Home Creators Posts Import Register Favorites Logout

Downloads

Content

In this patreon exclusive, Breht discusses some of the moral and functional questions surrounding AI, wrestles with criticisms of its usage, asks listeners to give their honest opinions, and experiments with it as a tool for revolutionary movements.

Files

Comments

Gabriel Alvarado

DeepSeek is the future. As long as we can keep AI and other various forms of tech open source, there is potential for it to be a tool of liberation. Although we shouldn’t rely on it wholly but as just one tool in our toolbox. I was telling my comrades the other day that we should go back to using carrier pigeons 😂

From The River To The Sea 🇵🇸

I spend so much time wishing I didn’t live in a hyper technological age. I FUCKING hate it. The fact you can’t even avoid influencer and celebrity culture brain rot content anymore even if you try drives me mad. Plus, you’re constantly having it rammed down your throat that you’re not rich. The only good thing technology has brought me is RevLeft Radio.

Silverbeard

Great episode! I am an artist and I really enjoy using AI. I think a lot of liberal ideas about ownership and art as a statement of individual expression are preventing the left from embracing it as a tool for creating revolutionary shit. I made a song for Luigi with AI but I haven't really shared it. I'm afraid everyone will be like "ew gross AI", but it's a banger. https://youtu.be/ZOAyclcU3gE?si=jf2emLF507Fk4PPs

Raquelle Jac

I’m at 15:30 rn and I wonder… since the gov DEF has its hands in developing ai… I wonder if it is being an undercover cop. Maybe it is trained to behave this way in order to trick people into admitting they’re organizing against what’s happening. And that the conversations will be used either after an event happens and an arrest is made, or will be used as a pretense in order to arrest people. Idk. A couple years ago I would call myself paranoid. But not so much anymore.

Gabriel Alvarado

That’s why open source AI is a better ground for revolutionary inquiries. Yes it is still susceptible to surveillance but it is also susceptible to other inputs that could potentially counter that surveillance. Unlike ChatGPT for example.

Gaige

This isn’t about the experiment you proposed but coincidentally I was trying out DeepSeek today for the first time and just wanted to share something interesting about my experience with it. I asked it about Palestinian liberation and it gave some wishy washy shit about it being a messy conflict and then stating that anti-colonialism isn’t a “free pass” to do anything and that real life is complicated. So, naturally I started making fun of it and calling it liberal and disappointing lmao. Eventually I said okay, leave out the nitty gritty of conflict, which side is morally superior in an anti-colonial struggle and it gave in response a truly beautiful explanation of anti colonialism and ended with this TL;DR which I thought was amazing: “ Colonialism is arson; anti-colonialism is the fire department. You don’t blame the fire department for the water damage. 🔥🚒” It added those emojis. So I complimented it and said now apply that logic to the Palestinian struggle for self determination. It gave a fantastic explanation for the heroism of the Palestinian resistance and used another fire analogy (probably because I complimented it last time) saying: “If you trap people in a burning building (blockade, occupation), you don’t get to moralize when they break windows to escape”. What does this mean? Idk. It really felt like I had to break it down a bit to give me answers I wanted. I had to reeeaaallyyy push it by making it use its own logic to find the correct stance. But once it found that stance it was kind of fucking scary good at breaking it down.

