(2025) Your recent similar, poorly received, post: Stop calling the Super Productionizer a 'baby blender' (2025) (20 points, 25 days ago, 8 comments) https://news.ycombinator.com/item?id=48613674
>It doesn’t work well for most things: “A car is just a tool, it matters how you drive it.
It does work. Cars still exist today. Society didn't decide to go back to horses.
>The entirety of all ethics involved in modern technological ecosystems and infrastructures rests solely on how a singular person chooses to use something?
No one claimed anything close to this statement.
>Tools, then, aren’t “neutral” in any way.
Tools may not be neutral. But tools are inevitable. They are a solver to people's problems.
> we can simply ask for art and it materializes before us. There is no struggle at all involved, thus the terrible labor of being an artist is removed!
Do not confuse the simplicity of early tools with there being no struggle to use something. The most efficient way to convey the exact vision of an artist is not purely a prompt.
Ngl I had to stop reading part way through. I’m no AI spokesperson, but the opinion expressed here is..
> “insultingly naive”, “overly simplistic”, “immature and self absorbed”
..exactly what I would expect from someone partly or majority through a PhD about tools. I like tools, their design, creation, and deployment more than average (though probably less than the author), but you need to come up for air, man.
Your interpretation of that phrase is oddly specific, highly esoteric, and completely different from EVERYONE not doing a PhD on tools.
> AI is just a tool
That word “just” seems to be doing a lot of heavy lifting for you. This sentiment is not intended to downplay the importance of tools on society. The generally understood intent of this phrase is to quell the current hysterics and to reassure anyone unfamiliar with the term _back propagation_ that AI, in fact, is neither alive nor sentient (in the sci-fi sense), and that it is “just” statistical modeling. Do not mystify the technology.
> a car/hammer is just a tool
While these are unanimously considered tools, I could entertain a long discussion and the “particular specialness” of tools and things that are inherently dynamical systems.
> what prototyping _should_ be about
The huge majority of the rest of your rant (or what I read of it) is awfully presumptuous and weirdly confrontational. Who can say what prototyping _should_ be? You of many people should understand that the creation and use of tools is contextual, sentimental, highly personal, intimate even.
> “But artificial intelligence… intends us not just to sit forward
From the first part of “The phrase”… AI is inanimate. Humans DO tend to anthropomorphize, and the whole point of saying “it’s just tool” is to remind people that… inanimate things don’t have an intent!
> our tools are using us
I find this notion somewhat trite. “The tail wagging the dog”. I’m not arguing that we aren’t impacted by the tools we use nor that the use and proliferation of AI is not impacting society, but it takes a PhD level of mental gymnastics to push that concept as far as you have in your rant.
The important takeaway though, is that you are a human, with agency and autonomy.
> it only matters how you use it
The implication of “How you use it” is that YOU GET TO CHOOSE. what you think, how you think, “how you use it”™, and even whether you use it! The choice is yours.
I don’t myself have a PhD, but seriously, do yourself a favor and come up for air. Read the parable of “The Empty Boat”.
It seems the author is instead arguing that we should say that AI is a destructive tool and we should not use it.
I was hoping the article would not label it purely good or bad, but 1) highlight that AI is not just a tool but a very powerful tool and 2) and therefore it very much matters how we all use it and how it uses us.
In this framing, we can see the things where it helps and hurts us and society and many levels at various intensities.
The article mostly just seemed to say how bad it was, and I don't think being critical of a new tool means to only see how it harms us but rather to see the potential full range of impacts.
I sometimes think its somewhat like opioid. Has its use, reduces pain, but then after extended use you also likely get addicted to it and starting having issues.
I mean I think sometimes it reduces pain. But it can also increase pain? I've definitely stayed up later than I've wanted to, or got into a fight with someone because I paid more attention to the AI coding agent than them, or had it convince me that I had made deep scientific breakthroughs only later to tell me everything was obvious and maybe tautological.
I just think it's a very powerful and complex tool that can cause a huge range of feelings and outcomes.
I don't think the article tries to argue against the good sides.
In my opinion it asks us to not just look at individual gain of using LLM's (which can be powerful), but also place it in context how we as humans built long-term sustainable societies. Most societies are currently ignoring that part for the most part, because they fear missing out.
Probably the hardest problem of LLM's is their ethics. As the author put it:
> Modern multi-billion parameter AI models are scaffolded on and made possible by the largest heist in human history: theft of everything that could be scraped from every corner of the digital spaces we share. Without prevention of and justice for this damage caused by current models, their use is highly fraught, ethically. We, as human beings, have developed complex social forms of intelligence when it comes to dealing with things like credit and provenance, two things that modern models are incapable of.
No one currently seems to have serious solution for this. Solving it the right way would be choosing the hard path. While some artists or writers are fighting for their rights, us programmers are collectively ignoring the issue. It is funny how most companies (and some individuals) only want to use LLM's of they don't train on their code. But is just fine to use some 40 years of collective open-source without any compensation or even acknowledgement.
This side of the story doesn't make LLM's less technically impressive, nor does it make the tools themselves less powerful. But unlike the industrial revolution, which some like to compare the AI boom to, during the industrial revolution humans built stronger productivity solely on innovation without strip mining all intellectual property in human existence.
I can see the benefits of LLM's, enjoy the technical capabilities and be impressed by the technology, and at the same time find these tools wholly unethical in their current form.
I'm of the opinion that it can be a tool if used as one, but that most people are currently interested in experimenting with various sci-fi visions. I don't ascribe any emotion or judgment to that, either. We should be doing the things we're excited to do if it doesn't cause too much harm.
Your comment focuses on usage. The post argues that we can't say "it's just a tool" because there are inherent traits that don't depend on usage.
> if it doesn't cause too much harm.
You touch something there: some of us think that simply using LLMs at all encourages developing / spreading this tech that inherently causes too much harm.
