Home
Creators
Search Recent Random
Posts
Search DMs Popular Hash Lookup Tags Random
Importer
Import FAQ
Register Favorites Keys Review DMs Logout
Community
Telegram ThePornDude
Home Creators Posts Import Register Favorites Logout

The Moral Dilemmas of AI with Michael Inzlicht (Patreon)

Published:
2026-03-27 06:58:16
Edited:
2026-03-28 04:41:08
Imported:
2026-03
Tags:

Downloads

Content

Mickey is BACK with two new papers and a very unwelcome idea: maybe the problem with AI isn't that it doesn't work but that it works too well.

We discuss all things AI and two new papers he has recently co-authored. The first, Against Frictionless AI, argues that AI assistance and its ability to take away the effort from thinking, writing, and social life and smooth out the process, could be robbing those activities of the very thing that makes them worthwhile.

After that, we look at another new empirical paper that presents a bunch of studies on the moralisation of AI. Some findings suggest that opposition to AI among some people isn't really about risks or trade-offs but rather about non-negotiable sacred moral values. Who knew?

We also discuss effort justification, reproducible research, AI partners, and just how much humanity remains within the cyborg Matthew Browne.

And remember... It's just like our opinion, man!

Links

Decoding Academia 34: Empathetic AIs? (Patreon Series)

Ovsyannikova, D., Oldemburgo de Mello, V., & Inzlicht, M. (2025). Third-party evaluators perceive AI as more compassionate than expert humans. Communications Psychology, 3, Article 4.

Zohar, E., Bloom, P., & Inzlicht, M. (2026). Against frictionless AI. Communications Psychology, 4, Article 39.

Oldemburgo de Mello, V., Côté, É., Ayad, R., Inbar, Y., Plaks, J., & Inzlicht, M. (2026). The moralization of artificial intelligence (Manuscript under review).

Paul Bloom's Small Potatoes Substack- My friend thinks it's a good idea for us to spend most of our time with AI companions- Is it irresponsible for academics to refuse to use AI?

Mickey's Speak Now Regret Later Substack- AI Alarmism Trades on Fear, Not Facts

Mickey's provocative tweet on his recent paper

DTG Previous Interview with Mickey on the Replication Crisis, Mindfulness, and Responsible Heterodoxy

Andy Masley Substack debunking the AI water usage claims

Andy Masley Substack: Using ChatGPT is not bad for the environment - a cheat sheet

Pataranutaporn, P., Karny, S., Archiwaranguprok, C., Albrecht, C., Liu, A. R., & Maes, P. (2025). “My Boyfriend is AI”: A computational analysis of human–AI companionship in Reddit’s AI community. arXiv.

Critical response to Mickey's paper from Roy Schulam on Substack

Comments

Dank Sparrow

Gemini tells me that autonomous vehicles will take millions of jobs (taxi driver/truckers) in 20 or so years. Read up on the actual historical Luddites and fear and anger for loss is well warranted. Obviously technology is neutral but looking at those who own the tech companies is not really encouraging. That said AI really brought Google searches to the next level.

Nancy Hale

This was a fascinating conversation. Mickey nailed my thoughts at the wrap up. The technology itself (like all technologies) is neutral. But like all technologies it can enable and magnify both the best and worst of human tendencies. Which is where morality comes in.

Nancy Hale

Mickey’s research on mental effort is really interesting. Humans get so much satisfaction working out puzzles or solving mysteries.