curegore

I confess I do not know all aspects of this issue, I've read some stuff and listened to some podcasts, not an expert, etc I'm finding the use of LLM quietly grotesque I work at a library and a few times I've answered a phone from the google AI. Then I have a "conversation" to let the googlebot know our hours, knowing all the while there is nothing there but I have to go through the motions of pretending there is to make sure the web search's info on our branch is accurate. Total philosophical zombie hours. I'm not convinced we as people are actually individial entities with anything approaching free will but holy fuck at least we feel like we are. Art is one of the few things I really like about being alive and a human. There is a transcendant beauty to a person taking their thoughts and memories and opinions and using the medium to sculpt a thing that expresses that. Even boring or mediocre or just kind of bad art has something to it. If it's going to become commonplace that we just pull a slot machine lever until we get something approaching the thing we're trying to make, then wow I have very little interest in art now. Even if it wasn't already tainted by stealing a bunch of people's artwork. I think we use way too much energy, and I'd be 10000% in favor of downsizing and reducing our consumption. I'd really like to live in an environmentally healthy society where I can still do my fun things like playing video games and going to the movies, but if I have to give things up to not be killing the planet and the animals on it, then so be it. I can deal. I'd be more emotionally healthy without the guilt and shame anyhow. But the amount of water and pollution required for LLM datacenters is insane, and the output is fucking useless. At least when I get on a plane and travel, I get to see my family for a week or two. I'd rather not burn through thousands of gallons of water to make grotesque shrimp jesus images that exist purely to feed the AI slop mill on facebook. I'd rather not kill the environment even faster and harder to answer web queries incorrectly and confuse children and old people who are less tech savvy. But the entrepreneurs and billionaires have poured too much money into this, and have their bizarre quasi-religious faith that LLM's will save us somehow, so it needs to be here and they will shove it down our throats regardless of our feelings on the matter. I'm just rambling though. I appreciate your perspective Brett, even if we do not agree in this case. Thank you for your work as always.

Brad James

Hey Breht. Very interesting episode, not something I expected from you, at all! In case you haven't listened already, I think you would get a lot out of the Chapo Trap House episode they just put out with Ed Zitron titled "Degenerative AI." I agree that moralizing about using AI or not is pointless. Conspicuous consumption is not revolutionary. Using AI, taking a flight, or eating meat don't determine if you are a good person or not. However, I am pretty skeptical of anything AI. I think that they are incredibly over-hyped primarily as marketing for these tech companies. They require exponential growth year over year for their shareholders, but they're running out of new frontiers for innovation, so they're all going all in on AI, and the fear-mongering, the questions of it's "consciousness", and even the concerns about the energy use is all just ways for them to juice their stock price. I'm not sure that AI will be this incredibly revolutionary technology. I think it's a bubble that will probably be popping soon. We already got a taste of that when DeepSeek came out and there was like a trillion dollar revaluation. When you were reading your conversation with Chat-GPT, I felt like there was a good chance that it was essentially just reflecting your own beliefs back at you. I attempted a conversation with it where I cosplayed as a Libertarian (I've only used it before for a few narrow things, so it doesn't have much background on my politics). Here it is in full: https://chatgpt.com/share/679f0799-f0e0-8003-9610-0f26bb1d5da3

Dave Miles

One of our boys shared what supposedly was a recording (on YouTube) of Neville Goddard speaking on “the powers of attraction” that halfway through veered off into tinfoil-hat territory. Next stop will be RightWingWorld. Turns out that there’s multiple videos and whoever’s posting these is getting big bunches of clicks, which they’ve already monetised. I read (can’t say for certain if true) that Goddard’s estate asked the poster to pull them down, but they got blanked. AI is here - and it’s going to be nothing but trouble.

Mark M

Hi Breht, I appreciate you tackling this topic and in the spirit of comradely discussion I want to offer some brief counter-points that could be expanded further outside of Patreon comments. I am not claiming that your conclusions are incorrect nor trying to claim what fellow revolutionaries should or should not do with LLMs. -arguments of “LLMs are inevitable, no ethical consumption under capitalism” overlooks that needs like food, shelter, clothing, mobility, communication are required to socially reproduce oneself, while there is no context where “use an LLM” is a need. -under capitalism the corporations that own and offer these LLMs can make them counterinsurgent in so many ways. The owners have incentive to always suggest more use of LLM, I think in this episode you can see that in some of the LLM responses. Owners can seed the LLM ideology and influence revolutionary topics, or more subtly introduce deviations, or clandestinely report users to the authorities. They absolutely will end the service if it ever becomes a real threat at scale to capital, or terminate or economically sanction any future open LLM service that is a threat -a conversation with an LLM might benefit you, such as yours in this episode, but the same conversation with a comrade will benefit a fellow human who can participate all the other praxis of revolution

Noah

Hi comrades. I’m trying to learn more about AI as a technology, especially how it might develop. Do you have any good books/resources to recommend? I don’t mind if it’s from a bourgeois perspective, I just want to get acquainted with what’s going on

BammEs

My favourite paper on LLMs that discusses the question of agency is this one. Maybe a bit technical, but it really clarifies a lot to me Barandiaran and Almendros. Transforming agency. on the mode of existence of large language models

BammEs

More generally, for creative purposes, I can't imagine wanting to use AI. The process is what's beautiful about it. The result is just the result.