Attempting to command well-meaning people to “stop doing X” as though you’re some sort of dictatorial authority is not only rude, but it’s also ineffective.
If you find yourself wanting to write like this, I recommend reconsidering. Persuade them instead.
you are yourself commanding a well-meaning person to stop doing X, like some dictatorial authority, with no attempt at persuasion. I recommend reading past the headline before making glib dismissals - spicy but it turns quickly against you.
>I’ve been thinking constantly about the common and casual phrase I’ve heard so often, “AI is just a tool - it matters how you use it.” This has been the rallying cry of tech-loving academics who no longer do their own research, tech bros who salivate over generative images of criminal depictions of people without their consent, and business-minded folks who actually don’t care about AI but see this as an opportunity to rake in more and more money for themselves.
Academics seem to like saying that considering any tool or technology to be value neutral, is naive.
And sure, technology choices have plenty of second order effects. Stopping any analysis at just how you are directly using the tool is probably insufficient.
At the same time, i still think that is part of how we use a tool. Society's choices about a tool (or even the choice to ban a tool) is still a part of how we use the tool and not instrinsic to the tool. I feel it all comes back to how we decide to use it. We can use nuclear technologies to treat diseases, power our cities, or we can use it to bomb places. We can chose to regulate the waste appropriately or not. Etc.
When i say things like its just a tool and it matters how you use it, i am claiming it is not inherently good or evil. It can have good or evil (or both) effects depending on how individuals use it and how society regulates it.
This is what is being refuted. The argument is that there is, in fact, inherent badness in the tool, which doesn't quite depend on how the tool is used.
I guess it depends on how you define tool. Like as far as i know, traditionally Molybdenum-99 (an important isotope in medicine) is made by fision of highly entiched uranium. The same thing a nuclear bomb does, just a little more controlled so there is no big boom.
Is that a different tool or the same tool put to a different use? I would lean to calling that the same tool, since its the same process. I think at some point if we divide tools too finely we just end up encoding the use of the tool in the tool definition and the whole thing becomes moot. Like is a gun used for hunting a different tool then a gun used for murdering? Maybe you could say a machine gun is different, but what if its the same model of gun?
[As an aside, there are very real arguments over if nuclear bombs are a net negative for humanity. For all the downsides, mutually assured destruction significantly suppressed major wars]
And then there's Project Plowshare, where we are so obsessed with justifying nuclear bombs that bombing yourself starts to sound like a sensible idea...
There is another platitude people spout on the matter. "Technology is not neutral." Usually without elaborating. Any correct statement can become a thought eliminating cliche over time.
What discipline, outside academia, could position "nothing is neutral" as insightful and helpful commentary?
So is doing nothing, consuming resources and goods created by others, and living in the default state of humanity: brutality, poverty, and early death.
If it sounds if I'm being glib in this statement, I am. This article is an unreasonable amount of words which eventually boil down to "we have choices in what we do": an exactly equivalent statement to "X is just a tool, it depends on how we use it." Which, of course, is the thesis he intended to rail against.
Right. AI is a tool that helps some people "make their situation better." I don't believe technology is value neutral either.
Say a tractor does the work of ten farmers, and the farm owner lays off those ten farmers. Is that the tractor's fault? It's just the farm owner's fault. But we usually say the tractor took the ten farmers' jobs, because the farm owner trusted the productivity gain and made the cuts. That's the real point.
I agree AI has that same dynamic, but I'd argue that people who write this kind of thing tend to come from the establishment. Just look at their background.
People treat "open source" as if it's inherently good. But honestly, they don't realize how open source can actually work in a pretty vicious way for non Anglophone countries. Do you know why? For people whose access to knowledge is limited by the English language, it's hard to sell software that offers less value than the open source stuff the Anglosphere gives away "for free." Now think about it. Can a developing country really produce that same kind of elite mental model?
In places that are ten or twenty years behind in technology, do you really think they can produce technology that surpasses Anglophone open source? At that point, learning English itself becomes a privilege, and that's where the technological asymmetry comes in. And the thing is, learning English in a developing country is expensive.
In that sense, AI actually has a more egalitarian side to it. That's why I find this kind of professorial binary thinking so naive. Whether using it is evil or good depends entirely on where you're standing. When a professor tells everyone else not to use it, it just looks like a "this is about my own livelihood" problem to me.
The entire culture of programming and IT is Anglophone. And to even get a proper understanding, you have to read academic papers. To really use Rust properly, you need to deeply understand polymorphism, starting from ad hoc and going from there. Do you know how much time it takes to really internalize all that? The starting line is different, so you can never catch up. In that sense, AI is an asymmetric tool.
It feels like people born on third base are saying, "What's so hard about getting to home plate?" I have no problem with people projecting their own emotional lines onto tools and creating echo chambers. But this logic that only human made work is inherently good, I just find that hard to understand.
This right here is the hypocrisy of the struggle that the establishment loves to celebrate. Some people can afford to stand on stage and fight the good fight, but others aren't allowed that luxury.
To the person writing that, technology might feel like oppression. But to someone else, it's liberation.
Honestly, what I find hard to understand about this piece is that it completely fails to recognize that what counts as a meaningful struggle is determined by your social and cultural position. Their worldview is just narrow. A person's institutional position shapes what they see as human labor and what they see as automatable labor. That's purely a matter of where you stand.
Who is stopping the third world from using and forking open source tools? I think its a nonsense excuse. Many of their people become accomplished developers when they immigrate to first world countries so foss is not the one at fault but crappy governments and crappy cultures are.
"AI is just a tool" is meant as a counterpoint to AI producing slop [1]. It isnt that there aren't ethical/societal implications with it as with everything we do in society.
[1] If you take the slop and manually fix/improve/verify it to production value it can be immensely valuable.