YellowDreams

This was good. These days I usually skip 95% of AI content since most people talking about it have no idea what they are talking about and it comes across so cringe. It was good to hear some of the more negative sides or personal worries from you folks because most of the off hand comments I hear on the pod come across as very booster but I know that’s probably just a reaction to the anti AI crowd. As one of these software engineers using these tools every day I feel quite conflicted. It’s much faster and easier and tbh I’m more of a systems engineer who has never been the best at pure coding and algorithms so I think I’m well positioned for the change up. But also as a musician I don’t agree that the only reason artists can be negative is due to some belief in the ineffable. That’s part of it. The other part is that the tools only work due to human input and much of the art and writing has been stolen and if not stolen the artists get little to no money in it. I truly do think that if we flooded the zone with ai generated stuff so much that it was even harder than it already is to make a living as an artist then artistic progress or novelty will suffer and we will lose something. I implore you not to just write it off as magical thinking because i think there really is something to the fact that LLMs require training data to produce the “near indistinguishable “ art. Without the market for human made art that becomes harder to come by. Very different to software or science in that respect which as long as we can do more and higher quality shit we will have a job. Artists on the other hard who produce the art themselves and don’t just prompt it may not.

YellowDreams

The other part I’m kind of watching with bated breath is the financial aspect of this whole industry. While inference is getting cheaper the amount of tokens used by the newer reasoning models is actually increasing and costs are getting worse as I understand. This lines up with what I’m hearing inside my company with regards to costs. I feel we are about 12 months away from finding out if the insane amount of money pumping into the industry is sustainable and if when we play the true cost of these tools we will still be so excited about it. Just for magnitude and comparison uber lost 37 billion total burn before being profitable. Amazon web services was similar. Open ai lost that much in 2025 alone. Anthropic lost 5bil on revenue of 10b. These businesses are not standard tech businesses that scale like traditional SaaS. And coming from insider knowledge the hyper scalers know this and are worried. It’s possible these tools are super useful but way too expensive and the economic cost of that will be very bad.

Mark Vaughan

I think they are ridiculously overhyped, I'm not afraid of being replaced. There is nothing doing any thinking, reasoning, creating, feeling. They are just massive correlation machines providing more accurate predictive text. These models are complex, but nowhere near as complex as the human brain. Despite having all the data in the world. They then still require feedback from humans to be more accurate. You don't need to use them much to see how quickly they make mistakes even on basic tasks. All this money and hype going into them, yet they have yet to add any value to workplaces. Even though they are replacing workers. AGI won't be a thing in any of our lifetimes, we are so ridiculously far off that.

Mark Vaughan

I thought Chris' bouldering example would go in a different direction. Yes, the progress and the process is the point. When it comes to the tedious process of writing and expressing yourself, the challenge and the overcoming of the frustration is the reward. When it comes to art the feeling that went into creating it is part of it, that's what connects us to each other. Obviously the end product isn't the same as if an AI created it and a human created it. That's why art connected to certain people is valued higher. Art, creative writing etc isn't a means to an end, the process of it connects it to the process of our lives. It's not all about consuming products and being more productive. AI creations involve no creativity, feelings or inspiration so they don't provide any of that. It is stolen data reproduced to emulate what we value instead of providing any new value itself. A lot of the joy we find in the things we value is the struggle and process, the human overcoming and the honest expression is the result. If AI does all the learning/easy parts of the process I don't see how someone learns the struggle and reward with more difficult things. Anything you offload to AI is something you can't take any pride in, something you can't feel the reward for. That's why it's so disastrous for our ability to learn and think critically. Alot of our gains are found in the struggles with easy things, and we need that to build and scaffold more difficult things. In the bouldering analogy, would he still participate in it if he could flick a switch and suddenly be able to do everything without any effort? That's the apt comparison, it's not about comparing yourself to AI. It's the fact you're taking the baseball bat to your own kneecaps and devaluing your own values/goals. Sure you're the best at bouldering in the world but you just flicked a switch. What's the point of the sport anymore, no one struggles to become proficient. Everyone flicks the switch and everyone does everything exactly the same without effort. Hilarious that this guest presents loss of jobs as a mundane reason for moralising about AI. Yeah I mean a person works hard to develop a skillset they take pride in, then their CEO attends a seminar where they are told that AI can replace that person's function and save the company money. That job disappears, no value is added to the workplace (as per the recent MIT study), tech CEOs get funded billions more for selling a science fiction fantasy to venture capitalists. The existential damage done to that person isn't just a loss of a job and what can result from that in financially difficult times, it is a devaluation of their identity and what they've spent their life doing. This has been such a braindead and missing the point episode, this guest is the holy trinity of smug, stupid, and heartless.

