And I still wonder if there even is any demand at all. Sure it’s used a lot, but most of that is free or highly subsidized services. And from startups that throw money at the wall and have no chance at all of ever becoming profitable. But there is demand from enterprise tho! Is there? The era of tokenmaxxing is over and enterprise is cutting down on AI costs after seeing spending balloon and no real increase in profits or revenue or even productivity. Some people claim to be more productive, but it’s not by a wide margin and the ROI is hard to see.

All AI is used for is filling the internet with slop. If generating an image of AI lobster Jesus would actually cost what it should cost, I wonder if they would still be generated as much. So these AI companies see high usage, more than they can handle, and claim high demand. They buy all the hardware, hardware producing capacity and force the construction of all these datacenters. Often by bribing officials, speed building, cutting corners and not caring about anything sensible like the environment. Instead of also investing in renewable energy, they just plop down gas turbines, because it’s cheaper and faster. And for what? So we can generate more slop?

But no worries guys, we’ll ask Grok how to solve climate change, not a problem! A little more fucking up the environment now doesn’t matter, we’ll get our magical genie to create a magical solution to all our problems. Which is so fucking dumb, it’s doubtful anyone actually believes that.

We already know how to solve climate change, we just choose not to because it would mean giving up some luxuries.

Spot on, this is the thing people often miss out.

These problems aren’t actual problems in the way most people think about. They are problems in the way that they are exercises. Like the Navier Stokes problem is reporting in the news like one of the biggest problems in maths ever. Leading people to think we are stuck on it and can’t progress in some way without a solution. But in reality the limitations of the equation are well known and don’t impact normal every day use. And in cases where they would, we have other tools at our disposal.

The goal of these problems is to challenge mathematical people and push boundaries. Having a proof pop out of nowhere is like having someone tell you the answer to a question, where the whole goal was to formulate the answer yourself. The answer itself is meaningless, that’s why you get 0 points on the exam unless you show the workings out. It’s like having a movie spoiled for you, the goal wasn’t to know the outcome, the goal was the experience the story yourself.

Working backwards from some proof might lead to insights, but it also might not and be harder than working forwards. It also might pick a certain path, where there were multiple and we would be blind to the other paths. Those other paths might lead to more insights, if we had explored them.

The way OpenAI specifically has approached this is also very weird and awful in my opinion. Like they set their sights on yet again ruining some aspect of the world with AI and then throwing billions of dollars at it to make sure it’s properly ruined. For what purpose exactly? Not to make any money, that’s for sure, they are 2 trillion in the hole and counting.

One of the big issues is, a lot of the hardware has actually been produced. Companies like Microsoft are sitting on huge stockpiles of hardware waiting for the data centers to become ready. These data centers probably won’t get built, so that hardware is just sitting there unused. It’s going to take a long while for manufacturers to switch back to making hardware for regular people again and even longer for the shortages to be fixed. I’d say if the bubble pops tomorrow, we’re still looking at high prices for at least two years. And realistically prices won’t go back down again.

Yeah it took some doing on the part of IT. It was only possible using a separate drive and special settings on their part. That way it could be fully managed, but still allow for the dual boot.

Alas the AMD can’t be disabled and the Nvidia card can’t do display output. It’s one of those RTX Pro things that are purely used for compute, not driving a display.

All engineers have those cards in their laptops, it’s very common. Those aren’t for gaming, they are required to run a lot of the more advanced engineering software.

Nvidia itself is fine, as long as you don’t also have an AMD gpu. My main rig runs a Nvidia card and it’s just fine. I use Arch btw

I have a laptop from work that runs Windows 11 and have the same experience with the unusable start menu. I wasn’t allowed to install Linux on the machine. I run Linux on every other machine I have and have used Linux since the 90s.

However after complaining and begging with IT I was allowed to try running Linux on it. It’s a managed machine, so I had to do a dual boot from a second drive IT installed for me. If it worked out and IT was happy, I could run Linux.

Unfortunately it was a total failure. The big issue is the CPU is AMD and has on-board graphics that are normally used for everything. However there is also a Nvidia GPU, I think an RTX Pro, which is used for acceleration stuff. In Windows this works well enough, the AMD handles most of the regular display stuff and is plenty powerful for that. The Nvidia gets used to offload certain workloads, to make sure the laptop is usable for certain kinds of software. On Linux this is a configuration that simply isn’t supported or working at all. Either it just won’t boot or do graphics at all or it’s extremely slow and/or unstable or with some effort the Nvidia chip can be completely disabled and just the AMD is used. Basically turning a $2500 laptop into a $1000 one and unable to run certain workloads. Now getting some of that software to run would be hard anyways, as it’s not typical software and only supported on certain versions of Windows. But I hadn’t even gotten to that stage.

