Intro
Conspiracy theorists and scientists have a really key thing in common: They both focus on questioning the world around them. The key difference is the supposed to be the philosophical grounding. Scientists are supposed to accept the key assumptions of Occam’s razor [1] and provide support to the academic institutions.
Right now, at least in the field of computer science, Scientists have abandoned upholding academic institutions. Completely and Utterly.
In my field of systems we’ve produced for the world the technical designs for everything from the smallest chip used in tiny GPS trackers to the scale of the internet. We know how to build these things and measure their impact. We know how to reshape all the technology that people interact with on a daily basis. We know how to do all this and measure it.
It might be weird that I’m seemingly conflating the people in CS Research and Industry, but that is the crux of the problem… Industry has gotten a LOT of people who do want to do Scientific Research, because there are genuine problems that require Research to solve.
These are also the same people creating the systems that actually run your ChatGPT or Claude query. They map your request through the internet to a specific computer somewhere figure out how to run this ‘‘LLM’” thing and respond with the output. All of that happens in a matter of milliseconds so it looks like the model is doing something constantly. So now ask yourself if there are all of these smart people working on this surely they must be paying attention to the massive environmental cost. Surely, all these smart people wouldn’t just leave us high and dry to die?
Spoiler alert history tells us that they absolutely will, not that they won’t try to fix it. Anthropic is the latest in attempts to make what they view as inevitable palatable. The recent response to the Superbowl commercial about ads is just the latest in enshitification. ‘‘We won’t ever add ads’’, seems to be a hollow and empty promise. The only system I trusted not to have ads was the governments, until a naked Fascist destroyed the separation that exists between politics and citizen support in the government.
On that topic the initial part of this post has been slightly disingenuous. There are actually a really good group of CS systems folks that do care about improving the daily lives of people. However, these people generally don’t work at any company you have ever heard of, or even the ones popping off in the Fortune 500 space. I’m taking about national laboratories. This is where the job of improving people’s lives by filling out the infrastructure of government actually happens. You want COVID vaccine data secure? National Labs. You want your electricity and your power cheaper, faster, and more efficiently? National Labs. You want nukes not to blow us all to kingdom come? National Labs.
So after this brief aside, reader, you might be wondering what exactly does upholding academic institutions mean? In a philosophical sense it means when scientists working on the cutting edge of a field observe that certain work could have negative impacts, the community as a consensus whole should come out and condemn the people engaging in such behavior. Condemnation serves as a warning to people who will consume and use that work. It also serves as a clear indication of what the best available scientific knowledge has to offer discussions on politics and practical regulation for the benefit of individual people. An example of this effort can be found in our sister field, Security. The security community sent an open letter about the fact that the elections were secure, to combat the fascist-in-chiefs ‘‘Big Lie’’.
Concretely, upholding the academic institutions means coming to terms with the fact that our choices might kill all of you. Because those same choices will pay for our lifestyle improvements while the rest of society discombobulates and people die. Eventually, the work we do will come back to bite us, but there are a lot of non-systems people that’ll have to die before we feel it. At the same time, every professor in my field already knows this and yet they will still submit or work on countless papers that speed up the destruction of society. I am not exempt from this critique, in order to graduate I will do the same. We know we are all gonna choke on carbon. We know that human contact will slowly be replaced by LLMs and everything can be monitored all the time. But where ethics should have us actually do something, we are content to cheer-lead the destruction. As we have been for nearly 2 decades, before the incidence of LLM popularity.
It’s key to understand, this is not a new problem. LLMs are just the new flavor of it. This rant isn’t anti LLM. They can genuinely be used to help people. I use them periodically to automate tasks that I can’t write tools for, and to help me understand certain code patterns I’m not familiar with, but they aren’t a panacea to everything. They are tools, some of the most useful in history perhaps. But at the end of the day a tool should be only as good or bad as its user. Today using any of the LLM services people are most familiar with is, very bad for the environment and for the users mental health, and countless other things.
How To Solve this Problem
Brief Commentary on Dumb Blogs
So I absolutely hate blog posts or articles that just say here is a problem, idk how to solve it, whoops. I also hate when journalists do that, and that’s probably because I don’t get the point of news that doesn’t explain how I should change a bad thing that is happening. So I’m going to tell you briefly what we can do as a society. Then I’m going to tell you what you can do, if you want to make up for this error in judgment that my entire field is collectively having. What I am not doing is telling you that you should do these things. Live your life how you want. But if you do want to have an impact on this problem, this is a prescription that you can follow to help. I take pride in being able to break down problems into digestible pieces, so if these steps don’t feel like something you could do on a day where you don’t have much to do, let me know in the comments (once I have them up).