This is also how I have seen this particular thought terminating cliche used. The problem with the framing is that on the one side you have someone complaining about unrestrained slop and then this thought terminating cliche is offered. Why yes, it matters how you use it and the complaint is that users are not refining the output enough before presenting it to others.
The thought terminating issue with the phrase is that it isn't just a tool. Once you automate its use (automated PR reviews, ticket fixups, etc.), it becomes a process as well. You are almost certainly purchasing it as a service and not as a tool. It also contains elements of randomness that most tools do not. These quibbling points are what triggers the cognitive dissonance to produce the thought terminating effect. It is quibbling so you cannot really argue that point but the point is irrelevant; the second part of the cliche seems to be agreeing with the complaint about slop and yet the sentence is offered as some counterpoint. This termination prevents the actual conversation about how valuable its use is when the quality standard is not lowered.
>Even a hammer, made of wood and iron, requires trees to be cut down and earth to be mined up. A simple hammer requires laws to be written about fair treatment of workers in multiple industries, sustainability of various biological and geological environments, and regulation about the sale and use of the hammer.
Well ... too bad that blacksmiths didn't have access to this amazing tools before those laws and regulations were put in place. A hammer doesn't need all these thing. The iron law of bureaucracy needs them.
When I read articles like this - what I observe is mostly rage born from the fear that the core of their identity - aka being smart in the general vague non threatening and non offensive way is rapidly depreciating asset.
I expected to find the following techno-critic arguments in this blog post, but did not really find them. Here they are, Hacker News!
[Tools do not exist in the void but in a society]. You can't study tools without studying the context that created it. Some tools work in some society, some others don't. One example, IIRC, are Ski-Doo with some native population that have Potlatch-like practices. Some anthropologist gave them Ski-Doo but you have to sacrifice something precious to give back, and soon they had to be burnt. In our society, anything expensive that require maintenance and repair is a bad fit as both "don't have the time" and the skills. We prefer "disposable" objects at the cost we know.
[Tools create potentials, society may realize them]. As hackers, we often like to create new potentials (eg. with Bluesky ATProto, or the anonymous Vuvuzela chat, etc.). We also envisioned that with electricity, then with 3D printing, that people would build many objects of their daily life directly at home. But, from all the created technologies, some potentials are realized, some not. Community Memory, a San-Fransisco pre-Internet electronic board, wrote that more often the potentials that favor people in power in our society are realized. That's Palantir - strong - business model.
[Tools convey intentions]. I said above that tools create potentials. These potentials are not limited to what the tool actually is, but it should encompass all the micro-choices made by ones creating it. To illustrate this idea, take a knife. Knife designed to cook on one side, and knife designed for hunting in the other side, do not look the same. Youtube, by not displaying a big bar telling you how much you already uploaded to their server convey the idea of an infinite storage space, etc. Additionally to the raw utility of your tool, you can convey ideas, intentions, suggestion, affordance, either consciously or unconsciously.
[Our tools maximize efficiency at the expense of everything else]. Efficiency, in this case, could be defined as reaching a specific goal, in a stable & defined context. The critics of efficiency improvement are multiple: often, what really matter is to reach an acceptable threshold on many many different goals. Additionally, most efficiency gain make the process more and more dependent of a stable defined context, each micro-change in the environment and the whole tool / production chain is disrupted. It could be summarized as the opposition between "industry" and "craft". It's also one argument of the luddite: they also had tools to build cloth, that were less optimized in term of speed to build a cloth, and harder to master. But on other goals, that were not considered at that time, like quality, they were better. They thought the better quality of their cloth will save them, when people will find how poor quality are the cloths made by the new automated machine. But it did not happen. Still, from a French perspective, crafts still exist there; it's what luxury brand sells, because rich people know how much better they are. William Morris experimented and published a lot on these topics, it's an interesting reference.
With the 4 ideas (tools do not exist in the void, tools create only potentials, tools convey intentions, we build tools to maximize efficiency only), I think we have some basis to build a stimulating critic of both AI, cars, and knife. But this part is left as an exercise for the reader :-)
People interested in these topics may read Ivan Illich, André Gorz, or Jacques Ellul.
I've read "amusing ourselves to death" and there the author also criticises the idea that new technologies are "just tools".
The invention of the telegraph changed how information is traded and the contents of newspapers, the invention of the TV changed politics, and so on.
The medium shapes the message and induces behaviours from us. We never thought about wanting to promt chatgpt, or watch people doing sports on a screen. But the possibility of doing these things makes us change.
LLMs obviously have and will continue shaping the world, and I am also afraid that it will create a worse future than the one that we had until now.
I think deployment should slow down and research should be financed and diversified.
The difference is that TVs and other past inventions weren't pushed on everyone at all cost. You could and still can live a mostly TV-free live. Workplaces aren't making you watch TV while you work. There is no one insinuation that you'll be obsolete if you don't embrace TVs. Your existing appliances didn't suddenly sprout a TV screen. Instead, people bought TVs because they saw value in them, at a time of their choosing.
I think the environmental aspect is interesting and worth discussing. Around the offices the common joke is that people will "just burn down a piece of the rainforest" when they fire up their AI to solve some complex problem. Which certainly isn't what the world needs right now, and you can't have the "tool" without also the massive water consumption in a world where not everyone has access to clean water. Though as the fatalism in the burning down the rainforest implies, people around here have sort of accepted that the world is going to get hot.
On the other hand. If we apply the same sort of fatalism to AI, then we can expect AI to lead to civil uprising and a world which will probably be a lot more sustainable once most of us are dead.
I don't think the automation is any different from what we've seen the past 150 years. Except that perhaps this time AI is the tool which is actually going to do to the office what the assembly line did to the factory.
Neither the power, nor cooling, requirements are necessarily environmentally harmful. As always ymmv.
First water; there are plenty of places on the planet where water is bountiful. If you site a data center well (say on a big lake in a rainfall area) then water is not an issue. An "using" water in this way doesn't affect water scarce areas.