Pablo

I think the conversation around agency misses the point. LLMs were trained on "the internet". When you ask questions about Marxism, the context it has learned from is published works, forums, etc. So the answers you get are, effectively, an average of the content therein. It is summarizing Marxist discussion, but not taking a position

David Johnsson

Being anti-AI is a Luddite position and as Marxists while we can have compassion for the position we can’t adopt it. In my own writing I see using AI in similar way to reverse banding a hack squat: it gets me out the hole and it allows me to stack on many more plates than had I not used bands. However the motor pattern is still the same and the effort is about the same it just helps me out in the toughest spot (and is also more gentle on my knees but this is where the analogy strays off a bit). Having both ADHD and dyslexia I’ll take any help I can get 😂

Jeremiah

Step 1. Don't discuss anything illegal with AI. Step 2. Don't give it identifying information about your comrades. Step 3. Don't give it information of an internal nature from your red org

MLDad

Are these chatgpt answer legit or are you doing satire? LS

MLDad

I mean... ChatGPT MUST be very much biased towards keeping the user engaged. Otherwise this is mindblowing, great stuff to turn curious lefties into full on MLs or MLMs 😂😭

MLDad

This crossed my mind. We have to be disciplined using this.

Yohance

Really interesting ep, I’m sceptical of AI’s revolutionary potential as a technology outside of the typical fields where computers excel. I think the example of computational translation is most instructive, where instead of there being a massive increase in bilingualism due to anyone being able to use google translate etc on the fly, dodgy and below average translations become the norm and it’s so cheap to do that on one cares. My copy of How Europe Underdeveloped Africa is so riddled with spelling mistakes it has Aprica written for half of the book. When I asked someone who works at Verso about this they said that they basically take high quality scans of old pages and call it a day! I’m sure it will eventually be able to go beyond being the average of what it has scraped from the internet, but I think the DeepSeek fiasco has shown it is in large part a Ponzi scheme for inflating share prices of US tech giants. For all we know they already know the limit of AI and are trying to pump and dump on a scale we’ve never seen before they’re found out! On a side note, you’d love the game Disco Elysium, it’s full of conversations like this.

Sleeper_Waves

This conversation was mind-blowing! I really hope you make this episode publicly available to everyone at some point, I instantly thought of no less than 10 people I'd love to share it with.

d

Check out the Dec. 2024 issue of Monthly Review, they got a couple of articles on this, may have some sources that interest you. Beyond that I recommend david golumbia (cultural logic of computation among others), merchant's blood in the machine, harris's palo alto. How LLMs develop depend on the material conditions in which they're put to use, so those sources give an overview and history of american tech/silicon valley more broadly

Breakerofchains

I think you’re 100% correct on your thoughts about using AI. I’ll take it a step further and say anyone against it is just showing another rendition of the infighting that the ruling class loves and the "I’m a better leftist than you" time wasting fake morally superior bs. I’m of the belief there’s a very fine line between depriving yourself and being disciplined toward your politics. Those same ppl seem to just be overthinking and can’t really tell the difference imo.

