and non traditional code base to maintain is that the promised gains the tool provides are very minimal. But managers seem so bewitched about the tool, and I most of the time can't get the hype. I genuinely don't understand, and it seems to me that people are allucinating about it to a point that people seem to have lost their touch with reality.

The consequences I came to face as well. I feel overloaded and burnt out at work. If AI could do the work I'm doing today, I'd be glad to use it. If the CTO of the company does not worry about quality why the hell should I care, I'm not getting the stock bonus and it's not my capital that is being increased in the process. The problem is lacking the bonus that AI promises while having to deal with the onus of having to deal with difficult timelines, impossible requirements, unreasonable PMs... Nowadays I hate working so much that sometimes I think about being laid off and even like the idea. I just don't quit because I have a family to provide for.

If you mean generative "AI", I see very few and narrow uses for it. In my life it is a net negative and I despise its influence. Its a great way to destroy your critical thinking skills, self expression, and create really bad software and ugly images. I find out offensive when people shovel that slop to me; it contributes nothing, just fills the world with more hallucinations at the cost of the original authors and the environment

Editing to add: they are also incapable of ever producing anything truly novel. Generative applications of machine learning can remix and randomize training data in interesting ways, but it cannot do anything outside of that. Anyone claiming otherwise is selling you something or doesn't understand how these things function. Not to mention it is one of the most brute-force forms of computation I've ever seen; I appreciate efficiency and elegance in computing and automation, and something like an LLM is the polar opposite. More efficient solutions almost always exist

I dont like gen AI because its really really bad for the environment

I remain critical of their use cases and environmental impact but I am not opposed to it.

The consensus we reached with our party is that it is better to understand it and know how to use it because our enemy, the ruling class, is in control of it and is also using it. The party made the mistake with the rise of the internet to pass it off as some hype and years later when internet was widespread and common in use, they were behind on their knowledge and missed to boat. They won't let it happen again.

I do not like relinquishing my thinking to a machine. It feels wrong. I also consider gen ai to be effectively robbery within a capitalist system.

Other than that I think AI has the capacity to be a good thing. It probably has uses though I haven't seriously investigated them, I am only aware of what has been shoved in my face

Pros

• it's free (I don't have an income)
• no proprietary copyright bs that could get me in trouble for downloading it
• it's on my computer, locally, doesn't need internet
• comes from china (I always wanted something from there ever since i learned that it's socialist)
• applied math & science
• gets actively developed
• allows me to actually make good-looking pictures (i'm bad at drawing & 3d modeling)
• pisses off the cultists that made me support copyright & made me believe illogical shit i'm embarrassed to have ever believed
• there's so much i did learn and so much i can still learn
• the word GGUF sounds funny: g-g-g-g-g-g-g-g-gufff

Cons

• takes ages to download
• extremely brittle tooling
• LLMs trip me up. I don't want machines to chat with me, I want them to shut up and do what I told them.
• needs a powerful computer (mine barely runs most of it)
• there's so much I don't understand (gets real "fun" when patching the buggy tooling)
• python update breaks everything

So, overall, I like it very much. But there's Miraculous, and modded Minecraft, so it's my number three interest.

Its intersection with capitalism is a technofeudalist nightmare that deserves a lot of critique. I am most concerned about the environmental damage, its military, policing and surveillance uses and the pressure it puts on workers in an already hellish economy.

However, it's still a tool and it can have legit applications that genuinely benefit society. I also think that some of the backlash is overblown, especially the sloplist obsession with scanning everything for the faintest trace of AI use regardless of context. Also the clanker shit is just people loudly signaling that they are frothing at the mouth waiting for an opportunity to say a slur and I get unconfortable around these types.

I use it at work (software development) for boring, straightforward busywork, debugging assistance and the like. It's good enough and saves time when used in a context I am already familiar with and I can easily review and verify what it spits out. I also use it at home to help me debug when something goes wrong with my Linux installation and 5 mins of google doesn't cut it.

