this post was submitted on 12 Sep 2026
978 points (99.0% liked)
Work Reform
17570 readers
287 users here now
A place to discuss positive changes that can make work more equitable, and to vent about current practices. We are NOT against work; we just want the fruits of our labor to be recognized better.
Our Philosophies:
- All workers must be paid a living wage for their labor.
- Income inequality is the main cause of lower living standards.
- Workers must join together and fight back for what is rightfully theirs.
- We must not be divided and conquered. Workers gain the most when they focus on unifying issues.
Our Goals
- Higher wages for underpaid workers.
- Better worker representation, including but not limited to unions.
- Better and fewer working hours.
- Stimulating a massive wave of worker organizing in the United States and beyond.
- Organizing and supporting political causes and campaigns that put workers first.
founded 3 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Practically, the only use of them right now is to inflate stock prices. No business benefits from this much compute other than hardware suppliers and construction companies.
As someone in the field, I can assure you that a great many businesses are realizing significant benefits. Our PR sucks though.
Are they? The fact that we aren't seeing more completed projects on github tells me its not useful for coding beyond peer-review and generating boilerplate. I've met 2 non-tech people who managed to get a MVP they have no way to maintain. Who is realizing significant benefits?
Github is not the real world. My information is based on our clients. A basic search produced the following...
XXXXXXXXXXXXXXXXXXXX
High-end, frontier AI models are driving major breakthroughs across crucial domains: Healthcare & Biomedicine Drug Discovery & Protein Folding: Models like AlphaFold predict protein structures in seconds, accelerating candidate target identification for complex diseases like Alzheimer's and cancer. Early Diagnostics: Advanced computer vision analyzes complex scans (MRIs, mammograms, pathology slides) to detect early-stage tumors and rare systemic abnormalities faster than manual screening. Genomic Medicine: Processing petabytes of multi-omic data helps design targeted therapies tailored to individual patient genetic profiles. Climate, Sustainability & Agriculture Precision Climate Modeling: Advanced neural networks simulate localized weather extremes, helping cities anticipate floods, severe storms, and wildfire trajectories weeks in advance. Grid & Fusion Energy Optimization: Machine learning manages real-time load balancing for renewable power grids and stabilizes plasma containment configurations in nuclear fusion reactors. Precision Agriculture: Satellite and drone computer vision algorithms analyze soil health, crop diseases, and yield forecasts to optimize water and fertilizer deployment. Scientific Research & Discovery Materials Science: Autonomous lab platforms test millions of theoretical chemical combinations to invent better battery chemistry, carbon-capture materials, and superconductors. Astrophysics & Deep Space: High-throughput processing identifies exoplanets, maps dark matter distribution, and filters noise from radio telescope signals. Accessibility & Public Safety Assistive Technologies: Multimodal models enable real-time scene description for visual impairments, instant speech-to-sign translation, and adaptive interface controls for mobility challenges. Disaster Response & Search Operations: Satellite analysis models process post-disaster maps instantaneously to route first responders and locate trapped survivors.
The thing is: The amount of compute being planned (it's often not getting built) can only have one true goal: automated mass surveillance, media manipulation and dismantling of what remains of the middle class, in addition of fully automated warfare. No other task than rolling out Flock 2.0 and the equivalent of Skynet needs that much compute. All the things you mentioned already work with the EXISTING compute. I agree that some additional datacenters will be needed, but for the things you mentioned a fraction of planned compute is already fine, it seems.
Fusion.
https://www.sciencedaily.com/releases/2026/09/260903064215.htm?hl=en-US
Medical
https://www.umaryland.edu/news/archived-news/august-2024/ai-speeding-discovery-of-new-drugs.php?hl=en-US
Infrastructure
https://developer.bentley.com/use-cases/itwins-and-iot/?hl=en-US
Environmental
https://rfcx.org/?hl=en-US
The adoption of AI to solve problems is accelerating. It's just not making headlines. Good news is just not rage bait. Fear sells clicks.
That's nice, but NOT the endgame. I do not contest that AI can be a positive thing too, but there are very powerful forces, concentrated mainly in California among the tech Billionaires, who want to use AI as a tool to ensure that their will cannot be denied.
The US was already using AI in the attack on Iran, minimizing human involvement in target selection and approving the strikes (thats how over 100 little girls died). Russia is experimenting with fully AI automated attack drones based on Nvidia hardware. Flock cameras are analyzing not only license plates, but data points like bumper stickers and what other devices are in the car, making every personal movement transparent even if you leave your smartphone at home. Home security cameras are using facial recognition (which they only pulled back because they were inept in messaging, but you can be sure this functionality will return, and with full access for the authorities). Laws preventing authorities from spying on citizens get bypassed by buying the required data from advertising companies. Social media is getting flooded with LLM-generated content designed to steer public opinion regardless of what the truth is. Entertainment hardware like Smart TVs is listening in on your conversations in your living room (and in many cases much further away), transcribing your words into text using AI and storing it - even if you keep your TV offline, your chats become company property - and therefore state property after a simple financial transaction - the moment your TV detects an open WIFI, or in the worst case just use technologies like LoRa to transmit the data home.