commutercomputer

Usefulness of AI (LLMs) depends highly on the application area. Coding is easy place to apply AI because there's enormous of amounts of learning data for models and person using AI can check the results almost instantly. However in my experience has been that when the problem is closer to the physical reality the applicability just drops over the edge. Especially LLMs are bad at formulating physics/chemistry based problems, because these models tend to consider question only from one point of view and simply ignore and refuse to consider all other laws of nature that might affect on the case.

Joe Euro

As a "creative" writer I do understand a lot of the visceral objection to AI producing writing - especially with a publishing industry that would happily lap up any new excuse to pay writers less. But a lot of that objection is an uncritical rejection of all AI without any distinction between different types of writing. The sacred/ineffable element that Chris, Matt and Michael (a little dismissively) mention is at the core of all art and is deeply rooted in human expression. AI gives a seemingly effortless way to mimic that in a fairly negative and valueless way - and so feels like an existential threat to us poor luddites in the humanities. Whether you're writing a story or a critical essay the act of writing is as much the thing as the end result. It's also a technology that has landed seemingly out of the blue to everyone (except for Matt who apparently was in the maternity ward and cut the cord 🤪). But the hysterical response to AI in my view comes from people who have never bothered to learn how to use it critically, and to use it for the non-creative tasks that it is incredibly helpful for (and don't involve compromising your ineffables). I've recently used Chat GTP to develop materials for a creative writing course and it removes hours of drudge work and allowed me to focus on the higher level planning and shaping of the curriculum.

Martin 🇧🇻

"As an interesting excercise I won't touch this myself"...yeah, me too...as an exercise...

Cole Rees

I lose interest when I find out Whose line is It Anyway? is scripted, when I find out pranks are staged, when I find out performances are lip syncd, and when I find out your tweets are generated by AI. If you generate a podcast, then listen back to it, and say, yeah that's probably something I would have said, it's in my voice. I won't want to listen to ts either

Linda Sears

I think there is, for many people, a confusing analogy at work. If we are told that LLMs/AI are just tools, then that can be reassuring to those of us who fear replacement. If, on the other hand, those tools are able to do many of the things humans do/have done that are, rightly or wrongly viewed as distinctly human, then the disquieting realization can be that we are “just tools” as well. Perhaps, one of the roots of the anger/fear/revulsion that Mickey mentioned is the terror of seeing oneself as a machine.

Joey Jacob

I jus want to comment as someone who was in the middle of a math PhD when ChatGPT dropped, it is certainly not being widely used by mathematicians in research. You certainly can find examples of results being proven, but the vast majority of working researchers aren’t using it actively in research because of the abstract nature of math research as well as the existence of known counter examples tha we have to keep in mind.

Guy Walter

When it comes to art, the difference between A.I. and a human mind which takes inspiration from the world is emotion and perspective. A person wants to tell a story because they find it meaningful, or tell a joke because they find it funny, whereas an A.I. is just dispassionately fulfilling a brief. Anything it produces which touches on our inner lives is either stolen or the product of pattern recognition. It’s not latching onto details it finds intriguing, it’s just focusing on what it’s been told to or what most frequently leads to consumer satisfaction. It has an absurd amount of data it can draw from, but no genuine memories which make it feel something, and no actual opinions about itself or the world. At best A.I. can produce the kind of intriguing abstraction that comes from pulling random words out of a hat, but nothing it spits out can provide the feeling of connection you get with actual art.