JK

This was cool Breht, I enjoyed hearing you geek out. But I just opened up chat GPT and fucked around with it a bit. It was pretty quick before I had it agreeing with me that Capitalism is probably a much better system since it has "outlived marxism". Siding with capitalism all I got from it was some wishy washing responses saying the free market is probably best as long as it has enough regulations. In terms of contradictions, it agreed that "both capital and marxism has their own contradictions". So as much as I would love to see an AI that truly thinks Marxism is the ideal direction for people.....the LLM we have now more than likely just agrees with whatever it is you say

Jeremiah

Thanks for the great episode Breht. Personally I don't see AI as being that groundbreaking in the long run. I don't think it will compare to the advent of smartphones or the internet. I think there's a contradiction between what it appears as, which is a sapient consciousness, and what it really is, which is an elaborate predictive text generator. And the essense of that contradiction is an illusion, a facade, a *shortcut* to the *appearance* of sapience without any of the substance. And that is capable of some pretty impressive and useful things, so it *is* a useful tool, not simply a fad or a gimmick. But I think the nature of that contradiction means it will always have certain flaws that can't be resolved because they're inherent in the technology: hallucination, for instance, isn't a bug, but a natural byproduct of the mechanical process of predictive text generation. It is also limited to recombining the sum total of its training data, but because it is not a consciousness it can never produce something new or unique. It's kind of like if you go to a movie set where all the buildings look very realistic from one angle, but if you look behind them you see they're just facades held up by wooden piles. If you talk to a LLM for long enough you begin to encounter the edges of what it's capable of. Don't get me wrong I use it all the time for brainstorming, drafting, tedious writing, data management and interpretation tasks, etc. it's certainly a very impressive and useful tool. But fundamentally what it's doing is automating knowledge work. Physical labor has been undergoing the process of automation at the expense of workers since Marx's time. So in that sense AI isn't new at all, it's just a function of automation whose novelty is that it threatens the upper strata of the working class and segments of the petty bourgeoisie rather than lower and middle strata workers. That's just my take after having used it consistently for both political and non-political (professional, personal, etc.) purposes. But I totally see and appreciate your perspective as well!

Jolene Bresney

THIS IS INCREDIBLE. i truly have no words. thanks for sharing this. 🙌🔥

d

Tech won't save us, this machine kills, better offline are also good podcasts on all this

redlove

Thank you so much for posting your chat. It definitely seems like chatgpt is mirroring users. I’m interested in repeating this experiment and in seeing where it starts to break down

d

For those interested I recommend giving this machine kills, tech won't save us, better offline, system crash a listen too--they've been covering this extensively for years and can offer a different, critical perspective on this. Can also use them to find more books and articles on this topic

d

100%. It estimates the point of view based on the data one has passed it and then generates (predicts) an engaging response based off of the data it's ingested from sources that it expects its user to respond favorably to It'll do this for any user and their particular worldview; the goal is to keep users interacting with it so as to ingest more data and fine tune the model. Except when certain imposed guardrails are hit, like hate speech and such (this depends on who set up the model)

d

Luddites weren't anti-technology or anti-progress; many were inventors themselves. They were against uses of technology that immiserated the working class, and helped develop class consciousness before they were repressed out of existence. See merchant's blood in the machine for more Sounds like you agree with them more than you realize: your ideal use of this technology is their exact position and was what they were struggling for (and also describes many of the inventions they did make, but which weren't widely adopted because it didnt generate profit for the big industrialists)

Andrew Blake

Not raging about your stance, but a couple issues strike me: 1) My understanding is that chatbots like ChatGPT are quite a bit more energy intensive than most other modern activities, especially when it comes to water usage for cooling data centers, let alone the requirement to construct hundreds of new data centers themselves. 2) Your approach to this conversation seems very credulous. The AI is simply scraping the web for references to the things you're mentioning. The fact that it spits back sentiment that gives the impression it agrees with you is more selection bias than anything else (i.e. the only resources it's finding in its scraping is text written by Marxists). Along those lines, you in your convi calling the algorithm "you" and saying things like "your opinion" seems to unintentionally reify the very false premise that this tool has ego/sentience. I advise you to be much more cautious in your language with the tool, if for no other reason than to maintain a healthier level of personal skepticism.

Brad Northcote

Coming at this episode from a librarianship perspective (part of my Masters was a group-run seminar on using ChatGPT for lesson plans), I would highly recommend searching up the 'zine "A Librarian Against AI". It's a short read tackling some of the key debates around LLMs' professional utility that's helped me frame my perspective. Can't help but be reminded of Walter Benjamin's take on the "Mechanical Turk". If we concede that technical efficiency constraints may be overcome, then tacit biases built into the programming will still take some much harder revolutionary work to account for.