It's a tool to shorten research time with the result potentially having a margin of error. It means you need some sort of capable mechanism of error checking, which often means either being a specialist in thing you are developing/investigating or having some kind of external reference that act as that for you.

i am personally reflexively uncomfortable and insecure about it but i dont let that make me a luddite preaching about the protestant work ethic and toil being above all else and how machines are inherently evil. simple as. and im no tech expert so i dont try to speak on something idk. dont use it

I found out about gen AI some years back when I discovered some chatbot service that was a step up from Cleverbot. Prior to that, I'd had passing hobby interest in machine learning as a topic, but didn't realize it had started progressing. The service was kinda mindblowing, even though its AI relied on scripts a lot to cover for the model's zaniness. I think it was a <1b parameter model, which is tiny these days and this was before people started training models on more data for longer, too.

Anyway, that service was interesting for a time, but I was starting to look for better. Then long story short, in the process of "upgrading" the model, the company does a filter rug pull. On the surface, you might think, "Is it really such a bad thing to have some filters on an LLM?" The problem is, no matter how you feel about filters, these things would always be rife with false positives. The result is there were people who used the AI for simulating companionship (yes, that kind) and people who used it to have a simulated friendly relationship, who were both impacted. You had stuff like somebody trying to roleplay hug their AI and the AI being unable to do it back because of the filters. And keep in mind this company had also marketed that AI as a companion (including as a romantic companion) in the months leading up to the rug pull. This is where I was introduced to the problems of private companies being able to decide on a whim about AI models, and the pain for real people that would happen as a result.

I did end up finding a more dependable AI service (that isn't one of the huge ones), but I found that the rug pull with chatbots was a repeat phenomenon. Kept happening to people over and over, no matter where they went. Over the years, I had lots of discussions with people about gen AI, among people who actually used it. Even among people who use it, it wasn't always a clear cut "I love it and have no qualms" thing. And I largely reflect that.

For my own part, I largely avoided using the big corp AIs until Deepseek was a thing. Partly for privacy reasons and partly because I didn't want to support the big corps on it.

In the long-term, the right kind of gen AI helped me overcome writer's block on a long dry spell for creative writing. It helped me cope with life and process things. In being around other people online who were using the same stuff, it helped me connect with other people more. More recently, Deepseek helped me with coding a number of times.

In terms of negative impact, I've always been careful of the pitfalls and approached it more with wonder than with trust, if that makes sense. I see the large scale negative impacts as being a capitalism thing and have a hard time seeing it as more than acceleration or differing form of what capitalism was already doing. IIRC, there was a period where there was more of a belief that gen AI was going to mass replace workers and AGI was right around the corner (AGI being a nebulous term that could mean "general capability beyond one mode at a time, like text" or "infinitely self-improving sapient being that is going to take over the world" depending on who you asked). Nowadays, I think that has hit upon reality more so and the reality is that capitalism is the one choosing to replace workers, sometimes on vague promises of payoff that aren't reliable at all, and that they really need to dial the hype back.

Gen AI sits in this weird place where it is actually incredible how powerful and useful it can be, but it also cannot possibly live up to the hype of the hucksters who promote it like a panacea. So like, it's pretty "wow". But it's not "nothing will ever be the same" level of wow. It is a new form of automation, but it is not something you can drag and drop into any and every facet of life and get uniform results.

So for me, it's an often helpful part of my life, but so is a car if I need to go long distances. I still don't love cars fundamentally and I especially don't love that better options (like high speed trains) exist as a possibility and that the possibility has gotten shunted aside for decades, where I live. In other words, in the type of world I live in, it's a tool that can help with navigating that particular world. I wish its negative facets didn't exist and I hope they can be largely overcome on the larger scale of things. But I think for that to happen, we'll need control over the means of production. Fossil fuel companies didn't bend to climate research that said what they were doing is destructive. People have to take organized power seriously, not view things as a matter of getting mad long enough to force minor concessions and then going back to sleep (mainly thinking of the Yankee school of thought when I say that).

