Robots are doing exactly all the things described because they are hazardous to humans.
LLMs are not robots.
Physical AI isn’t that good yet. It’s getting better, but we are still at the beginning. AI hasn’t been out of the AI winter for long; these things take time.
That being said, “dumb” robots already do a lot, and that will probably expand too with new AI support.
Real answer?
People who have access to the tools use it to do tasks they want to avoid. Sally from the civil team doesn’t work in mines or under the ocean, but she does spend way too much time on PowerPoint slides and emails she doesn’t actually care about. So she has Claude handle it so she can focus on her other work.
We have robots doing the dangerous jobs in this meme. Its just they don’t do 100% of the work. They also don’t do 100% of knowledge work either to this day. If you have an AI make you a power point and then you don’t edit it and do quality assurance on it you are going to have one jank ass powerpoint.
Even if the crazy miracles its doing with decomp & recomp work on video games lately: that still requires a ton of human involvement.
They might eventually be able to do 100% of physical labor though. And 100% of knowledge work. They just are not there yet. The former requires use to have enough physical materials though.
But…robots are doing all three of those things?
They do these things at night when they aren’t busy making powepoints.
The very premise of this meme is ridiculous.
Mining was one of the earliest adopters of automation. “Robots” have been doing mining jobs for decades, for centuries if you are generous with how you define “robot”. Mining companies keep searching for any job they can automate. Even the guys who drive the enormous mine trucks were replaced when they found they could let a machine drive the trucks instead.
The only difference is that nobody yet knew how to automate “knowledge work”. The latest generation of AI can simulate knowledge work, which apparently is good enough for some jobs. But, the fact that it hasn’t yet taken some of the few remaining mining and chemistry jobs is a clue that it has severe limitations.
Sure, you can automate the creation of powerpoint presentations, which is “knowledge work”, but is there a mining company that is willing to fire the guy who supervises the robots and replace him with ChatGPT? These are ruthless profit-driven companies. If AI could do that guy’s job, he’d be unemployed.
Current LLMs are limited by the fact that they can’t train during inference. i.e. They can’t learn. If we find something better than backpropagation or diffusion, then things are going to actually be different. At that point, the models would be able to compose ideas together instead of just predict the next token in sequence based on previous training.
Current LLMs are limited by the fact that they can’t train during inference. i.e. They can’t learn
Learning is much more than just training during inference.
True unsupervised learning is still in its infancy, and requires a completely different architecture than the current fancy auto-suggest boom. Unsupervised learning has been the dream since at least the 1950s when AI was an aspiration, even before transistors existed. Nobody has figured out how to do it yet.
If pigs learn to fly, then things are going to actually be different.
If you consider LLMs to be robots I guess? I don’t think robots are writing emails, like a robot is not sitting down at a keyboard and typing out emails in your average office, or any office at any scale really
I don’t care about the overhyped spambots, I’m talking about the plethora of amazing machines that mine, pick plastic out of the ocean, and process hazardous chemicals all day every day. That happens all the time, constantly.
Not anywhere near at equivalent scale.
“Artisanal Mining” (ie, press-ganged children forced into mine work in conflict zones) continues to be a lucrative industry in a way freelance accounting or contracting increasingly is not.
Automation is coming for the high margin jobs, while leaving humans to the low margin manual labor.
You’re comparing slavery to paying for skilled labor tho
Did this meme come from Facebook?
Are you nuts, those robots are expensive. We don’t want to get them dirty.
It gets hot in there too
Technology isn’t there yet. Don’t worry, our overlords cannot wait to replace us with soulless metal that cannot unionize.
Software vs hardware
turns out, a lot of manual labour tasks are too complex to automate easily. even if the work is repetitive and seems like it would be simple, there’s a lot of things that you can’t just stick a robot into and have it work
computer tasks, on the other hand… it’s a lot easier, and there’s much better training data, and the stakes are theoretically (and often in practice) lower.
now, as for WHY we’re not globally prioritizing automating the dangerous stuff completely, that’d be because money, I imagine
edit: I’m referring to semi-autonomous robots here, not general automation. automation in general is still going strong, as companies love to spend a few millions dollars to automate something so they save on many more millions of labour costs in the future. customers will buy a custom machine from us for a couple million and operator costs will halve while production doubles. the remaining half of the operator costs I mentioned to towards maintenance staff, which is generally a more skilled position than assembly line work. obviously my numbers are rough, but you get the idea
There’s irony in how easy it is to automate jobs that are generally regarded as “higher ups”, more intelligent as opposed to lower-tier jobs. And those managers and executives would likely be pretty bad workers.