Equally lots of electricity can be generated in environmentally harmless ways. Solar, wind, hydro are all clean. (Hydro depending on how and where.)
Yes, in the long run, it may make sense to put data centers outside the US. Norway for example has no problem cooling things down. And hydro is abundant.
Plus, water "consumption" is also variable. Water is used for cooling, but is not necessarily "lost". Its "used" in the sense of "made use of" but may also then be "used for something else".
Sure, it's possible, but we don't seem to live in a world where that is actually the case in practice.
The large AI companies are perfectly happy to draw water to the point that local residents no longer have access to clean tap water, and they are perfectly happy using dirty "temporary" gas generators for power.
We should judge it for its actual use, not its ideal use.
I think there's a larger problem of externalities (i.e. environmental costs) not being fairly represented in transactions. AI is just the latest and biggest symptom of this. If someone has a business process which involves generating a waste product (or utilising some resource) then that's only a problem if the cost isn't being paid by that business (and ultimately reflected in the price charged for the product/service).
Obviously that's not the whole problem - the other issue is the financial resources these companies have access to. Presumably if they wanted to, instead of pushing up the worlds prices for DRAM or electricity they could have distorted pretty much any market they wanted to given the money they have. If instead of datacentres they decided to buy the world's supply of coffee then presumably people would be paying $100 for a tin of coffee beans. You could introduce all sorts of restrictions and market controls but for me a better reform would be ensuring that they didn't have hundreds of billions of dollars to spend in the first place.
>Except that perhaps this time AI is the tool which is actually going to do to the office what the assembly line did to the factory.
I think this is probably right, but for the people who worked on assembly lines they (or more realistically their kids) could go find an office job to do instead. It's not clear what the kids of todays office workers are going to do for employment (if anything).
I wonder if this fatalism comes from a feeling that the enormous power of AI must somehow come with some sort of proportional cost. There isn't really any clearly known cost so maybe that's why people talk up the water and global warming and social impact fears and all that.
In comparison, does anybody say they're burning down a bit of rainforest when they go to the toilet (and flush it)? Or when they cook dinner at home with more waste heat than factory cooked food? Or throw away a pen that's run out of ink? The environmental impact of all the mundane things in life is far greater than using AI but the impact of each mundane thing is perhaps also far less impressive.
I think the fatalism comes from having done a lot in our personal lives to live more sustainable, voting for green politics (in Europe you can find those everywhere on the political spectrum). You'll even find a "if it's yellow let it mellow" sign in the bathroom of a lot of my coworkers homes. Yet seing little impact because of how we as a global society continue to burn the planet. I know my fatalism comes from this.
Am I not allowed to say that hammer is just a tool and it matters how you use it? Because it makes all classic sculpture art meaningless? No it doesn't.
The "car is just a tool" doesn't make oil waste problem any less problematic.
I am one of those people that say AI is just a tool and it matters how you use it. But also, GCC/Clang/Rust are just tools and it matters how you use them. Your mouth is a tool and it matters how you use it. When thinking about ANY tool you should think about the best way to use it.
But for AI specifically this saying is most important, because a lot of people treat AI as another person. "He's like a junior employee" is what I've heard in my company on meetings. It's not a he, it's not a junior, it's not an employee. It's a tool!
I see few people actively engaging with the thesis of the article so lemme restate it in another way:
We can agree that the invention of the car revolutionized transit, but when you optimize around cars it creates perverse incentives that warp society. American society for example is rife with these issues: Cities and towns that are fundamentally unwalkable, cars that are increasingly large and hazards to public safety and lower and lower expectations for drivers.
In that sense the AI craze is exactly the same. Even if you believe that AI will or has revolutionized coding (a take I firmly disagree with), the resulting craze is creating perverse incentives that are warping tech. Access to computing is becoming narrower and narrower as prices sky rocket, applications are becoming more unstable and unsecure by the day and more and more of our economy is reliant on AI gains that will likely not be realized. Not all tools reshape society for the better, much like how advertising was a tool for attempting to gain eyes on your product now it has become a dystopian weapon to sell your users as the product.
This is without getting into the actual issues AI has directly, just the way it's warping things around it.
The “just a tool” argument was doomed from the start. We discuss tools all the time and many are controversial. Meds, drugs, guns or gambling machines are only a few examples.
I’m not saying we should treat AI as one of the above; I’m saying that regulation is common and can be beneficial.
Note that commenters here generally praise regulators when eg smartphones are banned at schools.
Then again, AI is almost completely unregulated at the moment. Its expansion sometimes disregards laws. Defending use of AI in its current form is in opposition to regulating AI use.
It's still not obvious that AI is uniquely environmentally destructive compared to some other "non-essential" industries, and it gets even less obvious when that's weighed against its current and potential benefits. Where the datacenters are built could probably be better regulated but that issue isn't inherent to the technology.
I think all issues regarding intellectual property rights actually fall under automation, even if everyone could stop their works from being fed to AI, enough public domain/open source works will be used, and enough people will license/donate their work for AI training for us to eventually get to the same point, it would just take longer, everything would be ethically sourced and people whose jobs are getting automated would still be unhappy.
The only real problem is the consolidation of wealth and power, everything else feels like a cope or a distraction.
As I understand it, in America "it's just a tool" is shorthand for "guns should be regulated like hammers, i.e. barely at all, responsibility lies with the user not those who make, market and sell it"
Needless to say a lot of people disagree with that, because of the shootings. And internationally, a lot of societies regulate guns a great deal.
I'm not saying this to steer the conversation towards gun control, or to compare AI use to killing. Rather, I'm saying that the conversation here so far has has been confused because people have different ideas of what "it's just a tool" means, so they're talking at cross purposes.
To some it's an obvious statement of fact; almost everything man-made is a tool, AI is a tool like a pair of shoes or a guitar is a tool.
