Jamie Wheal, author of Stealing Fire and Recapture the Rapture, speaks in a clip from Pinchbeck's Breaking the AI Barrier course on whether human capacities can survive the current wave of technology. He builds the talk on Tim Wu's The Master Switch, which traces how information technologies from the telegraph and radio to the internet begin decentralized and utopian and end hegemonic and centrally controlled. Wheal extends the pattern to psychedelics, which he treats as an information technology, and says it is why he was skeptical of the psychedelic renaissance years before others were. He maps the AI debate into three camps and argues that the fixation on superintelligence crowds out nearer-term harms, pointing to Cambridge Analytica as an early and crude case.
Guest: Jamie Wheal
In brief
- Wheal argues, following Tim Wu's The Master Switch, that every information technology starts open and decentralized and ends consolidated under central control, so utopian rhetoric about AI should be treated with suspicion.
- He extends the same historical pattern to psychedelics, which he classifies as an information technology, and says it grounded his early skepticism about the psychedelic renaissance.
- Wheal sorts the AI discourse into three positions: boosters such as Sam Altman and Marc Andreessen, safety-focused figures such as Ilya Sutskever, and those who dismiss the systems as hype.
- He argues that framing AI only as savior or exterminator flattens the conversation and removes the mid-level harms that most need discussion.
- Wheal notes that promises of technology-created leisure have recurred since early industrialization and the postwar consumer boom without materializing.
- He cites Cambridge Analytica's algorithmically workshopped 'crooked Hillary' framing as evidence that even crude data-driven manipulation can shift an election, with deepfakes posing a larger threat to shared falsifiable reality.
Full text
that if there were a theme that was sort of a thread we could tug on, it would be something along the lines of maintaining, preserving, enhancing our humanity in the midst of this technology. Um, and so if it's okay with you guys, I'll just kind of uh share a few frames that are hopefully helpful. I mean, they're just at least some of the ones that I've been noodling on. Um, and one of the first is that I would really recommend if you haven't come across uh Tim Woo's book, The Master Switch. He is a professor of law um at either Columbia or NYU. I sometimes get that confused.
I think it's NYU. Um but his book, The Master Switch, I think is precient. And as far as like if I judge the value of a book by how often it comes to mind and the model I find is relevant and applicable, his is like top three for me in the last decade because in a nutshell and I and I actually I reference it and use it in the book a book I wrote called Stealing Fire which ironically enough ends with Prometheus and Epimetheus and Pandora. Um but but the idea is that all information technologies follow the same arc from telegraphs and radio to initial television all the way through the internet etc. And I extended the argument I think psychedelics are an information technology so we should expect them to behave in a similar way.
I think we could also plus one AI without a doubt but he says they all start out utopian and decentralized and they all end up hegemonic and centrally controlled. And so the early days of radio would be like everyone around the world with a little magic box can hear things and be informed and be inspired and democracy will rule and you know all the things. The same with telegraphs and radio, you know, all of all of the bits and pieces. So his historical take gets us off the hot takes and gets us off our current hype cycles, you know, but pro and con. And then just you you just see the pattern and being able to see that pattern for instance, you know, position with many other reasons, but position me on the skeptical side of our current psychedelic renaissance as early as 2016, 2017, 2018.
I'm like, "Guys, this is not going to go the way you think it's going to go and here's why." And people were like, "Oh, boo, you make me sad. I'd rather trip more." And you know, like twinkly lights and and and sure enough, you know, 2025, everyone's like, "You were right." And it's like, I wasn't right. This the pattern is obvious, right? So, I would really encourage considering I mean, you could probably ironically plug this into Claude or Chat GPG and get yourself a synopsis of Tim Woo's work at this point um to tell you exactly how we are. But I would say be aware of that and then be leerary of lip service to the utopian and decentralized versions.
They tend not to last long if and if they were ever sincere in the first place. So, that would be step one. Um, step two would be um, well gosh, I mean we can go happiest to saddest, we can do anything we want. Um, I would say let's work backwards from the kind of the the practical and tactical, right? So there is some description Sam Alman's in that world, lots of, you know, Mark Andre, various other go- go Silicon Valley boosters who are sort of like it will be so amazing. It will it will remove us of our drudgery. it will do all the great things and we will have more leisure time etc et we've been seeing that pitch that shtick since early industrialization including the 1950s and like look look what women will have all the time for now that we have microwaves and dishwashers and all the things and you know and and of course that free time or even CDs might cost $20 each now but they're so cheap to make and they're so much cheaper than vinyl and records they'll be we'll be selling them for $3 each soon and of course that never happened right there's all of these promises of that then sort end up unspooling.