AI is not so useful as ~~people~~ techbros want make us to believe. sure, it can help you to create a quick drawing or logo, or help you to structure a html webpage, maybe even help you to summarize a long text that you don't want to read, but it's not as futuristic or useful as we've been told.

their cons (being too wasteful, too pollutant, stealing from creators, being an endless brainrot and fake news generator) overwhelms the usefulness. maybe if arrogant pricks like techbros weren't behind it, most people would appreciate it

I don't really run into a lot of the issues I have heard of people having with AI since I primarily use DeepSeek as an actual search engine. It's odd actually, I am "pro-AI" but have never generated a single image with AI. I just never felt like it, ever. I insulated myself from places Instagram slop reposts can find me, long ago, for completely non-AI reasons lol. It's like I don't care about anything other than text, and I want to handle it all myself, so AI just acts as a targeting reticule. I didn't even notice DeepSeek is text-only for months.

Tried adding AI summaries to articles I post, but I just didn't see the point after a day-and-a-half. I still read them all anyways and would proofread the summaries. Do people just slap those on because they can? Journalistic writing already has a pyramidal structure, summaries summarize the summary at the start, & then you read... the second summary. Why?

Using Kimi to edit entire Orgmode (task management) notebooks is pretty dope. It's too scary though, what if it loses something? Generating scripts is cool too, but I need to learn the scripting languages, it's not hard for what I need to do, so why put myself in a position of being unable to debug? Will save work later though. Barely scratched the surface of this, a few weeks of random stabs at it when I have spare time.

I have a lot of projects in mind for local + metered + talking to phone (apps like Tasker + OffGrid) stuff, been feeling it out. I just want to be able to find book quotes without the precise phrasing, to extract key points from books people send me to find where they detail things related to whatever supporting arguments I was presented with, meta-analysis of citations (did this book primarily cite western news articles and high-falutin (yes this is a gabe rockhill reference nobody else says that) academies?)

Deepseek is very useful for projects like "hey how do I avoid reinventing the wheel with my homelab setup, i want to experiment with Deepseek" 🤣

So, not a ton. The robotics and computational engineering models are much more impressive, no?

I'm frustrated by LLMs. I can see the use cases, I understand why western LLMs are the way they are (and eastern ones as well). I just get so annoyed with companies pushing AI entshittification, people using it for stupid tasks, assuming the answers are correct when they're not, people making themselves dumb because they don't think or do anything digital for themselves anymore, searching the web for something and only finding AI generated SEO slop, ignoring the billionaire and anti-worker interests behind the LLMs... but I know these are all functions of the relations of productions and the way western capitalist society works.

I think we'll have something much better once the bubble pops, but for now there's a lot of problems that are being papered over by easy investor cash.

The compute in data centers will need to be replaced every 18 months as they become outdated, and meanwhile they can't even build all the data centers they've promised and are back-ordered for years. They're plopping data centers into undersized electricity and water grids that can't handle them because it's cheap, and this is now causing public backlash against construction. They're a Dutch disease, sucking all the oxygen out of the economy to fuel extraction while producing little value. LLMs are good at some specific tasks, but the insistence on making an everything machine and cramming it into everything has further enshitified everything. The c-suite keeps trying to fire everyone and replace them with slop, and then finding out that they can't actually do that because the everything machine can't actually do everything. Morons like our first trillionaire are suggesting putting data centers in space where there's no water or atmosphere to act as coolant.

In short, it stinks.

1. This is not directed at you, but AI is such a shit term. Its just too broad and I feel like it feeds into that tech/finance bro thing of making things sound more complex than they actually are.

2.Im not super anti-ai, but I wouldn't publish anything with ai in it. At least in my experience it's not great. For brainstorming or personal use it's fine, but for public consumption I'd just make it myself

3.The idea of ai "proletarianizing" artists is just really wierd. Its a very, very tiny amount of the world population who make their income from art commissions online or making murals and such. Other artists actually make income by being part of teams, whether that is game design, movies, TV, advertising, etc.