The whole experience reminded me of getting Linux to run on my laptop back in 2002. An unstable mess with half the features missing. Back then I loved the challenge and when I finally got the thing to ping over the PCMCIA wifi card I was super happy. Never did get the sound working on that laptop tho.

Obviously I aborted the experiment and have just accepted Windows 11. It’s fucking awful and I bitch about it every time I need to use that machine. But for now, there doesn’t seem to be a better option.

Nah I tried, the bot is super limited. It gives correct links and references, but gets stuck in a loop a lot. It states all Epstein files have been released with the legally required redactions made, with a link. It states members of congress can view unredacted files, except in cases where the documents were already redacted by the FBI prior to release. It also stated the FBI made a statement no current investigations into any person related to the Epstein files was ongoing and no further indictments were planned. It linked all of this, including that statement from the FBI made in 2025.

If you push further or try to get anything out of it, it just says it can’t and repeats the same information again. So as far as the bot and the FBI is concerned, the whole Epstein thing is wrapped up. Thousands of victims, tons of data collected, but it was all Epstein and Maxwell in isolation. Epstein is dead, Maxwell convicted, case closed. This makes no sense, but if you ask the bot how this can be, it states it can’t express opinions and just repeats the same spiel again.

I’m citing actual real trials, how is that theoretical knowledge. Please explain how a humanoid robot is actually any use? Because I fail to see it. Your timeline seems very optimistic, given there are currently only very tentative trial projects and they aren’t a big success as of yet. But somehow in 2-3 years there would be a big rollout?

As for your job, I’ve been an embedded systems engineer for over 25 years now, it’s what I have my degree in and have worked in. I’m usually more on the software side, but have in the past been a lead designer on fully automated manufacturing lines. Including the physical layout, electrical, all of the conveyers and robot/fixtures and of course the software. I’ve never worked in automotive, but I do know a lot about automation of industrial processes. These days I’m no longer involved in design, but instead fully on the software side. But I do know automation and I do know robots. And to me the whole concept of humanoid robots is fancy marketing to get investor money, not an actual serious solution to parts of industrial processes that are hard to automate.

You are just repeating marketing. Humanoid robots coming to car manufacturing! For one thing that industry is already heavily automated and further automation will need to be cheap to be worth it. Most of the tasks still done by humans are either very hard to automate, or simply too expensive. Often not because the task itself is hard, but the task can vary a lot, where automation loses flexibility. Another issue is access, where a human can easily snake into confined spaces and even get into the car without damaging anything, this is harder for traditional robots.

However humanoid robots solve none of these things. They are still expensive and have trouble with varied tasks, despite what the marketing says. They are also only slightly humanoid, still being mostly metal, so they can’t get into small places and don’t have the dexterity required. They can’t crawl into the car to get something done as they are large, heavy and those kinds of movement are very hard. The tasks could be automated with some kind of snake like arm with many dexterous manipulators on the end, like the ones used in medical applications, but they are too expensive.

A lot of the hype has been around a pilot program done by the startup Figure and BMW. But if we cut through all the marketing BS and hype for a moment. It was one single robot, in a cordened off area doing a single task. It was picking up large metal sheets and placing them on a welding fixture. Something a regular old robot arm could do easily, cheaply and without faltering. Reports are the robot tripped a lot, misplaced the sheets and was out of commission for a large part of the trial. Which was to be expected, given it’s a prototype. It was the second of three planned trials, which has been retired and I don’t think the third one has started yet. After those trials, it’s yet to be seen if BMW will continue with the program.

Xiaomi is experimenting with their robots in their electric car factories and have done some trials similar to BMW. But they have since shifted away from the humanoid design (although the marketing is still pushing this angle), in favor for wheeled robots with multiple arms. They are optimizing the robots for the more repetitive sorting and internal logistics tasks, which are already automated in many factories by traditional conveyers and sorting machines. The hope for Xiaomi is they can be more flexible in their setup, not requiring the huge machines other factories employ. They also hope to learn and develop their robotics technology so they can sell that tech. Given the low cost of labor in the regions they are located, it is most likely costing them money and the robots aren’t intended to replace tasks currently done mostly by humans in that industry.

Tesla is making a lot of noise, but all of their demos have been human operated puppets, or even a human in a suit. They say internally they are doing a lot and it’s all very wonderful, but have given very little evidence of that. They have had both a bunch of executives leave and a lot of their top engineers have left as well. Rumors are the project is a disaster, with people who used to work on car stuff suddenly needing to pivot to robotics. Those are completely different fields with different requirements and constraints. With Musk pushing the pivot from cars to robots, the people aren’t happy and the project is failing. Even the people that were good at robotics left Tesla to go work at Sunday Robotics.