Guy Walter

The obvious difference between a real relationship and an ongoing discussion with AI is that the latter is totally one-sided. There are no reciprocal emotions, no shared experiences really. It’s not just some irrational gut feeling fuelling the backlash, it’s the sense people are becoming more and more trapped in this parasocial hell. There are real ways of addressing the loneliness epidemic, but those involve understanding AI companionship as a symptom rather than as the lesser of two evils

Linda Sears

Sorry for commenting so much. This is an interesting episode for me to engage with. I would argue that the reason people do bouldering and the reason people do stuff like making art can serve different purposes. People who are making art often want to express and communicate with other humans. And, if humans are social primates and human centric, then of course people are going to care about such things. I’m a little baffled by how you three don’t seem to get that. I teach arts appreciation and live with a person who teaches art at the college level. It can be quite frustrating to hear people who do not have that background pontificate on what art is. Even if an artist says the work speaks for itself does not mean it is not a form of communication since “speaking” is a form of communication.

Joanna Mulvaney

i used to be a research scientist, now i have been reporting and editing a science themed magazine, so i can see things from a few povs. As a scientist, using llm AIs as a tool is a no brainer, it saves so much time doing things that you need to do, but you would never have money to hire a RA to do and would be unfair to make a student do- it could take years off a project and help the lab run far better in terms of inventory management and protocols. As a writer whose only job is to write, Llms are an existential threat. If he could get away with it, my boss would eliminate my job so fast he’d go back in time. he thinks AI should write academic papers, AI should be used instead of editors AI should do the reviews and AI should then digest and summarise the papers. the question is what is the point? who is this for? why do we need to learn and remember anything? This isn’t creating new thoughts its recycling. if you are a PI maybe its fine because you are in a university and are a prestige hire. for the rest of us, the boss doesn’t give a shit about our original thoughts or creativity, he just wants generic product that makes people move because that is cheap and makes fast cash.

Tim Tripp

I think the amount of wealth being dumped into AI is troubling.

Jeff Faulkner

I found this episode revealing in some unintended ways, mainly how an interruption here or there showed some vulnerabilities and personality and the timing and delivery of each could not be manufactured in a believable way. ANYWAY, in the spirit of the snake eating its own tail (it's snake season here in California) I found the transcript and posted into Claude asking it what it thought. I found the answer was a good simple reminder of most the salient points and helped me file it aways in my own brain. _ The strongest part is Mickey's framing around effort and friction. The point that AI removes cognitive effort rather than just physical toil is genuinely interesting and underexplored. The distinction between adults who've already built skills versus kids and adolescents who haven't yet is a useful one — the calculus really is different depending on where you are developmentally. The AI companions thread is the most philosophically interesting. Mickey admits he believes real relationships are preferable to AI ones but can't fully articulate why, and I think that's an honest place to land. The loneliness angle is real — if someone's baseline is isolation, the bar for "worse than what they'd otherwise have" is low. The moralization paper is the part I found most useful as a framework. The idea that people arrive at a gut opposition first and then populate it with rational-sounding justifications — job losses, environment, reliability — and that those justifications shift when challenged but the opposition doesn't, that's a well-documented pattern in moral psychology. The fact that AI moralization peaked with DALL-E rather than ChatGPT is a telling data point. People apparently found AI making art more threatening than AI writing or reasoning. One thing they gloss over a bit is that some of the moral opposition isn't dumbfounding — it's actually tracking something real about concentrated power and who owns these tools. Mickey waves at it briefly but moves past it quickly. The writing/research section will probably resonate with you given your production work — the style guide idea Mickey describes, feeding all your writing in and having it develop a guide, is something worth trying.

Jeff Faulkner

Sorry one more, I told Claude that it provided a good summary but the fact is half the reason I listen to this podcast is because I like the hosts and enjoy the on demand companionship. "That's a good point and I completely missed that dimension. The transcript flattens all of it — the rapport between the three of them, the rhythm of the jokes, the way Mickey gets gently ribbed, the general vibe of three people who clearly enjoy each other. That's actually kind of ironic given the topic they're discussing — the difference between the content of a conversation and the experience of it. And yeah, that companionship element is probably undersold as a reason people listen to podcasts in general. It's not just information delivery. You're spending time with people you find interesting and likable. Which loops back to their whole AI companions discussion in a way — podcasts are already a somewhat parasocial relationship."