To others it's a statement that corporations should be held blameless, and access is a moral right that shoudn't be limited on consequentialist grounds.
> To others it's a statement that corporations should be held blameless, and access is a moral right that shoudn't be limited on consequentialist grounds.
The old, what is better for the rich VS what is better for society.
You’re not doing this but to your example: Guns don’t have numerous other positive use cases, so AI is simply not a fair comparison.
And in the case of AI, given the media environment the fact that it’s being used to save millions of hours doing back office automation is never covered.
So yes this is a prime argument for AI. It’s just a tool with a myriad of uses that no one covers cuz they’re boring.
As to the rest of the authors argument, obviously I’m aware of its myriad negative consequences but just disagree that banning matrix math or NIMBY-ying warehouses where matrix math happens is somehow a solution to those problems.
You want to talk about policy fixes at the level of applications, I’m all ears.
I think another important parallel between guns and AI is that the people who say "it's just a tool" are discounting all the cultural context. Additionally and anecdotally, they tend to be the people who are most connected to that cultural element.
A gun is just a tool except it's also a symbol of countless things within American culture: freedom and independence and even masculinity sometimes. It's bound up with various cultural tropes and narratives that mean it it's only a tool in the same way that Mjǫllnir is only a hammer.
Similarly, we have literally centuries of cultural context to the idea of thinking machines, and it is a tool, but a lot of the people saying that also think it's their girlfriend or god or just something that will entirely upturn society and remove [entire class of people].
"It's just a tool" is accurate but disingenuous, knowingly or not. There's a lot more bound up in a gun or AI than there is in a can opener.
I find it hard to see AI in a positive light when it is the root cause behind my friends not being able to afford gaming pcs and consoles, every thing is getting more expensive. I use it at work (because it is mandatory) and objectively speaking it has helped me, but the negative effects are hard to ignore.
But really, blaming AI (software) for this is a similar situation to "old man yells at cloud".
The software didn't want more RAM. AI companies did. Not to mention that RAM companies may have purposely raised prices and blamed it on AI, I've heard about a class-action lawsuit against the big three RAM companies for coordinating a price increase.
If you don't want to use AI at work, you should really file a complaint. Again, the software didn't do anything, someone else did something to force it on you.
I'm not saying this as something to anger people, but because hating AI has no effect on whether you make AI optional at work or stop big tech companies and RAM manufacturers from price increases. Software has no mind or consciousness, it has no sense of being evil and purposely doing these things. People do, and people do it on purpose. People made AI, after all. It didn't spawn into existence.
The root cause of not being able to afford PCs might just be RAM companies themselves, along with some real demand from AI companies.
69 comments
[ 3.6 ms ] story [ 36.3 ms ] threadIt does work. Cars still exist today. Society didn't decide to go back to horses.
>The entirety of all ethics involved in modern technological ecosystems and infrastructures rests solely on how a singular person chooses to use something?
No one claimed anything close to this statement.
>Tools, then, aren’t “neutral” in any way.
Tools may not be neutral. But tools are inevitable. They are a solver to people's problems.
> we can simply ask for art and it materializes before us. There is no struggle at all involved, thus the terrible labor of being an artist is removed!
Do not confuse the simplicity of early tools with there being no struggle to use something. The most efficient way to convey the exact vision of an artist is not purely a prompt.
> “insultingly naive”, “overly simplistic”, “immature and self absorbed”
..exactly what I would expect from someone partly or majority through a PhD about tools. I like tools, their design, creation, and deployment more than average (though probably less than the author), but you need to come up for air, man.
Your interpretation of that phrase is oddly specific, highly esoteric, and completely different from EVERYONE not doing a PhD on tools.
> AI is just a tool
That word “just” seems to be doing a lot of heavy lifting for you. This sentiment is not intended to downplay the importance of tools on society. The generally understood intent of this phrase is to quell the current hysterics and to reassure anyone unfamiliar with the term _back propagation_ that AI, in fact, is neither alive nor sentient (in the sci-fi sense), and that it is “just” statistical modeling. Do not mystify the technology.
> a car/hammer is just a tool
While these are unanimously considered tools, I could entertain a long discussion and the “particular specialness” of tools and things that are inherently dynamical systems.
> what prototyping _should_ be about
The huge majority of the rest of your rant (or what I read of it) is awfully presumptuous and weirdly confrontational. Who can say what prototyping _should_ be? You of many people should understand that the creation and use of tools is contextual, sentimental, highly personal, intimate even.
> “But artificial intelligence… intends us not just to sit forward
From the first part of “The phrase”… AI is inanimate. Humans DO tend to anthropomorphize, and the whole point of saying “it’s just tool” is to remind people that… inanimate things don’t have an intent!
> our tools are using us
I find this notion somewhat trite. “The tail wagging the dog”. I’m not arguing that we aren’t impacted by the tools we use nor that the use and proliferation of AI is not impacting society, but it takes a PhD level of mental gymnastics to push that concept as far as you have in your rant.
The important takeaway though, is that you are a human, with agency and autonomy.
> it only matters how you use it
The implication of “How you use it” is that YOU GET TO CHOOSE. what you think, how you think, “how you use it”™, and even whether you use it! The choice is yours.
I don’t myself have a PhD, but seriously, do yourself a favor and come up for air. Read the parable of “The Empty Boat”.
I was hoping the article would not label it purely good or bad, but 1) highlight that AI is not just a tool but a very powerful tool and 2) and therefore it very much matters how we all use it and how it uses us.
In this framing, we can see the things where it helps and hurts us and society and many levels at various intensities.
The article mostly just seemed to say how bad it was, and I don't think being critical of a new tool means to only see how it harms us but rather to see the potential full range of impacts.
I just think it's a very powerful and complex tool that can cause a huge range of feelings and outcomes.
