I spent my life as a researcher in the ML space (now called AI) because I didn’t want people to have to deal with the same hardships I did. I saw the incredible strides we made in search and creative tooling. Nobody else will have to spend days rotoscoping a character. We will be able to find any picture just by searching for a blurry, low resolution one.

But it’s turning into a system that is used to funnel money from the most vulnerable population into the coffers of the wealthy. Instead of equalizing, it’s exacerbating inequality.

As I see it, a model trained on pubic data should be a public good. You can charge people at cost for inference and training, but unless you provided the data you’re obligated to give the weights back.

There’s a lot of hostility to AI and I think a lot of it is ill informed or ignorant. People are opposed to data centers, but they’re focused on the “AI” part, when in reality said data centers use less water than laws and serve more than just AI. They give you Nerflix and search results and YouTube. (Not to say it’s never an issue, just that there’s more nuance. Water constrained areas are not a good place for data centers.) I’m personally more opposed to data centers because it prevents me from owning my own hardware. I want to buy a thing and use it as I desire. I don’t want to pay a subscription fee to use my own computer.

There are a host of other objections, but I’ll leave it here for now. There are a lot of good reasons for hating LLMs, but mostly people jump on the bandwagon and hate them for stupid or inaccurate reasons, preventing us from having productive conversations.

everyone i work with that uses it to generate text has lost the ability or patience to read anything, even very light text written by a human.

at best they skim, but mostly they demand in-person conversations so they can be spoon fed the contents of the documentation.

gone are the days of “this meeting could’ve been an email” because they won’t (or can’t) read the email.

so yeah, i hate it but am required to use it for work. which makes me hate it more.

If we’re really, really lucky, the crash might take a few more industries with it, which could buy us a bit more time with regards to the climate collapse. If we’re rather unlucky, it’s just mindlessly accelerating suffering.

It has its uses. Generative stuff is ok **as long as ** you keep it to yourself. Its sometimes easier to ask llm for explanation on some obscure thing / to take a look at what information on system logs could mean. Its neat gatget. Kind of on par with some kind of useful tool but too unreliable to be of real use. And even if its useful to some, the cost and looming threats are not worth getting on the bandwagon (environment, if its successfull you wont be able to afford it, if its not capabilities will be reduced greatly if not completely and its not much use for you then). And most of all, if its as successfull as they want to sell to investors; have fun being unemployed and looking for job with other couple billion people as once the rich get theirs you wont get yours.

Other people who lose their minds because of it are the problem imo, though ones that do shouldnt have been on anykind of position that matters in the first place.

The small help it gives does not equal the vast evil done to the environment, communities, and actual makers and creators.

Broadly speaking, fuck AI

Currently, it’s a skill amplifier. You can’t code for shit and rely solely on it, it will give a shit product. You know how the whole codebase works and use it as needed? It will speed you up drastically while maintaining the quality you want.

Agents are a nightmare. They introduce a variable completely uncontrollable by you to your project, which will make you lose track of how everything works, resulting in a shit product.

I see its future on more optimized models that can run well locally on lower hardware, not those that use entire datacenters to power it. The big AI companies still do not generate profit and are doomed to fail in long term, but AI itself isn’t going away.

I feel like it’s making people dumber. I keep getting screenshots of Google ai summary from people as if that’s the answer and they’re confident in it. I keep finding those are not correct. For me, ai has created more work because I can no longer trust my colleagues statements.

There are a lot of things that AI can do better than any human.

Thinking, writing, conversing and art are NOT those things.

Let AI analyze, not create. End the slop.

Artificial intelligence is good at doing boring repetitive tasks or point out things for a human to look at when there is too much data for us to scrutinise.

If you do not have the skill to verify the output you are almost certainly getting garbage out of it.

The fact that they cram AI into everything is because they are throwing AI at the wall and seeing what sticks. Spending millions on a 1% chance that you will make billions makes economical sense if you can afford to lose.

“Everyone is trying to use a Magnifying Glass to do everything from change flathead screws to preparing food, from levering open crates to surgical incisions.”

In other words: The base algorithms and techniques are very interesting and have valid applications in a narrow range of fields, being useful for pattern recognition, generation, and analysis.

The problem is everything else. Everyone is trying to use that one tool for things it is neither effective nor practical nor appropriate to use it on, and then being shocked when this has widespread negative outcomes for both themselves and everyone around them. Separately but parallel, the drive to find useful applications of it (in areas where there aren’t) is causing companies and individuals to engage in grossly unethical behavior.

Even in the fields where it is most useful, you have to be cautious about not treating it as a “truth machine”. A speaker I heard put it simply: “AI cheats.” It will latch on to the most prominent patterns, not necessarily the ones you want it to analyze.

We’ve been using AI for decades and while generative AI is widely despised and mostly deserving of that despite, lumping all AI together while calling out the harms is doing a disservice to a lot of useful AI techniques and tools.

Its shit and I’ll be glad when it gets flushed.

AGI doesn’t exist and is not being developed, but the idiots with all the money and power believe it is. Their stupidity and insistence on feeding all our resources to this stuff is gonna kill us all.

On one side…

On the other…

I’m a shameless local LLM shill. I have one loaded right now, experimenting with it.

I also thought Sam Altman was a psychopath before that was the popular opinion, and have never subscribed to OpenAI or even used their free UIs. Not once. Not even the “original” GPT 3.5.

I don’t plan on ever doing it. I will rail against them.

In other words, the sides aren’t mutually exclusive.

One can hate “AI” with a burning passion, using the term AI pejoratively. And be passionate about local ML.