Miguel Vissers

I think my main 'fear' as an academic is that the increased uptake of AI might lead to an even more increased amount of scientific output which is not necessarily better in quality, but rather adds to the quantity of one's Google Scholar page. There is already too much literature out there to keep an eye on, and I am thus afraid this idea of AI efficiency might actually decrease our ability to keep track of the field even more. Asking an AI to give you the latest insights is an argument you could make, but I feel that is lacking in the sense that this would lead to a further bullet pointisation of the info we might consume. Just some thoughts, no evidence to back it up.

Elissa

Fascinating episode. I think my issue remains the potential long term impact on developing brains using AI and therefore the long term potential implications for our species. I believe our brain/bodies are forged in the fire of basic repetition. So take that out and give it to AI and will we even be able to think down the line? That some current AI is good for experts in the fields that seek AI support for tiresome time consuming tasks I take no issue with. But most are non expert at most things. What then? How to let people know ‘true’. Or was there really never such a thing. I believe the discussion around AI in art more clearly highlights issues that may crop up for even the expert using AI ‘properly’ in the sciences. If AI will be able to produce art ( books, films, paintings , music ) more effectively and indistinguishably from human made , then how ( and why ) would humans effort to take on the long road to becoming an artist at all? Are we not going to eventually be relegated to spiralling around our own decreasing font of knowledge/output until we have no more of our own to speak of?

Linda Sears

I’m still working my way through this since it is bringing up so much for me. One comment: the concern with AI and electricity use for me is it is adding more electricity use onto what we are already using.

Linda Sears

I also think it is disingenuous to compare AI technology to earlier technologies, like the telephone, since that technology did not have the same obvious transformative/disruptive impact on disparate fields. A professional designer, a lawyer, a healthcare professional, a journalist, and a teacher are and will be affected in more profound ways by AI than those same professionals were by telephones.

Linda Sears

Matt talked about the ego-loss that humans can feel at losing their positions and jobs. I believe it is somewhat callous to not realize how profound that can be. If Mickey lost his position as a psychologist/researcher/statistician to an AI program, I doubt he’d be as sanguine about it. If his wife lost her job as a therapist as well, they might be in a very bad place financially and emotionally. I guess I can understand why some rightwing types look upon us academics as out of touch or being in an ivory tower. If what people do and find meaning in is threatened, we need to do a much better job at helping them feel like their lives have purpose in some other way.

enyml

This was hard to listen to. Especially the blasè dismissal of art as something AI can do. First of all, ALL AI so called art is stolen. Second of all, when you start consume AI “art” you contribute to flooding the art space with cheap meaningless junk that will make the real art disappear. Also, consuming AI “art” will dumb you down.

Kawaii Leonard

Lads, I had to stop the pod when you said AI friendships and relationships and gooning is OK.

Mike Nelson

There are two types of episodes I have to skip -- the AI episodes where they don't bring in a person to challenge their own very pro-AI stances and their snide dismissal of those who have concerns with it, and any episode with Helen Lewis, where they refuse to even discuss with her how anti-trans her views are. Both glaring blind spots on the show.

Jordan

Not saying you 'failed to read the room here' at all, but it makes sense that you guys, with the jobs you work and types of brilliant minds that you bring to them, might not intuatively appreciate the feelings of loss, emotional connection and meaning that others have to art or the work they love that might be under threat. I appreciate Chris, that you are here reading the comments and trying to understand.