In my opinion it asks us to not just look at individual gain of using LLM's (which can be powerful), but also place it in context how we as humans built long-term sustainable societies. Most societies are currently ignoring that part for the most part, because they fear missing out.
Probably the hardest problem of LLM's is their ethics. As the author put it:
> Modern multi-billion parameter AI models are scaffolded on and made possible by the largest heist in human history: theft of everything that could be scraped from every corner of the digital spaces we share. Without prevention of and justice for this damage caused by current models, their use is highly fraught, ethically. We, as human beings, have developed complex social forms of intelligence when it comes to dealing with things like credit and provenance, two things that modern models are incapable of.
No one currently seems to have serious solution for this. Solving it the right way would be choosing the hard path. While some artists or writers are fighting for their rights, us programmers are collectively ignoring the issue. It is funny how most companies (and some individuals) only want to use LLM's of they don't train on their code. But is just fine to use some 40 years of collective open-source without any compensation or even acknowledgement.
This side of the story doesn't make LLM's less technically impressive, nor does it make the tools themselves less powerful. But unlike the industrial revolution, which some like to compare the AI boom to, during the industrial revolution humans built stronger productivity solely on innovation without strip mining all intellectual property in human existence.
I can see the benefits of LLM's, enjoy the technical capabilities and be impressed by the technology, and at the same time find these tools wholly unethical in their current form.
> if it doesn't cause too much harm.
You touch something there: some of us think that simply using LLMs at all encourages developing / spreading this tech that inherently causes too much harm.
- developing all but the most complex software
- maintaining a >100k LOC codebase on your own
- generating illustrations and animations
My iOS codebase has 95k LOC + another 40k tests, and it is working great. It has been months of work.
If you find yourself wanting to write like this, I recommend reconsidering. Persuade them instead.
>I’ve been thinking constantly about the common and casual phrase I’ve heard so often, “AI is just a tool - it matters how you use it.” This has been the rallying cry of tech-loving academics who no longer do their own research, tech bros who salivate over generative images of criminal depictions of people without their consent, and business-minded folks who actually don’t care about AI but see this as an opportunity to rake in more and more money for themselves.
And sure, technology choices have plenty of second order effects. Stopping any analysis at just how you are directly using the tool is probably insufficient.
At the same time, i still think that is part of how we use a tool. Society's choices about a tool (or even the choice to ban a tool) is still a part of how we use the tool and not instrinsic to the tool. I feel it all comes back to how we decide to use it. We can use nuclear technologies to treat diseases, power our cities, or we can use it to bomb places. We can chose to regulate the waste appropriately or not. Etc.
When i say things like its just a tool and it matters how you use it, i am claiming it is not inherently good or evil. It can have good or evil (or both) effects depending on how individuals use it and how society regulates it.
I have yet to hear a compelling counter example.
This is what is being refuted. The argument is that there is, in fact, inherent badness in the tool, which doesn't quite depend on how the tool is used.
That’s picking at a particular level in the hierarchy of application that amounts to sleight of hand, even if unintentional.
If you said “a nuclear bomb”, the more specific instantiation focuses the purpose and likely impact in the world to a narrower, more purposeful goal.
Simply: nuclear technology might be used to treat diseases, but I am pretty certain nuclear bombs are not.
Is that a different tool or the same tool put to a different use? I would lean to calling that the same tool, since its the same process. I think at some point if we divide tools too finely we just end up encoding the use of the tool in the tool definition and the whole thing becomes moot. Like is a gun used for hunting a different tool then a gun used for murdering? Maybe you could say a machine gun is different, but what if its the same model of gun?
[As an aside, there are very real arguments over if nuclear bombs are a net negative for humanity. For all the downsides, mutually assured destruction significantly suppressed major wars]
What discipline, outside academia, could position "nothing is neutral" as insightful and helpful commentary?
So is doing nothing, consuming resources and goods created by others, and living in the default state of humanity: brutality, poverty, and early death.
If it sounds if I'm being glib in this statement, I am. This article is an unreasonable amount of words which eventually boil down to "we have choices in what we do": an exactly equivalent statement to "X is just a tool, it depends on how we use it." Which, of course, is the thesis he intended to rail against.
Say a tractor does the work of ten farmers, and the farm owner lays off those ten farmers. Is that the tractor's fault? It's just the farm owner's fault. But we usually say the tractor took the ten farmers' jobs, because the farm owner trusted the productivity gain and made the cuts. That's the real point.
I agree AI has that same dynamic, but I'd argue that people who write this kind of thing tend to come from the establishment. Just look at their background.
People treat "open source" as if it's inherently good. But honestly, they don't realize how open source can actually work in a pretty vicious way for non Anglophone countries. Do you know why? For people whose access to knowledge is limited by the English language, it's hard to sell software that offers less value than the open source stuff the Anglosphere gives away "for free." Now think about it. Can a developing country really produce that same kind of elite mental model?
In places that are ten or twenty years behind in technology, do you really think they can produce technology that surpasses Anglophone open source? At that point, learning English itself becomes a privilege, and that's where the technological asymmetry comes in. And the thing is, learning English in a developing country is expensive.
In that sense, AI actually has a more egalitarian side to it. That's why I find this kind of professorial binary thinking so naive. Whether using it is evil or good depends entirely on where you're standing. When a professor tells everyone else not to use it, it just looks like a "this is about my own livelihood" problem to me.
The entire culture of programming and IT is Anglophone. And to even get a proper understanding, you have to read academic papers. To really use Rust properly, you need to deeply understand polymorphism, starting from ad hoc and going from there. Do you know how much time it takes to really internalize all that? The starting line is different, so you can never catch up. In that sense, AI is an asymmetric tool.
It feels like people born on third base are saying, "What's so hard about getting to home plate?" I have no problem with people projecting their own emotional lines onto tools and creating echo chambers. But this logic that only human made work is inherently good, I just find that hard to understand.
