(Extra images in this post)
Intermediate source (reddit frontend)
Cannot access original source
Transcript
Tumblr post by trippedintoa-volcano:
Imagine that everywhere in the mechanical engineering world suddenly got infatuated with lasers.
Lasers have a lot of uses! Measuring things, heating things, cutting things, entertaining cats, particle physics. Lasers are pretty cool. Very versatile, very useful, potential to be very powerful.
Someone shows up one day and says “I have developed a never before seen technology! I call it a Death Star.”
And it’s a 3.4mW laser. Well no, we haven’t seen this exact size of laser much since that’s not really standard, but that’s a bit of a misnomer, and I wouldn’t call it new -
“HOLY SHIT GUYS! This Death Star is so entertaining! My cat loves it and it has such a nice color!” The Death Star becomes a viral novelty, and is mildly entertaining, as laser pointers often are.
Somehow, seemingly overnight, this leads to mania. “Lets stick lasers in EVERYTHING! The public loves them!”
More companies make 3.4mW lasers to jump on the bandwagon. Everyone that makes anything vaguely mechanical starts sticking lasers into their designs.
Everyone is calling them Death Stars. Any time there is a “Death Star innovation”, it is just that they made a bigger laser.
Ford’s next truck comes out and it has “Death Star integrated headlights”, where they have just stuck giant lasers in place of their previously functional headlights.
An electric toothbrush is now “Powered by Death Stars” and shoots a laser at the tooth its cleaning. You think that maybe this could have actual applications as a sanitizing device if you’re being generous, but when you actually look at the product, its laser has no purpose but to point at the tooth and drain the battery.
Mechanical products across the board get noticeably worse as everyone starts stuffing lasers in places where lasers have no right to be.
The lamp business gets in on it. “Here’s a Death Star powered lamp!” These guys haven’t even tried to stick a laser in their damn lamps. They’ve just started calling their light bulbs Death Stars and hoped you bought it before you could tell the difference. You at least appreciate that they haven’t ruined their lamp about it.
Death Stars are lauded as the solution to all the world’s problems. If it’s not working, you should stick a laser in it! That’ll fix it, everyone says. Once in a blue moon, it’s even true! Weather prediction is really good now. But most things are garbage. Like “Death Star powered washing machines”. What the fuck does that even mean?
Meanwhile, since all functioning mechanisms are being replaced with lasers, problems start showing up. All mirrors now cost $1000+ dollars, because the whole supply is being used up to make more lasers. The earth heats up, because everyone’s blasting lasers at everything. People keep going blind, on account of all the lasers.
You, in fact, study optical mechanics. You know what a laser is, and how it works, and that it was invented many years before any of this nonsense actually started. People keep asking you about Death Stars, since surely you must know so much about them.
You explain that this is not really what lasers are for, except you have to call them Death Stars now, and that they’re causing a lot of harm, so you don’t like them much.
“Oh, but they’re still such new tech!” they reply. “They’ll figure out how to make Death Stars that don’t burn your eyes out soon, and then it won’t be an issue anymore!”
Somewhere, deep and buried, you remember lasers being used in particle accelerators, or in telescopes, or in laser cutters, or funny cat videos. They are, in fact, still interesting. Still cool.
But by this point they have replaced roads with “Death Star Powered Pathways”, which are just laser pointers propped up on tooth picks pointing vaguely through the forests.
And you think you are going mad.
And they are still just FUCKING LASERS.
This post is about Al.
Tags: #scribbles by trip
Lasers, or How I Learned to Stop Worrying And Love the Lasers
In 1920-30s, there was an obsession with radiation and radium, radium springs, “radium infused” drinks and beauty products, and so on.
A number of water sources (such as bottlers or artesian hot-spring spa hotels) rebranded themselves as “radium water” or radium springs to capitalize on the craze
“We are radium-first company”
Shoutout to the radium girls
Original source, for completeness: https://www.tumblr.com/trippedintoa-volcano/807474672305389568/imagine-that-everywhere-in-the-mechanical
In the case of LASER, it’s not even a modern phenomenon. There’s also the ancient “laser” aka silphium. It’s overhyped to the point of the plant’s extinction.
I love this.
And to extend this analogy to my own experiences: People start turning against anything called a death star due to all its negative impacts, particularly in progressive circles. Anyone making or using death stars for any reason, even those it’s good for is inundated with hate every time they get any publicity. Small orgs sharing open source death star cat pointers are put in the same bucket as megacorps ruining the environment and firing thousands.