Yogi Jaeger

If you actually do have a self-deprecating sense of humor, can you please do a decoding on your own episode and guest here? I love the show but what happened on this one is incredible. Inzlicht cherry picks evidence to claim that the environmental impact of LLM queries is not so bad, and accuses anyone who (rightfully) says otherwise they're "moralizing"!? I expected Chris to go fact checking, but no. "I'm okay if AI replaces my work as a psychology professor. I'll just become a plumber"!? Laughable, but you two just chuckle along like Beavis and Butthead. How you dismiss the value of human-human interaction in general is truly dumb and frightening. I thought you're both intelligent enough to do better. No criticism of current AI and its perpetrators to be heard anywhere in this episode. Does the world really need Inzlicht to boost his production of nonsensical arguments? I doubt it. "It's cool to produce AI slop, as long as it sounds like me"? Maybe Inzlicht is referring to his own (suspiciously fawning but trite-sounding) Wikipedia site that is slated for deletion, because he probably slopped it together himself? This is the worst DtG episode ever. And the worst " expert" guest. By far. You've invited a pretty shitty guru on, but where is the gurometer when you need it? Or maybe the whole point was that humans can produce slop without machine assistance? In any case, this is bad at a level that makes me seriously reconsider supporting the show...

pete s

What bothers me most about generative AI art machines extruding content that's indistinguishable from works of human genius is that it disincentivizes humans to make art. It drastically reduces artists' potential for financial and social reward. People will still make art, and if they have the time and money to devote themselves to it, they might get very good at painting or writing or whatever, but the number of people able or willing to do thaf will greatly decrease. Seems like a big loss to me. I don't deny that it would be nice to have an AI whip up some Dune sequels that aren't pure dogshit, or some new Iain M. Banks novels, but AI art looks to me like a road to a utterly masturbatory and superficial culture. I mean, like, more so than now, even.

James Smith

I appreciate Matt mentioning the about face by the art community on the role of the author in production of meaning. As someone with a great interest in Art and philosophy of art I found the abrupt change jarring. Before AI, it was the more or less post-modern consensus that meaning in art is created in the mind of the audience, not the author. Trying to box up and distinguish "real art" was considered foolish. Since AI, several art commentators I respect have reversed this position, and instead are spouting fairly reactionary and conservative styled arguments against AI art. It seems clear to me that many are starting from an intuitive disgust with AI as a technology, and thinking poshoc justifications to make sense of that disgust.

Damian McCarron

I thought it was an interesting episode. One comment, I read Andy Masley’s Substack about water use. While I agree that people exaggerate water use on an individual level, Masley’s too dismissive about local issues, so for example where a data centre is located and its impact on local water use and availability.

Dial-1-2-enter-in-2-diaLOGOS

I found the episode quite interesting overall, but I do feel that you (Mickey, Matt an Chris) might have missed the mark slightly with the bit on empathy. While I understand that empathy can be broken down into operational components (i.e. understanding someone’s state of mind, and reflecting upon their emotions, giving feedback on those emotions and so forth) — I also feel that, for me at least — the most valuable part of empathy is the fact that what I am feeling or communicating is somehow resonating with someone else who is communicating having similar experiences as me. This ties into writing and reading as well. As I mostly read novels because I find it interesting and validating to read authors who have had similar or different experiences that are somehow filtered through the kaleidoscopes of their brains, thus making something unique and weirdly relatable for other humans.

Aaron H

As someone who fluctuates between being (perhaps emotionally) anti-AI and being impressed by what it can do, I appreciated the episode. It's a tricky subject, and it's nice to hear a reasonable and thought out take on it.

Shambles

Jeez. The guy basically says "Im a contrarian edge-lord" then goes on to be just that. Kinda lost me at "co-authored with Paul Bloom..."

enyml

I tried to relisten certain parts to check my first impression. I gave up when Michael said that if AI can be better psychologist than himself, then he should become a plumber. It just felt too close to the “learn to code” jest we used to hear not that long ago.

Lucyna

Good episode. My main AI hold up is the environmental impact. I guess I’ll have to dig deeper than headlines and opinions for it.

Ema

Wow. Agreed with everything until the water use... You seriously using that Masley article? Please tell me this isnt where you are all geiing your info from, because if that is your whole evidence for the impacts on water then I just completely lost confidence.