There are other project ongoing, but the marketing and hype is a far shot from the real world. And like I said, those humanoid robots don’t really solve the main issues why humans are being used for those tasks in the first place. The path Xiaomi is on seems much more promising, with robots designed for the task instead of looking like humans.

Car manufacturing has been at the forefront of innovation and automation for a very long time, but I have severe doubts they will actually deploy humanoid robots in a large scale and have it be useful. Show me a robot who can duck into the car and connect some connectors underneath the dashboard and I’ll reconsider. And then do it in a cheap, flexible and repeatable way, so it’s better than just having a human do it. Because a high end surgical robot could do it and with a couple of years of programming it might even do it all on its own. Until some engineer decides we need a different kind of plug and we need a retrofit. Just having a human do it makes much more sense.

Edit:

And I forgot to mention one of the biggest issues with these humanoid robots, they are extremely unsafe. With a traditional robot it’s very simple, it’s anchored to one place or moving on a rail. This makes the range it can operate in very limited, which in terms of safety is a good thing. We setup physical barriers, light screens, warning light to signify operation, camera surveillance and emergency stop buttons at useful places. Even large physical disconnects which are clearly labeled and visible for use in maintenence. Inside that box it’s a dangerous place for soft squishy things like humans, one unexpected move and the human is gone. But outside that box is perfectly safe.

These humanoid robots are designed to be used in human spaces, that’s the whole point of making them humanoid in the first place. But in terms of safety this a now an issue, as you can’t have humans and robots near each other. It’s fully mobile and can go anywhere the humans can, you can’t have stuff like physical barriers or light screens. Camera surveillance is harder and a emergy stop is difficult as well. These robots are no joke, often weighing hundreds of kg. Easily enough to severely injure someone if they make a wrong move, step on someone or fall over. So the only real way to use them is to have them confined to a small area and treat that the same as traditional robots. This makes the mobility kinda useless, as it can’t go outside the safety zone. It also makes it hard to design manufacturing processes around these. Usually there are parts of the process that employ robots and parts that don’t. For the parts that don’t, humans are free to walk around and do their thing, everything is designed such that the robots do what they can and then leave the finishing steps up to the humans (although stationary robots like inspection rigs are used in the final stages). If we suddenly start mixing robots into those human steps, the whole thing needs to be redesigned, probably leading to efficiency loss.

In other areas, other types of AI are already enabling robotic automation of factories and in the near-ish future might make self-driving truck fleets practical. This would be devastating because factory workers and truck drivers make up like 30% of the US workforce – and again, there are no signs of new classes of jobs that could absorb the out-of-work workers, and no plans for a social safety net or wealth distribution.

While I agree with the sentiment this is happening and no social safety net or wealth distribution is on the horizon, this has absolutely nothing to do with AI. This is a marketing talking point pushed by people like Elon Musk in order to drum up investments. A lot of the automation that is happening in factories is just a continuation of the same automation that already started back in the 80s. Advancements in LLM technologies can benefit some parts of this automation, but often only tangentially. The US has also been lagging behind on places like Europe, where labor costs are higher so automation pays off faster, and China where the tech is cheap and produced locally, so it costs less to automate. The US is now catching up and there’s a lot of AI washing going on.

The biggest issue is the misrepresentation of the technology. LLMs are a very interesting technology and can be very powerful, but also has huge limitations. However anyone claiming it’s actual intelligence is simply wrong. It is marketed as being the AI we’ve known from sci fi stories, but it simply isn’t. It can do what it can (predict the next token based on the tokens that came before) and that’s all it can do. The people behind it will claim it generalizes to actual intelligence and thus will allow for human like robots, actual Star Wars droids, but that simply isn’t true.

Now this isn’t to say stuff like machine vision doesn’t benefit from more powerful GPUs with more memory for example. The investment made by the AI industry in this regard does help a lot, although the increased prices don’t really help. Stuff like machine vision also gets AI washed, just slap the label AI on it and get investments, but it’s basically the same technology that’s been around for a while. A company like Tesla claiming their “AI” “self driving” hardware somehow generalizes to humanoid robots is pretty dumb and just the latest in the long line of Elon grifts.

But like I said factory automation is proceeding rapidly, just like it has since the 80s. I feel like the impact of AI on this is pretty minimal. The warehouse robots used by Amazon for example are extremely capable and are dumb as bricks. Which is a good thing in a warehouse environment where you need stuff to be very reliable and predictable and not prone to random behavior like LLM based stuff almost by definition is. Those warehouse robots are also not humanoid, because that would be very dumb. Often factories and such are limited by their design having to accommodate the human form. Designing such places from the ground up for automation and using robots shaped for the job is much better.