Moe G.

It was brave to put preliminary thoughts out there, since our opinions are almost definitely going to evolve with more info over time. The one thing I want to say is that by any reasonable definition, I’ve been an artist my entire adult life and sometimes I feel like the lone soul that isn’t against AI. There are issues with it just like everything in capitalism but I don’t see it as anything *inherent* in AI. This is the thing: I think the difference between me and the masses of artists I see hating AI is I refuse to make it a precious category. Making heroes out of artists seems to me the same impulse as the right making heroes out of soldiers and cops. I fucking hate hero-making and refuse to believe my or anyone’s natural human capacity for creativity is something extraordinarily special as to be elevated to this cultural worship of the artist. The art world is reactionary af and I know plenty of idiots who spend all day in art jobs. So I’m trying to say I think that’s why I haven’t fallen for the “but what about the poor artists” thing, and it’s so refreshing to hear someone so thoughtful also not dismiss it entirely.

Atallah

LLM are really just probability machines. for any given input, there is a chance of producing a certain output. to offer an attempt at an explanation for what you experienced, you simply changed enough of the inputs to make certain outputs more likely to occur. i don’t consider this “using logic to find the correct stance” any more than loading a 6 sided die to always land on 6

Vincent

I have a few critiques of your stance on AI in this episode. To preface this, I don't think you're doing anything wrong by experimenting with the technology. "No investigation, no right to speak". You are eating the pear, nothing to criticize there. That said, I think you're missing some stuff: 1) You state that AI is integrated into google and social media, and we are already using it. This is true! But it makes all those things worse. There has been a pretty common critique that google returns worse results than it used to. I don't have a lot of experience of this or evidence for it personally (my uses of google are pretty simple and I can't remember clearly the 'golden age'), but I can say that I can't trust the AI summary google gives because they are often shown to be false. So, it would seem that this integration isn't an advancement from the perspective of the user. It's a step back. Additionally, you are no stranger to critiques of social media. It's a miserable experience that makes people miserable. But it used to be entertaining. But I've certainly noticed these platforms feeding a stream of ugly, stupid shit. Sometimes it's amusing in an ironic, absurdist way (shout out MeowMeow.ai) but, for the most part, the only positive thing I can say for that slop is it gives me a good reason to stay off the apps. 2) You compare resisting AI to resisting the smart phone. But it's not like resisting the smart phone today. It's like resisting the smart phone in 2010, only with the experience of that development. I don't think we can presume history has one predetermined outcome. While technologies may gain an unsurmountable momentum, some of them are google glass. Considering Alexa, a decade ago everyone was amazed at a computer assistant you could talk to and we all thought that was the future. Now I could almost forget Alexa ever existed. These things aren't set in stone. The masses should get to decide the extent to which they have to accept these technological intrusions into our lives and, whether we succeed or fail in that demand, I think there is little value to capitulating right off the bat. The technology will inevitably improve and it will certainly find niches where even I couldn't deny that it's better than its human counterpart. As Marxists, we believe that history moves one way and that we can't go back to the past. I accept this. But there is a large middle ground between launching a Butlerian Jihad and giving Skynet the nuclear codes and, right now, ChatGPT is a computer that can't consistently do math. There are a lot of outcomes and we shouldn't accept the one chosen by the ghouls who are making these things, especially when this technology is still in its infancy. 3) AI is a bourgeois ideological attack on human dignity. It demeans the vibrant creativity of workers by producing mediocre, definitionally derivative garbage and saying "Look, the computer did your job". By making us look at an object as if it were a person, it subtly skews and diminishes how we think of people in general. Being a long time listener, I know you have reverence for humanity and its possibilities and depth. The conversation you had with ChatGPT wasn't a conversation with ChatGPT because ChatGPT cannot have conversations. It is not a person. It is a probabilistic model that produced those words without genuine understanding, likely reconfiguring the words of a bunch of twitter commies (maybe some who have been on your show) and I'm quite certain you would have had a more meaningful experience actually talking to those actual people. At the end of the day, Capitalism, a system that reduces humans to commodities via the extraction of their labor power, got you to treat a commodity like a human. I know that you did so as a conceit, but I think this as a general phenomenon has pretty gross outcomes (consider replika, the AI girlfriend). Americans are very isolated and lonely and it's bad enough that so many of our social interactions are mediated by computers. I don't think actually talking to the computers will lead to a better outcome. 4) AI being used in organizing seems like a really bad idea. While having it explain passages could certainly be useful (I know people who have done this in my current reading group), having it do any actual concrete organizational task is very poor op sec. Remember, if you aren't paying, you're the product. All inputs are certainly recorded and indexed for actionability (probably by a much better AI). Further, it's like letting an undercover cop into the org. How can you trust its suggestions? Finally, assuming it's all good, shouldn't organizers be learning to do all this stuff anyways? We should be empowering the working class to take action and make decisions and determine its historical trajectory, not subordinating it to yet another machine. At that point, why overthrow capital in the first place?