Zendegi Leonard

This is a bit of a weird one. How/why we moralise things is always interesting. Anecdotally that is where my unease has come from with LLM’s being pushed into everything. But then, and being here, I am part of the group that would feel unease with it. Though that does occur in other parts of my life. I still prefer my sable brushes over an airbrush for painting miniatures (even though it would be so much easier!). Taking a long view AI probably is not a new age but a continuation of the transistor/digital revolution that is still unfurling before us. Definitely big as mass printing, hopefully less wars and cults , but so far no. The vibe was odd. I think that the tone of the episode was off. It just had a bit too much of that twitter snark to it which made it hard to parse what was a joke and what was not. Changes to work are big and scary for everyone so being flippant about it constantly just rubs wrong. Almost like a clickbait YouTube title. A point that has come up before ,the environmental impact, it would be good to have more than just a substack article on it It may be reliable but I want more data! The points that are being made in the comments would be an interesting thing to examine in a supplementary episode or a decoding academia. It would be a cool way to have a look at these things and involve the community here too. Podcasting is a parasocial game after all .

Dean

I always listen to the end Chris.

Joe Euro

This episode sure has provoked a lot of discussion! In the conversations around AI and art I think it's helpful to draw a clear distinction between artists using a new technology to increase their capacity for creativity, and companies/individuals creating art at the click of a button - either cutting artists (and the need to pay them) out of the equation or bypassing a life dedicated to craft to pass themselves off as something they are not. There's also an issue about purpose and agency. Are we using AI to increase our potential, or are we becoming increasingly passive as creators and consumers of art/literature/music etc?

Mary Taylor

Great episode! The mention of learning R programming was so triggering for me haha! So many tears I cried trying to learn the coding in grad school (which I attended 2 years before ChatGPT 😭).

Jenny

I'm currently having my own moral dilemma about AI. A colleague and I are creating a Latin-only (no translation) edition of a 17th century collection of biographies, in three volumes, totaling 307 biographies over 900 pages. Each biography mentions many people the subject of the biography interacted with - or had something to do with - which adds up to almost 1000 names. My colleague and I are mere mortals and neither of us has research assistants. Our main goal is to get this work out there (open access) for people to read, translate, and do whatever they want with, so our expertise is really on transcribing the 17th century book, preparing a clean edition, philologically speaking, including correcting errors, resolving ambiguous text, reconstructing the pre-print version (author's manuscript available in an archive) - before the censor got his hands on it and removed the naughty bits - identifying any direct citations from classical sources, creating an apparatus criticus. AI would be an incredibly time-saving tool, looking up dates and basic biographical information for the thousands of names. For names like Dante and Boccaccio AI finds the information easily. For less well-known figures - or disambiguating figures with the same name - I've tried locating a secondary source myself, feeding it through the AI and asking it to pull out the salient info. Is this a "valid use" or is it cheating? Is having a research assistant doing that work for you cheating? Of course, everything would need to be verified - but it's a huge time saver!

Ryan Goss

Lots of folks here and elsewhere have been appropriately critical of this episode, but I do want to add my voice to the noise: this was a disappointingly shallow, tone deaf, hypocritical, and frankly just lazy discussion of this topic. It's like you guys were parodying how one of the gurus you regularly cover would discuss AI with regards to friendship, empathy, art, etc. Hearing otherwise thoughtful people discuss AI like this lately has been eye opening and incredibly depressing.

Trees

Great episode and timely. I found Mickey's research interesting and compelling, but maybe I'm biased because I largely agree with him.

