AI is a...noisy topic. People I am naturally allied with on essentially everything have succumbed to a madness. I need to rewatch Neverending Story, but the situation is very much like The Nothing.
The Nothing is a supernatural force that appears in the novel The Neverending Story and its adaptations. The Nothing is described as a formless entity that is composed of real-world human negative emotions, such as disillusionment, disbelief, and humans' lack of imagination. As a result, the entity negates the existence of things and people in Fantasia. While the novel describes it as a formless being, the film adaptation depicts it as a raging storm, but the filmmakers explained that the Nothing is actually an invisible force, and the storm is a consequence of its presence, like the mist that surrounds the Nothing in the book.
I hope and expect this will cool off, the linear algebra-based thinking simulators we call LLMs will become normal technology, and most people will cough and shuffle when you ask them about some of the things they said back then.
I've been the person going mad back when I was younger and less thoughtful. I get it. It's not a terminal situation: it comes and goes with each beat of history.
I don't have the patience for tackling the misinformation running rampant right now, and more thoughtful and patient people than me are doing that anyway. I've tried over the last year or two and it just leaves me exhausted.
My alternative is to go a level up and share my toolkits for thinking about stuff, because it's why I so often find myself at odds with otherwise natural allies these days. No amount of self-checking with those tools has led me to any conclusion other than: no, actually, the world has gone mad:
I've seen people describe the last few years as a boycott, like I'm crossing some picket line by experimenting with LLMs. There's no real boycott. Just people yelling past each other.
I would have joined a real boycott on the merits even if I think many of the anti- arguments are misinformed. Give me a specific target, achievable demands, and a strategy informed by the actual history of labor movements, and I'm there.
My own early attempt at starting a career was cut off by automation, so I get it. No one was organizing aspiring IT professionals against the move away from on-premises hosting to the cloud. You don't see me yelling at you for using Gmail or a VPS instead of self-hosting. There's a lot to like about the cloud, as there is with generative AI. Jobs were moving in to computers in datacenters long before Google published the transformer paper in 2017.
What we have now is fear directed like an energy weapon at people who have nuanced views on technology. Your fear doesn't become activism just because it's loud and angry. Aimless fear with no end in sight that ignores its own ineffectiveness. Useless. This is matrices and manifolds, not Target or Harry Potter. It's not going anywhere, but we could still change its course as it winds its way into our lives.
Which is to say: Twitter-style microblogging and social media sites with algorithms aligned against human interest have driven popular public discussion, starting around when "let's go to Twitter" went from something people mocked Don Lemon for saying on CNN to just the way it is.
So much of how we perceive the world, now and for a while, is mediated through people whose perception of the world is filtered through these platforms and the aforementioned madness. It's ambient now, in the same way online transitioned from a place you went to the default state. No V.90 handshake announces in shrill triumph that you've jacked in to the future. It's no good to say you aren't on these platforms. The people who make the decisions that affect you are, or are influenced by people who are.
We're inside the second decade of the madness. You could start the timeline with GamerGate and what agitators learned from it and carried into other misinformation campaigns. Or you could peg it to the revelations around the so-called Internet Research Agency which, if I were of a mind to create conspiracy theories for fun, was behind GamerGate, and Adam Baldwin is secretly a Russian agent. But we don't insinuate such things here, not in the context of thinking critically. Given two possibilities, Adam Baldwin as Russian asset and Adam Baldwin as gullible and irresponsible, the latter survives Occam's razor even in a world with actual proven Russian assets manipulating public discourse.
The 2016 election happening around the same time Twitter turned on the algorithmic timeline for everyone made what should have been a transient vibe into a persistent vibe, and it's taken me a while to fully accept that. Every attempt to catch old Twitter's vibe, like Mastodon and Bluesky, has succumbed to the madness as people migrate with the same ill behavior and produce the same environment even without a misaligned algorithm. There are better forums I've neglected where pro- and anti- sides of any topic are held to a higher standard, or where communities have continued to develop after people abandoned them for Twitter and others, so I'll regroup there for what I tried to get out of Twitter-style microblogging.
The habits of thinking I developed in part from reading about Dr. Sagan's baloney detection kit in The Demon-Haunted World: Science as a Candle in the Dark have served me well over the years. I think everyone should develop a toolkit for thinking and navigating the world, the same way you should develop an actual toolkit.
It's okay to be wrong, but it's not okay to keep being wrong once there's enough stuff in retrospect to know you were wrong and update. Which is to say if you come to me with an unfounded or over-fit argument or article about AI, I will politely hand it back and walk off shaking my head. Because it's not fair to hand me such a lopsided situation where I would need to spend 10x or more effort addressing all the ways it's wrong before trying to salvage a point from it to discuss.
It's exactly like handing me some text from an LLM you asked about something. You couldn't be bothered to think critically about it and form your own opinion, and now you want me to deal with it? No. Go back and build a better toolkit.
Ed Zitron is straight out. I'll hedge a little on most of the most obnoxious anti-AI personalities, but you can't be so wrong so often for such a long time and get another shot without an unmistakable change in epistemology.
A short list of better people to follow:
Cory Doctorow – I don't always agree with his exact framing of AI issues, but that's a normal and healthy thing in this kind of situation. We can at least agree there's a difference between a machine made of linear algebra and Sam Altman while sharing a strong mutual dislike of Sam Altman. Read The Reverse Centaur’s Guide to Criticizing AI.
And I'll never forget when one writer turned to me and said, "You know, you prompt an LLM exactly the same way an exec gives shitty notes to a writers' room. You know: 'Make me ET, except it's about a dog, and put a love interest in there, and a car chase in the second act.' The difference is, you say that to a writers' room and they all make fun of you and call you a fucking idiot suit. But you say it to an LLM and it will cheerfully shit out a terrible script that conforms exactly to that spec (you know, Air Bud)."
These companies are desperate to use AI to displace workers. When Getty Images sues AI companies, it's not representing the interests of photographers. Getty hates paying photographers! Getty just wants to get paid for the training run, and they want the resulting AI model to have guardrails, so it will refuse to create images that compete with Getty's images for anyone except Getty. But Getty will absolutely use its models to bankrupt as many photographers as it possibly can.
A new copyright to train models won't get us a world where models aren't used to destroy artists, it'll just get us a world where the standard contracts of the handful of companies that control all creative labor markets are updated to require us to hand over those new training rights to those companies. Demanding a new copyright just makes you a useful idiot for your boss, a human shield they can brandish in policy fights, a tissue-thin pretense of "won't someone think of the hungry artists?"
The key difference between Cory Doctorow and the average public AI skeptic is I don't read in frustration making little notes on every paragraph, look at it, and give up any notion of responding. He has a difference of opinion on certain arguable things, not a fundamental difference in way of thinking or inclination to actually think that leads to a separate and incompatible reality.
Simon Willison – The person to follow if you want to keep up on what LLMs are actually capable of and what they can't do yet. His annual reviews are good reading if you want to get current.
Andy Masley – The person in this list hardline anti-AI people are most likely to foam at the mouth about, but who is most useful as a central anchor between different views. Long before AI was a going concern, smart people were warning about a rise of ecofascism in the coming years as climate change becomes tangible for more people. Those years have come, and Andy's defense of liberalism is a good read if you think it's a good idea to keep space for different views even at the cost of advancing your own in the short term.