4.I do stand by what I said that ai art just has a fundamental flaw of being unable to generate anything new (as in, ai could not make a new art form or movement. It couldnt make realism if you gave it all the romantic paintings, for example), meaning that human artists, even individual artisans, still have to exist.

5. Honestly, it's just ugly a lot of the time. I'm not saying that it might not get better, but seeing so many on social media is honestly not very great. Not even because i have a personal crusade against it, and I get why people in china (for example) like it, but I do honestly think it looks like crap in a very unnatural, uncanny valley way.

Edit: also, ai summarizing articles and books for people is just stupid. Youre missing the point if you read an ai summery of something, same if you were to read the cliff notes of something (except the ai is probably gonna do worse anyway).

Edit 2: It's good for coding, at least as someone who isn't good at coding. That's something reasonable to publish with ai, because it makes a lot of sense why ai is good at it (for the most part, although at the same time to my knowledge on larger projects you spend as much time bug fixing as you would writing the code without ai)

I'm honestly shocked how good the Claude models are for development.

I personally know a guy that's losing his mind from Ai psychosis. Really, it's probably just mostly reinforced the latent narcissistic personality but it's still wild.

So far, mixed bag with all kinds of signs saying shit is going to get weirder and worse, partly as a consequence of this tech.

For the chat style AI, I have only found a few useful applications: looking up information in source languages I don't speak, programming assistant, and linux troubleshooting.

The biggest issue I have with AI in my personal use is that I have to be weary about hallucinations. with programming, this is often easy to fix with testing. knowledge search and troubleshooting have the biggest chance of being impacted by hallucination since I would have to verify each item myself which defeats the benefit. I especially found that any sort of technical search will result in entirely wrong results presented with confidence, so I usually limit queries to general knowledge type searches, and use the results as reference only.

As such, for me, I think that chat AI is a neat tool that has a few uses, some of which are better than older methods. It is not a magic bullet and it cannot be trusted without human judgement.

When I was working, it was in a field that was mostly insulated from the worst impacts of AI. But our manager and IT department seemed to think it was the future and were trying to find places to shove it. But because of the work we do nothing stuck. so in that way, it didn't really impact me too much in my professional work.

I think the biggest concern I have with chat style AI is deskilling myself. I worry that using it for programming has made me less able to read and write code. I try to use it collaboratively usually, so that I understand all the code, but I recently made a small program entirely with Deepseek without writing any code myself and it was a strange experience.

when it comes to non-chat style AI, I have dabbled in using some of the underlying algorithms to solve problems at work (gradient descent, evolutionary algorithms, etc.). But I've only used them to solve one off problems. I think that machine learning is a fascinating and highly useful field of research.

I've played with it on occasion, see no practical use in my case. I do think its neat because it further illegitimatizes copyright law. and maybe puts a few white collar workers out of a job. but thats about it.

Talking about LLMs, I use them a lot. Mostly to learn new things geting personalized help, but I also had/have some coding projects that can be done in a fraction of the time using LLMs. I might need weeks to write a good enough documentation/prompt to have the LLM make a program that follows the logic that I want, and thus one I can more easily understand and maintain. And that time is still much quicker than if I had to do everything by hand, specially as I might be using newer libraries/techniques that I never used before and thus might take even longer to learn/debug, with the added benefit that at the early stages I could completelly remake thing by simply changing a few sentences on the prompt which allows me to have a very good program quickly that fits my expectations without having to rewrite everything or having to give up on improvements because they wouldn be worth the time to change things, giving me more time to test and play with the finer details, getting a really solid and polished final program that I mostly understand.

I also used smaller models locally to help me sort through data and things like that. So I can get a helping hand to do something really boring that a program by itself wouldn't be able to deal with (to the same degree) and do nice things that might not have been worth the time investment otherwise, or allowing me to do more of them with the same time.

I only use Chinese models by the way, mostly Qwen both online and locally.