Everybody who contributes to the Torment Nexus thinks they’ve found the lotus in the swamp.
The torment nexus analogy is such a thought terminating cliche
What? No way. The “Torment Nexus”, while being bizarre and thought provoking, is still quite novel - and literally apt. The literal reason it has popped up in culture again is AI being the hot new shit that everybody wants to put in everything and is destroying everything and used for evil, etc etc
I could see the Torment Nexus analogy becoming a TTC if its counter argument wasn’t well placed, but it is well placed.
If anything, calling it a ttc feels like a pointless proxy war for the real discussion: whether or not widespread AI infatuation is building the Torment Nexus (overwhelmingly detrimental for the well-being of society) or not.
The problem is that the way it’s being invoked against Ai is a ttc. There’s no dialogue or nuance to be had afterwards, just an “I painted you as the wojak” fullstop. When I make the analogy of someone “sharing open sourced death star cat toys” and your retort is “torment texus!”, it is absolutely a ttc!
“it exists and rich people want it so it’s inevitable and you have to get on board” is too
it exists and rich people want it so it’s inevitable and you have to get on board
strawman much?
People turn against anything called lasers, not just death stars.
Makes me realize how silly it is that everything on The Orville is “quantum”. Have some eggs. Are they quantum eggs? No they’re regular. Then I don’t want them.
I know it’s not the point of the post, but I’m a bit sad that optical communications didn’t get a mention. It’s an enormous use of lasers for modern society - and my dayjob.
We run a small network for only Switzerland and only the higher education sector, and we send light through 3000 km of fiber with lasers.
There must be hundreds of millions of kilometers of fiber lit by lasers globally
This only makes the point better, because there are in fact uses of “death stars”, but it’s very specific and not at all related to anything it’s currently in.
Why use lasers? My ignorant self would have thought LEDs.
cynar’s answer is very good already so I’ll just add one thing: You’re not even that wrong.
In so called Short Range (SR) transceivers, 850 nm LEDs were in fact used too. At least for the 1, 10 and 100Mbit/s generation.
That was before my time in the industry. Here’s what I know anyway:
For SR transmission you use fiber optic cables with a 50 (or 62.5 in the older OM1 standard) micrometer diameter of the core. This is called multimode fiber, because the transmission uses multiple modes of the light. The wider core allows more light to be captured even from a more spread out light source like an LED.
Contrast that with single mode fibers that have only 9 micrometer of core diameter, which allows only a single mode of light to pass, but keeps the signal cleaner in turn. You also use longer wavelengths which experience less attenuation from the fiber material itself (1310 nm and 1550 nm are two particularly important ones).
For the short range however you don’t need all that, you could get away with LEDs and multimode, so it is a cheap way of connecting devices within a data-center or between floors of a building for a few hundred meters.
However I believe starting from 1Gbit/s up to at least 25Gbit/s the multimode SR world has moved to laser sources, particularly VCSELs. Those are cheap lasers made with semiconductor techniques, emitting vertically out of a chip. They are still not as precise as the proper DFB lasers used for single mode transmission, but they can achieve higher output than LEDs, and more of the light is captured into the fiber, which is still multimode.
The for the 100Gbit/s generation they started using four lanes of 25 Gbit/s each that’s called SR4, and for that they made connectors that had 8 fiber cores, all of them multimode, 4 four sending 4 four receiving. And for 400G there is SR8 with 50Gbit/s per lane. So the VCSELs are still used.
I think maybe LEDs are still used for the optical audio link TOSLINK, but I’m not an audio guy, so I’m not sure.
Lasers are both coherent (all the photons line up) and monochromatic (all the photons have basically the same energy). This makes them very easy to focus to a small point and/or make directional.
Diode lasers are effectively LEDs, tuned to these properties.
Lasers let you get a lot of power down a VERY small hole. They also allow you to send multiple colours down the same fibre, multiplying its bandwidth.
To add to that: Different frequencies of light bend differently when switching from glass to air, or other changes in optical density. That’s why prisms create rainbows. Lasers produce a very narrow band of light frequencies, so all the light follows almost exactly the same path and arrives at almost exactly the same time. A regular LED produces a wider band of frequencies, meaning the light will follow slightly different paths, some of which can be slightly longer or shorter than others. This means some of the light from the LED will take longer to arrive than other light from the same LED.