And and when we hear the pros and cons, like is AI going to save us? Is it going to kill us? Or is it overhyped and it's just stocastic parrots predicting words and and all of this is So you almost think of there's sort of three legs to this stool, right? You have the pro- AGI hypers and that's that's everybody from the altness to the Andre and of course like any Ponzi scheme or legit thing they have to do that because that's what justifies crazy valuations insane investment schedules all that kind of stuff. This is a winner take all and we are delivering a silicon gone.
Conversely, you have the Ilio Sutskyvers and the various other um you know the sort of they believe it's that powerful also, but they believe it's potentially for ill and we need to put the brakes on it. So you have they sort of agree of how powerful and worldchanging and and disruptive this is going to be. They just think one is going to one believes it's going to save us most likely. One is fairly certain their P doom, right? Is it's going to wipe us out. And then you have the kind of third leg of the stool that are like this is hype and hot air and it's not ever going to do the thing that everybody says it's going to do, right?
And so you're kind of like I don't know about you guys, but it can be kind of schizophrenic depending on who you listen to last or who seems most confident on Twitter or on a podcast or whatever and you're like I don't know whether to or go blind, right? So just to sort of map that and be like that that exists, right? And then to say, okay, um, if we zoom in on that, um, what are some helpful frames? Because when this is so new and so novel that most of us are grasping at straws or to what's the handiest analogy, right? And people talk about it's like the invention of the printing press but more.
It's like this, it's like that. And quite often we go to the Matrix, we go to the Terminator, we go to Skynet, we go we we go to our sci-fi analogies. Um, and we often think, is it going to wake up, be omnicient, and be intentionally or unintentionally malevolent to the goals and agendas of us. That's the whole paperclip maximizer thing. And, you know, will it just accidentally kill us or will it purposely kill us like HAL 2000 in 2001, right? And I think that um there's several things we can discuss here, but I would propose that either it's I mean it's supernaturally intelligent and it's going to kill us or save us almost sort of flattens our conversation and and removes some mid-level and lower level stuff that I think we absolutely should be talking about because you realize that even the that dramatic an argument behooves and benefits the folks trying to raise money right now, right?
and and they are stoking those fires and they are hyping that. But if you just back it off, you're like, "Hey, let's just assume it stays exactly at its level of competence that it has right now." So, no step functions, just it is what it is right now, more or less. We've already seen it hollow out education, right? We've seen it hollow out the entire first two or three rungs of employment, you know, with Amazon and Salesforce bragging about, you know, 30 to 50% of their work is already done by by AI coders, etc., etc. And of course, you only need cedar senior coders to then inspect the buggy code, but it's all weird blackbox spaghetti code because they just write it in some very strange and non-intuitive ways.
So, how do you become a senior developer after there are no more junior developers coming up the ranks, right? There's there's all sorts of problems with it, but let's just say it is what it is. Um, I am I don't mean to say I'm less concerned isn't isn't an accurate sentence. I just have no idea about 3 years from now, 5 years from now. But what I can say is it's plenty wobbly and dangerous now. Because in the same way that Tim Woos like look at radios and look at telegraphs and look at the march of this pattern. You can just go back to 2016 and look at Cambridge Analytica, right?
Cambridge Analytica was a super weak but still highly effective algorithmic big data quasi AIish approach to hacking politics and the bo the body politic in our conversation and if anybody knows like why didn't you vote for Hillary well I don't know she just seemed a little shifty she just seemed a little I mean I just didn't trust her right I mean those Clintons that da da da da da and then oh push push oh crooked Hillary she was kind of crooked right I mean those what are those emails right you realize that crooked Hillary was algorithmically workshopped by Cambridge Analytica and then pushed out through all the conservative threads. It wasn't a it wasn't in a sooie generous. It wasn't an original cultural meme. It was algorithmically optimized, weaponized, pushed out, right? And tipped that election.
And if that election hadn't gone the way it did, we would not be exactly where we are right now. So you're like, hell. So that was like AI light and it had that kind of insane knock-on effect. So just picture now we're on the cusp of deep fakes and lack of truthiness etc etc and a fundamental lack of knowability and mutuality and you know and again both visually and with video form as well as infinite text as well as AI slop as well as all of this stuff. Do we just become dissociated from shared actual falsifiable realities? And do we end up killing each other or killing ourselves?
Never mind whether there's a mustache twizzling how 2000 making a bunch of paper clips.
Automatic transcript, lightly edited for readability. Originally published on the Liminal News channel.