I used to despise the thought of AI in the workplace doing my job for me
but after seeing what apparently above-average employees output… I understand why it seems appealing
ex: sent my coworker instructions that made note of particular things, and a template file with clearly documented instructions inside the file. they used the wrong template file, not the one I explicitly sent the filepath for. they asked me to review the file. I made note of the issues that cropped up because they didn’t follow the instructions. they asked me to do a final review. I noticed that some of their fixes to my notes did the opposite of what I asked, or were just ignored. like, yeah, AI can do just as well as that. but it can’t do it properly, and that’s what we need
You get to replace a lot more cost when you swap engineers and artists for AI slop machines than you do when you swap manual labor for actual mechanical robots.
The people with the money to make the bots know this too.
When work is done by the underclass, human life is cheaper than maintaining complex automation.
BYD very famously invested in high end manufacturing equipment early on in it’s history. But, in trying to maintain, upgrade, and expand the equipment alongside increased production demand, the senior managers discovered that this was significantly more expensive than just hiring and training more local machinists to build the cars manually.
They ended up downsizing their automated team and hiring/training a huge number of skilled professionals. This allowed the company to grow far more rapidly, while still implementing new changes to their models with minimal re-engineering of their production floor.
Now, BYD is one of the largest car markets in the world, and they are slowing moving back to re-automate jobs that have become consistent and routine.
Yeah, I recently watched some chinese robots dancing on america’s got talent. What a world. smh.
Allegedly the reason can be found in the published Epstein files. But the people told me never pointed me to an exact portion, though it was something in the lines of:
Can you operate these image recognition and similar systems in reverse? That could be used for art generation, and we could dethrone the left from its ivory tower.
Is the implication that this quote comes from inside somewhere in the Epstein Files?
Robots do, there is anlot of automation now in big mines, Copper be extracted in El Teniente mine in chile, in the middle of the andes and the “miners” can sit comfortably in a control room in the city of rancagua with a cup of coffee.
EDT: Altough part of the mine is still manned.
Those machines are not robots in the sense that they can do stuff at least partially autonomously. Everything is remote controlled, i.e. There’s a human in the loop at all times. Present day robots can’t even do the simplest tasks without being teleoperated.
Who cares as long as thr dangerous work in the mines is not handed out to children and other victims of capitalism anymore?
oh and that mine I cite, El Teniente is state owned so the profit turns into hospitals and schools.
I wasn’t saying automation is a bad thing, i just said it’s not the same as robotics.
Robotics has been a field long before the current AI hype started. It’s been about sensory and precise motors and steel mobility under all sorts of conditions.
I think you’re conflating the term “robotics” with “autonomous robots.”
Robotics is a very, very old term.
Automation of robotics in an uncontrolled physical environment is so, so hard to program for.
This is why we got advanced language, music, art, speaking models first. There’s SO much of that data (they stole and pillaged every scrap, but the fact is that it exists) that models can be trained on.
There isn’t vast stores of raw sensor data that maps physical environments to servo, hydraulic, pneumatics correlated with optical, heat, tactile feedback etc.
If you think about how uncoordinated toddlers are, and how many years it takes to map all the inputs humans receive and learn how that translates to our environment so we can move around without much thought, you’ll kind of get what I mean. That information just hasn’t been collected on a large scale in a way that machine learning can train on it.
This is the underlying challenge of autonomous robotics. We’ll get there, but for now we still need humans controlling pretty much everything that doesn’t have an ultra-controlled space like factory floors.
I was in a grocery store the other day and they had some sign on a Roomba looking thing that was struggling to navigate the produce section. I can’t imagine how hard it would be to create a machine that can reliably and safely mine ores autonomously.
That’s just not true. I can’t speak for mines, but warehousing and collecting garbage from the ocean and other domains do have fully autonomous systems.
Humanoid robots aren’t really used anywhere, but that’s cause they aren’t really a functional product yet.