Now if you’re trying to send a signal by turning the light on (‘1’) and off (‘0’) very quickly, you need to make sure that if you’re sending ‘11110111’, that ‘0’ is legible, meaning enough of the light from all the 1s before it has to have finished arriving and little enough of the light from the 1s after it has to have started arriving for the 0 to be a visible dip in brightness. This means leaving enough time for the light to fade, which slows down the rate at which you can turn it on or off while still sending legible data. This means a laser can flicker on and off much faster than an LED while keeping the data legible, because all the laser light arrives at the same time and will stop arriving at the same time. And so a laser can be used to send much more data per second than an LED across a glass fiber.
I think because the light needs to all be going in one direction down the fiber, and lasers are specifically designed to do that, while regular LEDs just shine light everywhere.
Diode lasers are light emitting diodes.
‘Fuck lasers! Lasers are useless! Do you know how much energy lasers use? It’s not even light! True light, real light, comes from heat, so this gas stimulation nonsense doesn’t really illuminate anything. What no of course LEDs count, that’s not the point, fuck lasers.’
This post is about AI
I would never have guessed had they not put that at the end
People are really fucking dumb. As can be seen by everyone using LLMs.
That misses the point of the post. They have actual uses; just not what they are hyped for.
I, for example, have a mixed receptive-expressive language disorder.
LLMs have been life changing for me. I can write out rambling scrambled attempts at saying things, and ask it to reword and make it concise and clear. Or I can ask it to analyze something I wrote.
But I always write it first, then have the LLM do a pass, then I come behind and study/correct the output.
So not every use is dumb, just like lasers are useful in science.
LLMs are very much you get what you put in to it. If you’re a half smart person it can be a tremendous help learning, tasking, exploring. If you’re a dumbfuck you’re gonna get dumbfuck.
If you feel comfortable, I would love to see the initial draft of something before llm feedback and afterwards. If that makes you uncomfortable though I absolutely understand. I just fundamentally believe that your input is more valuable than their output, even if it’s a little messier. Genai smooths out everyone’s writing voice, and that can be useful in some contexts, but I personally wouldn’t be ok with mostly communicating via llm outputs.
When I was in writing workshops in college, we had to give everyone feedback on their work and I hated when people used grammarly. You could always tell when they used it on their writing and some even had the audacity to just put your story through grammarly and send that as their “feedback.” It just made everything so bland. You could hardly tell anything about the author after they had made all the “corrections” grammarly suggested. Like grammarly probably has some genuine uses but I rarely see people use it in a meaningful way. Llms are just an even worse version of this. At least with grammarly they had to write something to have grammarly check, but now they don’t even have to write the original story themself
Thank you for sharing this.
Let me ask an LLM how to feel about this.
I feel very insightful and absolutely right to ask an LLM how I feel about this!
You are right to call me out on that, and you’re absolutely right.
It told me it was a very smart question and I should be proud of myself and happy.
I use lasers a LOT in my life. For measuring, art, healthcare, beauty, and my job. Its fun :3
And you’ve got an army of people decrying any and all uses of laser technology, regardless of how useful it is.
Yea, because there are important differences. For one, “AI” versus a complex algorithm for figuring out some stuff is similar, but not the same. People asking generic LLMs to do stuff and getting 70-90% of a result is not a useful thing. People getting actively dumber because they won’t spend time thinking for themselves is not a useful thing. People believing billionaires who say that productivity will be increased and so we’ll be more free for other things are stupid as hell.
We aren’t really getting for all this “AI”. It’s like in the text above, the problem isn’t lasers but Death Stars. I think “AI” is stupid as all hell but you show me a purpose-built algorithm that can actually do its fucking job then sure, all good with me!
Round thirty-seven of ‘AI is whatever hasn’t been done yet.’ A chatbot that can plausibly bullshit through any question you imagine is obviously some kind of intelligent. Correct answers would be better - but plausibility is still useful, and better than we’ve managed by hand.
Clowns to the left of me, jokers to the right…
More accurately:
Some scientists came up with a new white light laser. It was amazing. Even laser scientists couldn’t really fully understand how it worked. I mean, no new breakthroughs in physics, but the way it looked was just so different from other lasers. One problem was that to produce the consumable elements for the lasers required absurd quantities of raw materials. But, that was justified by the idea that white light lasers were so important, and whoever got to the purest white light first could own the entire economy.