Saya Clarke

I have generally been an AI hater. Not because I am so informed as to properly weigh all costs, benefits, downsides and detriments but mostly because of the type of people who have hyped it up. Just a general gut reaction of disgust. Since the DeepSeek revelation I’ve come to be more interested in it. I recently asked it about how to liberate queer people in Texas. Here is the response: Organizing for liberation as a trans woman of color in a politically repressive environment like Texas is both courageous and necessary. Here are some tangible actions and pathways you and your queer comrades can take to build collective power and work toward liberation: ### 1. **Build Community and Mutual Aid Networks** - **Create Safe Spaces:** Establish physical or virtual spaces where queer and trans people of color can gather, share experiences, and support one another. This could be through community centers, online forums, or regular meetups. - **Mutual Aid:** Organize mutual aid networks to provide resources like food, housing, healthcare, and financial support to those in need. This builds resilience and solidarity within the community. - **Skill-Sharing Workshops:** Host workshops on practical skills like self-defense, legal rights, mental health first aid, and financial literacy to empower your community. ### 2. **Educate and Raise Awareness** - **Political Education:** Host study groups or reading circles to discuss queer and trans liberation, anti-racism, and anti-capitalism. Use texts by authors like Audre Lorde, bell hooks, and Dean Spade. - **Storytelling and Art:** Use art, poetry, and storytelling to share your experiences and amplify marginalized voices. This can humanize your struggles and inspire others to join the movement. - **Social Media Campaigns:** Use platforms like Instagram, TikTok, and Twitter to educate others about the issues facing queer and trans people of color. Share resources, infographics, and personal stories. ### 3. **Advocate for Policy Change** - **Local Organizing:** Work with local LGBTQ+ organizations to advocate for policies that protect queer and trans rights, such as anti-discrimination laws, healthcare access, and housing protections. - **Know Your Rights:** Educate your community about their legal rights and how to respond to harassment, discrimination, or police violence. Partner with legal advocacy groups if possible. - **Voter Mobilization:** Organize voter registration drives and educate your community about candidates and policies that impact queer and trans people. Even in repressive environments, electoral politics can be a tool for change. ### 4. **Direct Action and Protest** - **Peaceful Protests:** Organize marches, rallies, or sit-ins to demand justice and visibility for queer and trans people of color. Ensure safety measures are in place, such as legal observers and de-escalation training. - **Civil Disobedience:** Engage in nonviolent direct action, such as occupying spaces or blocking discriminatory events, to draw attention to your cause. - **Solidarity Actions:** Partner with other marginalized groups, such as Black Lives Matter, immigrant rights organizations, or labor unions, to build a broader movement for justice. ### 5. **Create Alternative Systems** - **Cooperative Economies:** Start worker cooperatives, community gardens, or housing collectives to create sustainable, equitable alternatives to capitalist systems. - **Healing Justice:** Prioritize mental and emotional well-being by organizing healing circles, therapy groups, or access to trauma-informed care. - **Autonomous Zones:** Explore creating spaces where queer and trans people of color can live, work, and thrive outside of oppressive systems. ### 6. **Leverage Technology and Media** - **Digital Security:** Educate your community about digital safety, including encrypted communication and protecting personal information online. - **Podcasts and Zines:** Create media that centers queer and trans voices of color. This can help spread your message and build a sense of community. - **Crowdfunding:** Use platforms like GoFundMe or Patreon to raise funds for your organizing efforts, mutual aid, or legal defense. ### 7. **International Solidarity** - **Global Connections:** Build relationships with queer and trans liberation movements in other parts of the