The people behind all of this are the same people financing far right parties all over the world, because a divided world is easier to dominate.
Society is under an AI-powered full scale assault. I hate it, because I know how much positive stuff this technology can enable, just as you documented. But giving up all our rights and living in an hyperscaled version of 1984 is not worth it.
Agreed. Really.
And just as the labor movement kicked back against the robber barons, we need to do the same. Some beheadings are fully in order. Would you be surprised to hear I'm an anarcho-syndicalist punk? A few weeks of a global general strike will bring those billionaires into line. Let them eat their crypto and stocks as workers share solidarity.
As far as warfare goes, we have banned nuclear weapons and have maintained a global consensus on that. Why not AI weapons? It's about time humanity outgrew war altogether. Perhaps this could be the tipping point. A drone swarm over a few senators homes would change hearts and minds quickly. Governments need to fear the citizens again.
But I digress
AI based weapon systems do not have the shock value of being a weapon of mass destruction. They are only an automated version of basic guns and explosives (that are available pretty much everywhere, including homemade variants) on a mobile platform. And the nuclear weapon proliferation treaty is pretty much dead in the water, with multiple nation states either already having acquired them in the last 2 decades or being so close to having them that the only thing holding them back is political in nature, and these barriers are easily jumped over.
It took WW2 and the potential to sterilize and glass over the entire planet to make countries agree that nuclear disarmament is needed. I do not see the necessary willingness to prevent "automated guns". In a world where nuclear proliferation treaties are crumbling, i would not bet a dime on a worldwide regulation of AI-powered weaponry, especially when looking at how many fucks the big countries are giving when the UN or ICC is speaking.
The only wild card in this situation is that basic FPVs are so cheap and easy to build out of commonly available consumer goods that it can't really be prevented that common people might build them (or outright just buy them), and the people with better technical knowledge will also be able to build their own AI weapons. That's like "the democratization of violent force". Seeing how smart the average person is makes my head hurt when trying to envision such a future.
Once, the crossbow was condemned as being too barbaric a weapon. Until humanity outgrows its desire for violence, new ways to kill are inevitable. This is a problem with human nature, not devices.
Is gunpowder culpable for artillery?
The power of a crossbow to cause mass casualties is pretty low. Even conventional small arms have limited potential when used as single weapons, and there is always a human deciding to pull the trigger, who might disobey illegal orders to kill the innocent.
When implementing an fully automated AI powered kill chain, there is noone feeling responsible for the actions of these machinations, because the emotional distance is too large. AI systems do not have the concept of an atrocity if built for this purpose. And the killing only stops when someone disables it - but i'd say that if someone was unscrupulous enough to have started something like this, he will keep it going just to ensure he won't be punished for it.
I'd say the actual potential to cause harm to innocents of such a system is much larger than the threat of nuclear war, because the latter can only be started by an actual madman who would have in most cases to fight his own staff who actually wants to live. The former i can see being implemented by run of the mill politicians and generals, or even some cartels.
Not sure what about this even is AI. Its a software development platform specialised in Infrastructure projects. I guess the reality capture/mapping API uses AI, but most features are independent from that.
A lot of could, maybe and potentially in an article written during the starting hype in 2024, great source. It would be great if it actually worked, but most studies ive seen so far came to the conclusion that the predictions the AI made were based on entirely wrong assumptions.
Machine learning is a great tool and has been around way before the current hype. It also has very little in common with what people usually mean when they say AI.
The only somewhat proper example you delivered. Sounds interesting.
Let's face it, the examples you've shown have very little in common with LLMs, Generative AI or Agentic AI. Most of it could be done with machine learning algorithms we've had before this started or are just nice gimmicks that didn't revolutionize anything unlike what you claimed. There definetly are a few interesting applications for specialized AI models, but the tradeoffs simply aren't worth it.
For every 5 years that AI brings us closer to Fusion power, it brings us 15 years closer to a cyberpunk esq dystopian future. Its great that I don't have to worry about power costs anymore once I'm locked up cause some predictive algorithm said my kids would be against the current regime.
Not to mention that most of the actually useful applications are developed by independent researchers and have little to do with the big AI Industries that build these datacenter. You know, the ones that get the actual hate
AI is a broad field. The same equipment runs ML, molecular combinatorics, factory robotics, symbolic processing, image analysis, and mega-multivariable processing. I don't even see LLMs as significant consumers of the available processing power. That's the foam on top of the ocean for the large buildouts. A use for the unused capacity leftover in between actual workloads. But it is the visible part. The fact that a digital twin of a bridge doesn't spark your imagination is an indicator of this.