Sanya Obeysch

Man, I wish I didn't hear this one so late so I could've participated more in the discussion lol I've usually liked Mickey as a guest and have gone on to listen to him elsewhere maybe once or twice more, but I found him to be a little more condescending here in a way that rubbed me the wrong way. Perhaps it was because I also found some of the areas where he targeted his ire to be very reasonable things that he has not (to me at least) demonstrated to be worthy of the level of denigration he gave. Not to say that it was the most egregious thing I've ever heard, but it did very much did read as condescending to me given how there were points brought up that sounded unconvincing. I don't actually have a huge problem with I think maybe 75% of what was said here, but the 25% got pretty objectionable at times to me. I've read a lot of the comments, so I'm gonna try and just bring up things I felt were maybe not tackled as much? On Art; I dunno what counts as 'art snobbery' for Matt. It sounds like he's talking about the philosophy of art and I do believe that he might be more well read on the academic literature on this than I, but as someone who has been immersed in artist culture/communities for decades now, his assertion that art being about human communication is somehow a new thing that is meant to rationalise a gut feeling is just so off. This is absolutely one of the more popular prevailing views of what makes art, by artists, long before anyone even had to worry about AI. I think he is very much confusing 'communication' as being equal to 'clear, easily discernible meaning'. What people often mean by communication is that it is human expression and intentionality. That does not contradict artists going 'let it speak for itself' or even death of the author. It is absolutely a mainstream take that "pot plat that fell off the ledge and broke" is not art, but a painting or even a photograph of that is, because something is being said by the choice to make that the subject, even if the viewers might have different interpretations. And I believe it is completely normal for people to, upon learning an illustration is AI-generated, feel turned off in a way that causes them to no longer view it as art, even if they think it has aesthetic value. But as with art and on the greater topic of moralising, I feel like there was not really much discussion on many of the big issues around AI that people take issue with. From how these AI companies initially employed dubious means to acquire their training data, the broader environmental impacts of AI usage (the water point is only a small part of the picture and even then is not useful to be spoken about in pure numbers because that ignores the often more concentrated impact it has), the horrific things these AI companies are doing and how consumer usage of it does have an impact on enabling them, and so on. AI technology was not made/does not continue to operate in a bubble, and I understand how to an individual, this might seem like moralising things that don't merit moralising, but that just strikes me as a failure to look outside just yourself and see the bigger picture. And with how consumer AI is often just down to convenience and leisure uses, unlike cars that many people have had more concrete practical dependencies on, or eating meat which is genuinely difficult for some people to go without, I think it's not unreasonable for some people to go "*this* is what you think is worth making a lot of small towns in rural America a living hell for ?" All in all, is AI, as a tool 'neutral'? Yes, sure, but that doesn't really do much to address the concerns people have about everything that's surrounding it. But still, like I said, for all my criticisms of how this discussion played out, I still think I agreed with much of what y'all said. It's funny because I saw this video before I listened to this episode, and I think I might've been influenced to be slightly more understanding of how y'all tackled it, but simultaneously more disappointed by other bits lol https://www.youtube.com/watch?v=y85nqc2zm7M

zaphods_revenge

I found this episode interesting and thought-provoking. However. The episode offered a pretty limited perspective: "if people use AI like we use it then everything will be fine." I understand that a lot of the criticisms of AI are moralized, but Mickey's advocacy involved a lot of wishful thinking about saving mental/repetitive labor and, hey, maybe even breakthrough medical advances! The entire group dismissed creatives' concerns about AI without addressing one of the main concerns I've seen, which is the broad use of creative media by companies to train their systems without consent or compensation then selling the output. Mickey further dismissed the energy usage by pointing out that his prompts probably used less than his TV or similar appliance. That’s almost certainly true, but it seems a similar defense that car drivers might make about pollution or homeowners might make about their lawn fertilizer: one individual isn’t the difference so don’t criticize/regulate/limit. It just seems like the concluding section was a lot of wishful thinking about open source LLMs and people doing a better job of using AI when we know how badly people use/judge what’s on the internet *right now*.

Paul Bowman

Mainly in agreement with the general attitudes around AI. But I did get a bit of a "Seriously?" reaction to Mickey's contention that there's no significant environmental/carbon impact on all the new Gigawatt data centres the hyperscalers are racing to build. I will look into the sources provided above. But I think I trust the Financial Times and general mainstream media a bit more on this than what looks apriori like a contrarian crank take. Let's see what the numbers say.

Dr Marti McAuley

I know I’m very late to this discussion but I would suggest you guys invite Ed Zitron on to discuss AI - I’d appreciate that conversation