This right here is the hypocrisy of the struggle that the establishment loves to celebrate. Some people can afford to stand on stage and fight the good fight, but others aren't allowed that luxury.
To the person writing that, technology might feel like oppression. But to someone else, it's liberation.
Honestly, what I find hard to understand about this piece is that it completely fails to recognize that what counts as a meaningful struggle is determined by your social and cultural position. Their worldview is just narrow. A person's institutional position shapes what they see as human labor and what they see as automatable labor. That's purely a matter of where you stand.
[1] If you take the slop and manually fix/improve/verify it to production value it can be immensely valuable.
The thought terminating issue with the phrase is that it isn't just a tool. Once you automate its use (automated PR reviews, ticket fixups, etc.), it becomes a process as well. You are almost certainly purchasing it as a service and not as a tool. It also contains elements of randomness that most tools do not. These quibbling points are what triggers the cognitive dissonance to produce the thought terminating effect. It is quibbling so you cannot really argue that point but the point is irrelevant; the second part of the cliche seems to be agreeing with the complaint about slop and yet the sentence is offered as some counterpoint. This termination prevents the actual conversation about how valuable its use is when the quality standard is not lowered.
Well ... too bad that blacksmiths didn't have access to this amazing tools before those laws and regulations were put in place. A hammer doesn't need all these thing. The iron law of bureaucracy needs them.
When I read articles like this - what I observe is mostly rage born from the fear that the core of their identity - aka being smart in the general vague non threatening and non offensive way is rapidly depreciating asset.
Barrage of uncorrelated arguments ranging from economics, power, climate.
When the argument is multi dimensional like this, it usually means the author is struggling and desperate to make a point.
I’m not sure what kind of person finds these interesting and invoking curiosity.
[Tools do not exist in the void but in a society]. You can't study tools without studying the context that created it. Some tools work in some society, some others don't. One example, IIRC, are Ski-Doo with some native population that have Potlatch-like practices. Some anthropologist gave them Ski-Doo but you have to sacrifice something precious to give back, and soon they had to be burnt. In our society, anything expensive that require maintenance and repair is a bad fit as both "don't have the time" and the skills. We prefer "disposable" objects at the cost we know.
[Tools create potentials, society may realize them]. As hackers, we often like to create new potentials (eg. with Bluesky ATProto, or the anonymous Vuvuzela chat, etc.). We also envisioned that with electricity, then with 3D printing, that people would build many objects of their daily life directly at home. But, from all the created technologies, some potentials are realized, some not. Community Memory, a San-Fransisco pre-Internet electronic board, wrote that more often the potentials that favor people in power in our society are realized. That's Palantir - strong - business model.
[Tools convey intentions]. I said above that tools create potentials. These potentials are not limited to what the tool actually is, but it should encompass all the micro-choices made by ones creating it. To illustrate this idea, take a knife. Knife designed to cook on one side, and knife designed for hunting in the other side, do not look the same. Youtube, by not displaying a big bar telling you how much you already uploaded to their server convey the idea of an infinite storage space, etc. Additionally to the raw utility of your tool, you can convey ideas, intentions, suggestion, affordance, either consciously or unconsciously.
[Our tools maximize efficiency at the expense of everything else]. Efficiency, in this case, could be defined as reaching a specific goal, in a stable & defined context. The critics of efficiency improvement are multiple: often, what really matter is to reach an acceptable threshold on many many different goals. Additionally, most efficiency gain make the process more and more dependent of a stable defined context, each micro-change in the environment and the whole tool / production chain is disrupted. It could be summarized as the opposition between "industry" and "craft". It's also one argument of the luddite: they also had tools to build cloth, that were less optimized in term of speed to build a cloth, and harder to master. But on other goals, that were not considered at that time, like quality, they were better. They thought the better quality of their cloth will save them, when people will find how poor quality are the cloths made by the new automated machine. But it did not happen. Still, from a French perspective, crafts still exist there; it's what luxury brand sells, because rich people know how much better they are. William Morris experimented and published a lot on these topics, it's an interesting reference.
With the 4 ideas (tools do not exist in the void, tools create only potentials, tools convey intentions, we build tools to maximize efficiency only), I think we have some basis to build a stimulating critic of both AI, cars, and knife. But this part is left as an exercise for the reader :-)
People interested in these topics may read Ivan Illich, André Gorz, or Jacques Ellul.
The invention of the telegraph changed how information is traded and the contents of newspapers, the invention of the TV changed politics, and so on.
The medium shapes the message and induces behaviours from us. We never thought about wanting to promt chatgpt, or watch people doing sports on a screen. But the possibility of doing these things makes us change.
LLMs obviously have and will continue shaping the world, and I am also afraid that it will create a worse future than the one that we had until now.
I think deployment should slow down and research should be financed and diversified.
That's the thing to keep in mind, it's not something an individual can "fix".
That's why policies and laws are so important.
On the other hand. If we apply the same sort of fatalism to AI, then we can expect AI to lead to civil uprising and a world which will probably be a lot more sustainable once most of us are dead.
I don't think the automation is any different from what we've seen the past 150 years. Except that perhaps this time AI is the tool which is actually going to do to the office what the assembly line did to the factory.
First water; there are plenty of places on the planet where water is bountiful. If you site a data center well (say on a big lake in a rainfall area) then water is not an issue. An "using" water in this way doesn't affect water scarce areas.
Equally lots of electricity can be generated in environmentally harmless ways. Solar, wind, hydro are all clean. (Hydro depending on how and where.)
Yes, in the long run, it may make sense to put data centers outside the US. Norway for example has no problem cooling things down. And hydro is abundant.
Plus, water "consumption" is also variable. Water is used for cooling, but is not necessarily "lost". Its "used" in the sense of "made use of" but may also then be "used for something else".