I ran your wall-o-text through bullshit.js and it made quite some good replacements!
So you do think that using compute power to distill wisdom from complex information is a useful function.
Of course, but that has nothing to do with your text. Nothing in it was "complex", or "information" (in the sense of being useful); it was just corporate mumbo jumbo, which I just happened to verify using the tool.
Just so you learn some nuance: there are degrees of magnitude in all aspects ranging from expenditure, availability or soundness, between something like using a scanner and OCR to recover information from a damaged book, and "using AI".
No point in replying. Everyone here is convinced that AI is exclusively a scam to inflate stock prices and brings 0 benefits or productivity gains in any field ever nor is it going to anytime in the future.
This is not to say that there isn't a bubble or that AI will solve all problems in the world, but it does have plenty of use cases and many are already in place. For instance many 100s of likely rape/violence cases have been discovered at my local hospital by running the xrays and scans through AI, which hadn't been identified by the doctors by themselves.
Cool, you invaded the privacy of numerous people using AI... you must be proud.
The issue isn't that AI brings no value - it is the cost of that value. And many people aren't falling for the extremely overrated value that companies are claiming. Promised value coming from people who have a very long track record of not being truthful, lying for their own profit no matter who it harms. And then there are some who would rather not ruin the environment for those unfulfilled promises.
But sure, throw an anecdote out to prove everyone else wrong.
Altman and Musk are verified shit. This is common knowledge all across the industry.
Public companies don't set their value - shareholders do. Insiders know the value hasn't arrived yet. We build because customers pay us to build. The customer fronts the construction costs based on the signed contract. Investors buy or sell based on public knowledge.
Nobody's privacy was violated. Privacy laws are very strict in Italy and before any research is done on health data you have to sign and tick the box to specifically allow it. The data was also not handed to any company as the research was run by a local university.
How was anybody's privacy violated? Also I didn't do anything given that I just read this on a newspaper and am not in any way connected to the hospital.
You are judgmental and ignorant. You think, as many other people here on lemmy, that you are right about everything all the time. I don't deny that the AI industry has very big flaws. I fully agree and I believe that in the long term we will need public ownership of AI companies. I just disagree with the fact that AI is wholly bad and brings 0 positives.
Public ownership will happen, the moment the nation states are implementing a fully autonomous police state, and in extension using AI for automation of the military kill chain - for external AND internal threats.
I hate the fact that this technology also has good sides, because they are used to steer us all into a fucking nightmare.
Imagine if people said the same when telegraphs were invented, or phones, or the internet, ...
All these technologies provide police state material, and non-democratic states do become police states (China, Russia, Iran, ...). But there also democratic states that despite using these technologies are not police states.
The struggle for freedom is a political issue that has been there since the beginning of times and will continue to be there. It's politics, AI won't change this.
AI enables a new form of fully automated implementation of disinformation, societal violence and lethal force, completely disconnected from all morals, ethics and second thoughts a person might have. When the tools of suppression are not wielded by a human that might decided that things are going too far, there also isn't any risk of these tools getting turned against whoever activated them.
If you cannot see that, then you are either already one of their victims through societal manipulation or one of the people switching these systems on.
This discussion is not only for today's participants. I also speak to the lurkers. I believe that this fad of hating AI will pass. That over time the benefits will become obvious.
I also do not conflate AI with the capitalism surrounding it. Just like the industrial revolution brought forth great progress through greed and social malfeasance, this revolution is also a messy birth. But no one can dispute the benefits that came when horse power was replaced with horsepower. In the end, this will be the same.
In the meanwhile, we can hold the industry to high standards. Sound levels. Water usage. Energy consumption. My company supports setting standards on all these. That would produce a level field as we already do it volunterally. I'm all for suing bad actors into bankruptcy. Using on-site gas turbines for primary power is a crime against humanity. As is draining the water supply for cooling.
Great points.
Especially regarding the environmental aspects, I feel like the whole thing could just be solved by zoning datacentres as industrial and applying slightly stricted regulation.
AI doesn't kill the environment by itself, it just needs a lot of energy and there needs to be some planning put into providing clean energy for datacentres without spinning gas turbines or raising energy prices immensely for the average Joe.
These guys built in nowhere, ND.
https://ir.applieddigital.com/news-events/press-releases/detail/123/applied-digital-announces-250mw-ai-data-center-lease-with
Why? Energy surplus in the region. So much electricity that the grid couldn't accept it all. From wind farms. These are the kind of places to target.
If that's true, can you be more specific?
The NDAs are Very strict. Fired and unemployable is the least of it.
This is a primary reason for all of the "trust me bro" vibes. Anyone with real info is gagged.
Do you manufacture guillotines? That's the only industry that is realizing actual, genuine benefits here.
Everyone else supposedly making money is in a circular ponzi scheme.