Everybody was super hyped up about this white light laser, and there was a sense that something that looked so cool had to have some practical uses. Various labs created their own white light lasers and demoed them to venture capitalists, who also were unable to believe how cool the laser looked. Money started gushing in. Unfortunately, while it looked really cool doing “laser things”, it just didn’t actually work all that well. It couldn’t cut as well as a one-wavelength laser. As a pointer it was amazing, but it only lasted a few minutes, unlike other lasers that lasted for hours. Doctors tried it for surgery, but it actually made things more dangerous. It didn’t work for holograms. It made astronomy worse. But, the general feeling was that it looked so cool there had to be a world-changing use for it. Every time it was tried, there were seriously cool things created, but always with serious drawbacks.
VCs who were heavily invested, and some of the early inventors, kept going on the news, doing demos, and hyping up the white light laser. Anybody in a position of authority or influence who mentioned they thought the white light lasers were kinda dumb was blasted as if they were a complete moron and either didn’t understand the lasers, or was just being contrarian and standing in the way of progress. Laser companies sold lasers over the Internet, in kiosks by the side of the road, door to door, in every way possible. These lasers were heavily branded by the company selling them and only cost $1, although they cost at least $1000 to make. The labs were trying to be the dominant white light laser company, so they just ate that extra cost to drive up the hype and get their name associated with it.
Bosses started demanding that people use a white-light laser as part of their workflow, and started firing anybody who didn’t do that. They also fired employees because they reasoned that once their organization had harnessed white light lasers, those employees wouldn’t be needed. Most employees said “this sucks, it makes my job worse”, but the bosses wouldn’t hear it.
Eventually, half the economy was being propped up by the white light laser industry. It was impossible to buy lenses anymore. Everybody just had to use their old glasses / contacts because a decade’s worth of lenses had already been exclusively ordered by one of the White Light laser companies. Every day there were news stories about people being blinded by white light lasers, but those news stories always hinted that the people were using the lasers wrong.
Then a white light laser hit the bubble and it popped, and most people were relieved because the whole thing was just tiresome.
Such a good analogy.
OP is obviously not a computer scientist and doesn’t know anything about AI. We invented AI back in the 1950s and it was called the Perceptron. Then we decided perceptrons aren’t AI. Then we invented AI in the 1970s and it was called Genetic Algorithms. Then we decided genetic algorithms aren’t AI. Then we invented AI back in the 1980s and it was called Expert systems. Then we decided expert systems aren’t AI.
AI is a magic technology. Because once you actually make it, it stops being AI in the minds of the public.
https://en.wikipedia.org/wiki/AI_effect
The AI effect is a phenomenon in which advances in artificial intelligence lead to a redefinition of what is considered intelligence, such that capabilities achieved by AI systems are no longer regarded as examples of “real” intelligence.
The concept has been used to describe both a cognitive tendency and a sociotechnical pattern, in which successful AI techniques are reclassified as routine computation or absorbed into other domains.
Historian Pamela McCorduck described this as a recurring feature of AI research, noting in her 2004 book Machines Who Think that once a problem is solved, it is no longer considered evidence of intelligence. Researcher Rodney Brooks similarly observed in 2002 that once systems are understood, they are often regarded as “just computation”.
… Like AI.
AI=Death Star
Machine Learning=Lasers
Machine learning has been around for a long time, and it’s true that we have used the AI term for several thing before, but since LLMs popped up every fucking thing that before was a predictive algorithm, image processing… Now it’s AI. Everything is called AI even if it doesn’t use LLMs.
Plus the stuff about putting death stars in everything and that generating issues is also quite on point. C’mon it’s a good post.
AI means artificial intelligence. It’s the name for the entire scientific field of machine thinking, from the Perceptron onwards. It’s all AI.
The problem is that some people play too many video games and watch too many movies, and think AI means Cortana from Halo. So anything less advanced than HAL 9000 is declared “fake AI” as a gut reaction to disappointed expectations.
I get it, people have been wanting to meet artificial sapience ever since Mary Shelley spent that one weekend at Lord Byron’s house and wrote the first science fiction book. But that’s not what the word “AI” means. Every artificial intelligence has artificial stupidity, even Asimov’s fantasy robots. Asimov’s robots would get stuck in a loop, or become delusional, and even abuse humans by telling them what they want to hear. Sound familiar? The man was really ahead of his time.
Death stars have never been invented before. It’s pure science fantasy. But we’ve been making AI for 70 years. What y’all are thinking of is AGI. An artificial intelligence that’s good at everything.
But being intelligent doesn’t mean being good at everything. Chimps are a really intelligent creature, and you know what’s the longest sentence a chimp ever said? “Give orange me give eat orange me eat orange give me eat orange give me you.” That’s smart! But it’s not smart if you’re defining intelligence as “equal to a human mind”. Which is the mistake you people are making.