The large AI companies are perfectly happy to draw water to the point that local residents no longer have access to clean tap water, and they are perfectly happy using dirty "temporary" gas generators for power.
We should judge it for its actual use, not its ideal use.
Once AI becomes a commodity this is the sort of thing that will become a differentiating factor when consumers choose one over the other.
Obviously that's not the whole problem - the other issue is the financial resources these companies have access to. Presumably if they wanted to, instead of pushing up the worlds prices for DRAM or electricity they could have distorted pretty much any market they wanted to given the money they have. If instead of datacentres they decided to buy the world's supply of coffee then presumably people would be paying $100 for a tin of coffee beans. You could introduce all sorts of restrictions and market controls but for me a better reform would be ensuring that they didn't have hundreds of billions of dollars to spend in the first place.
>Except that perhaps this time AI is the tool which is actually going to do to the office what the assembly line did to the factory.
I think this is probably right, but for the people who worked on assembly lines they (or more realistically their kids) could go find an office job to do instead. It's not clear what the kids of todays office workers are going to do for employment (if anything).
In comparison, does anybody say they're burning down a bit of rainforest when they go to the toilet (and flush it)? Or when they cook dinner at home with more waste heat than factory cooked food? Or throw away a pen that's run out of ink? The environmental impact of all the mundane things in life is far greater than using AI but the impact of each mundane thing is perhaps also far less impressive.
Am I not allowed to say that hammer is just a tool and it matters how you use it? Because it makes all classic sculpture art meaningless? No it doesn't.
The "car is just a tool" doesn't make oil waste problem any less problematic.
I am one of those people that say AI is just a tool and it matters how you use it. But also, GCC/Clang/Rust are just tools and it matters how you use them. Your mouth is a tool and it matters how you use it. When thinking about ANY tool you should think about the best way to use it.
But for AI specifically this saying is most important, because a lot of people treat AI as another person. "He's like a junior employee" is what I've heard in my company on meetings. It's not a he, it's not a junior, it's not an employee. It's a tool!
We can agree that the invention of the car revolutionized transit, but when you optimize around cars it creates perverse incentives that warp society. American society for example is rife with these issues: Cities and towns that are fundamentally unwalkable, cars that are increasingly large and hazards to public safety and lower and lower expectations for drivers.
In that sense the AI craze is exactly the same. Even if you believe that AI will or has revolutionized coding (a take I firmly disagree with), the resulting craze is creating perverse incentives that are warping tech. Access to computing is becoming narrower and narrower as prices sky rocket, applications are becoming more unstable and unsecure by the day and more and more of our economy is reliant on AI gains that will likely not be realized. Not all tools reshape society for the better, much like how advertising was a tool for attempting to gain eyes on your product now it has become a dystopian weapon to sell your users as the product.
This is without getting into the actual issues AI has directly, just the way it's warping things around it.
I’m not saying we should treat AI as one of the above; I’m saying that regulation is common and can be beneficial.
Note that commenters here generally praise regulators when eg smartphones are banned at schools.
Then again, AI is almost completely unregulated at the moment. Its expansion sometimes disregards laws. Defending use of AI in its current form is in opposition to regulating AI use.
I think all issues regarding intellectual property rights actually fall under automation, even if everyone could stop their works from being fed to AI, enough public domain/open source works will be used, and enough people will license/donate their work for AI training for us to eventually get to the same point, it would just take longer, everything would be ethically sourced and people whose jobs are getting automated would still be unhappy.
The only real problem is the consolidation of wealth and power, everything else feels like a cope or a distraction.
Needless to say a lot of people disagree with that, because of the shootings. And internationally, a lot of societies regulate guns a great deal.
I'm not saying this to steer the conversation towards gun control, or to compare AI use to killing. Rather, I'm saying that the conversation here so far has has been confused because people have different ideas of what "it's just a tool" means, so they're talking at cross purposes.
To some it's an obvious statement of fact; almost everything man-made is a tool, AI is a tool like a pair of shoes or a guitar is a tool.
To others it's a statement that corporations should be held blameless, and access is a moral right that shoudn't be limited on consequentialist grounds.
The old, what is better for the rich VS what is better for society.
And in the case of AI, given the media environment the fact that it’s being used to save millions of hours doing back office automation is never covered.
So yes this is a prime argument for AI. It’s just a tool with a myriad of uses that no one covers cuz they’re boring.
As to the rest of the authors argument, obviously I’m aware of its myriad negative consequences but just disagree that banning matrix math or NIMBY-ying warehouses where matrix math happens is somehow a solution to those problems.
You want to talk about policy fixes at the level of applications, I’m all ears.
A gun is just a tool except it's also a symbol of countless things within American culture: freedom and independence and even masculinity sometimes. It's bound up with various cultural tropes and narratives that mean it it's only a tool in the same way that Mjǫllnir is only a hammer.
Similarly, we have literally centuries of cultural context to the idea of thinking machines, and it is a tool, but a lot of the people saying that also think it's their girlfriend or god or just something that will entirely upturn society and remove [entire class of people].
"It's just a tool" is accurate but disingenuous, knowingly or not. There's a lot more bound up in a gun or AI than there is in a can opener.
The software didn't want more RAM. AI companies did. Not to mention that RAM companies may have purposely raised prices and blamed it on AI, I've heard about a class-action lawsuit against the big three RAM companies for coordinating a price increase.
If you don't want to use AI at work, you should really file a complaint. Again, the software didn't do anything, someone else did something to force it on you.
I'm not saying this as something to anger people, but because hating AI has no effect on whether you make AI optional at work or stop big tech companies and RAM manufacturers from price increases. Software has no mind or consciousness, it has no sense of being evil and purposely doing these things. People do, and people do it on purpose. People made AI, after all. It didn't spawn into existence.
The root cause of not being able to afford PCs might just be RAM companies themselves, along with some real demand from AI companies.