This doesn’t take away from the big issue where automation has been proceeding rapidly since the 80s with no regard to any impact on society. In the past a large amount of the labor force was working in factories. With those jobs either being automated away or just moved off shore, it has pushed the population into the service industry. If that industry is now also automated and off shored, it leaves very little room for regular people to get a job. This should absolutely be talked about as a big issue. But hopefully we can talk about it without repeating billionaire marketing bullshit.

Meh I’d argue journalism has always been this way.

In the past it was more separated perhaps. We had the front page of the paper for breaking news, then the next few pages for in depth articles, there was the whole sports section etc. But there was often also a gossip section with news about celebs, there were the comics, general knowledge and background pieces and often fluff based on rumors. Some papers opted to be more serious and not include such things, often made by a printing house that also made straight up tabloids. Some papers focused exclusively on financial stuff for example. But even back then we also had our magazines and tabloids, some of better quality, most of lesser quality.

These days with websites it’s become more common to just have everything on the same website and have the users figure out what they want or don’t want to read. I’ve seen sites that usually post nothing but fluff and summaries of other peoples work, sometimes put out excellent articles with a lot of own research and sources. Often written by experts in their fields. But in order to fund those pieces, they need a lot of filler to keep the views and money coming in. Having just a little bit of info, like a rumor and writing an article about it has been an important skill since at least the 80s.

It’s easy to point out the lower quality as an example of how shit the world is, but I can tell you from first hand experience we had some shit journalism back in the 80s as well. There were certain kinds of newspapers that were known for that lower quality. What newspapers you were subscribed to said something about you as a person and your place in society. Often people who read only the serious stuff looked down on the common folk with their rag newspapers. And the people who read the more common papers looked at those people who only read the financial stuff for example as boring and stuffy.

Why do you think there has been such an emphasis on hacking with LLMs lately (especially by OpenAI). They figured out all those vulnerability databases were an excellent source for training material. In the past they scraped those, but just for general language training. Now they’ve specifically trained the models on the information within. Some team figured out how to use that data to train a model and have testing scenarios automated so they could write a good reward function. It wasn’t that they figured out the models are good at hacking, they ran out of content and found a new source of good data.

With all the books they’ve been scanning I wonder if the next thing is going to be a writing assistant or editor or something like that. Even though writing good books is an art form and the actual writing down of the words is the easiest part (still not easy tho).

These companies are starving for content and they’ve not just poisoned but absolutely destroyed the content well that is the internet. Given they were already hitting diminishing returns hard, it doesn’t matter too much to them probably. But more compute and storage has also been hitting diminishing returns hard and customers are complaining about the cost. So they are getting a bit desperate on how to improve these things at all.

Yeah except this is a tiny little plane and a 737 is huge. So those comparisons don’t make much sense.

That’s one of the things that frustrates me most about this whole AI thing. I fucking hate it and I want it to die, I wish it were never created in the first place. But from a tech enthusiast and a maths nerd point of view, it is super interesting.

Like the performance of these models is shit compared to a real person doing actual work. But if we think about what we are doing on a basic level, the performance is way beyond what I would expect it to be. I wouldn’t expect it to be able to form a coherent sentence or scale as well as it does (even though the resources required to run these is still very high).

It could have been really cool shit people did studies on and played around with to explore the math. Cool little play models we could let go on a bunch of data and see what it did and how. Something for a small group of nerds and experts who are into that kind of thing, for the sake of learning and nothing else.

But no, somehow it got transmorphed into “AI”. And marketed like this actual learning almost sentient computer system that can replace all workers. You can ask it anything and it will give PhD level expert answers. Oh and it’s run by a handful of the most vile men imaginable who pour all of the world’s money and resources into it, all so they get to be god emperor of the world. Fucking terrible.

I’m going to guess you have a “Don’t thread on me” flag up, but also a “Thin blue line” flag at the same time?

Even if it were all above board, it’s just gambling and should be heavily restricted just as all other forms of betting.

Or the economy crashes by AI replacing everyone and doing an absolute terrible job, causing stuff like supply chain collapse and communication/power outages.

I have yet to see someone explain how this whole thing actually works out. Just a dozen different scenarios that turn out terrible for almost everyone. The only scenario that works out is if a handful of people like Altman absolutely rule the world, become de facto dictators as they are the ones that control the few AI companies that run everything. And by work out, I mean it works out for that handful of people. For the rest of the world it’s probably just about the worst case scenario. But you just know it’s the scenario scammy Altman is hoping for.

You see the internet is like a bunch of tubes