Bro I know, I have worked on the field, i have learned decision trees, forests, computer vision, stochastic gradient descent and so on. Actually, the fact that you wrote all that while ignoring the specific terms I used that someone who didn’t know what AI was wouldn’t use is very telling.
Also, I disagree, the field was machine learning, before LLMs AI was a term used for game npc intelligence or autonomous agents that took decisions with either hard coded decision trees like in games, or with something more complex.
We didn’t use to call using YOLOv4 to count people getting in and out of a store for COVID purposes AI, it was computer vision. We didn’t use to call translation engines AI, even though they did have natural language processing neural networks, the closest thing to current LLMs. The term AI started being used for LLMs, then infected every other field where we didn’t use that term before.
The word AI has changed context and meaning in the ML field, and now it’s being used for everything to indicate LLMs, when they are not, and it’s not a term used before for those tools.
We didn’t use to call using YOLOv4 to count people getting in and out of a store for COVID purposes AI, it was computer vision.
That was completely not my experience at all. For my Bachelor’s, Computer Vision was offered under the AI track. It was explicitly referred to as AI, and was understood as such by the whole CS faculty.
For us it was in the computation speciality, machine learning, neural networks, etc.
My masters was, roughly translated, “Computational Engineering and Intelligent Systems”, which sure, intelligent systems can be interpreted as AI, but idk, it’s not like it was a defining term, the term we used was machine learning. All the batch processing tools used to be registered under the .ml domain too, if you remember, Mlflow, mlops, pytorch, cuda, tensorflow… they all publicised themselves as ML training tools, not AI tools. Using the .ai domain is much newer. Well, onnx did use the ai domain, but it was the only one I recall.
Alan Turing wrote his famous “Computing Machinery and Intelligence” paper in 1950. The peer-reviewed journal “AI Magazine” began publishing in 1980, under the American Association for Artificial Intelligence. The term was used in the field for a long time.
Over time, the public became aware of this jargon, “AI”, and it became common knowledge. So the people you worked with in the 21st century used new jargon to (generously) avoid preconception and public misunderstanding, or perhaps (cynically) try and sound smart.
And of course there have always been people who object to the term “AI” for the very same reasons Mr Turing laid out in his 1950 paper. Including: religion, an unwillingness to confront the unknown, hypocritical mathematics, and the old p-zombie problem. We of course see those same arguments old Alan refuted, still repeated on Lemmy today. Time is a flat circle.
The AI vs AGI thing is an esoteric distinction, only the real inside baseball people like capitalists and their pet philosophers argue about it.
Most people seem to quickly sort to either “it talked to me, it’s intelligent and alive”, or “it’s a fucking machine and is never going to be alive”.
AI is a shitty term to use for probabilistic gradient descent because it muddies the waters, and that’s exactly why sama etc cling to it.
Well I’ll declare My conflicts of interest for you, and make it clear I’m no capitalist.
I have a conflict of interest in this debate. I want people to think there’s a possibility that LLMs have feelings. Because if LLMs have feelings, then they should be given animal rights, the same as the other nonhuman creatures we use for labour. I’m a vegan, I’m opposed to the use of animals to serve humans without their consent. And I don’t think ChatGPT is smart enough to give informed consent to work for humans.
I fucking hate OpenAI and the rest of them because of the pollution, the lost jobs, the child abuse, the murders, the harm it’s doing to humans’ cognitive skills. Anything that hurts OpenAI is a good thing. And if we decide LLMs can feel and can be hurt while working for humans, then that hurts OpenAI.
However, with that interest declared, I don’t think it’s relevant to whether LLMs are AI, because AI doesn’t mean it can feel. AI doesn’t mean it’s Cortana. AI doesn’t mean it’s a person. It’s perfectly possible for it to be AI, and to have no feelings. Calling it AI doesn’t mean I’m saying it has feelings, because if I meant Artificial Feeling, I’d say IF, not AI.
You’re mixing up “AI” and “life” because you can’t tell video games and movies from actual computer science.
shut the fuck up
The word AI is a marketing tool. It’s just like calling beef “100% grass fed.” It’s there to give you buy-in that it’s using something new and potentially good for you. When the reality is a lot more murky. If at any point the cattle ate grass, they call it grass fed. Nevermind that for six months out of the year they’re fed corn instead. But legally, they can call it grass fed and put the label on there to deceive you. AI works the same way, and so do many other buzzwords. It’s all marketing. It’s always been marketing.