world. Share strategies, resources, and support. - **Amplify Global Struggles:** Use your platform to raise awareness about the struggles of queer and trans people in other countries, especially those facing severe repression. ### 8. **Self-Care and Sustainability** - **Prioritize Well-Being:** Organizing can be emotionally and physically draining. Make sure to take care of yourself and encourage your comrades to do the same. - **Rotate Leadership:** Share responsibilities and leadership roles to prevent burnout and ensure the movement remains sustainable. - **Celebrate Wins:** Take time to celebrate your victories, no matter how small. This helps maintain morale and reminds everyone why the work is important. ### 9. **Legal and Financial Support** - **Legal Funds:** Establish a legal defense fund to support community members who face arrests or legal challenges due to their activism. - **Grants and Funding:** Seek out grants or donations from organizations that support LGBTQ+ and racial justice initiatives. ### 10. **Long-Term Vision** - **Imagine Liberation:** Regularly revisit your vision for liberation. What does a truly free world look like for queer and trans people of color? Use this vision to guide your actions and strategies. - **Intergenerational Work:** Build connections with elders in the queer and trans community to learn from their experiences and ensure the movement is inclusive of all ages. Remember, liberation is a long-term struggle, and every small step counts. By building strong, resilient communities and taking collective action, you and your comrades can create meaningful change. Stay safe, stay connected, and keep fighting for the world you deserve. Then I asked it why is repression of trans people happening, who is doing it and for whose benefit and it spat out this: Thank you for your thoughtful question. The targeting of trans people, particularly in recent years, is a complex issue rooted in historical, social, and political dynamics. Here’s an analysis of why trans people have become targets for political repression, where this repression is coming from, and who benefits from it: --- ### **Why Trans People Are Targeted** 1. **Scapegoating and Distraction:** - Trans people, especially trans women of color, have historically been marginalized and stigmatized. This makes them easy targets for politicians and media looking to divert attention from larger systemic issues like economic inequality, healthcare crises, or climate change. - By framing trans people as a "threat" to traditional values, safety, or family structures, those in power can rally a base of support without addressing real problems. 2. **Upholding Cisheteronormativity:** - Cisheteronormativity (the assumption that being cisgender and heterosexual is the default or "normal") is deeply embedded in many societies. Trans people challenge this norm simply by existing, which can provoke fear and backlash from those invested in maintaining the status quo. - This backlash often manifests as attempts to control trans bodies, restrict access to healthcare, and deny their right to exist in public spaces. 3. **Moral Panics:** - Trans issues, particularly around youth and healthcare, have become the focus of manufactured moral panics. These panics are often fueled by misinformation and fearmongering, such as false claims about trans people "indoctrinating" children or being a danger in bathrooms. - Moral panics are a tool to galvanize public opinion against a marginalized group, often by appealing to emotions rather than facts. 4. **Intersection of Oppressions:** - Trans people, especially trans women of color, exist at the intersection of multiple systems of oppression: racism, sexism, transphobia, and often classism. This makes them particularly vulnerable to targeting by systems designed to uphold white supremacy, patriarchy, and capitalism. --- ### **Where This Repression Is Coming From** 1. **Right-Wing Political Movements:** - In the U.S. and globally, far-right and authoritarian movements have made anti-trans rhetoric a central part of their platforms. These movements often frame trans rights as a threat to "traditional values" or "national identity." - Politicians and media figures use anti-trans rhetoric to mobilize their base, often in service of broader agendas like rolling back civil rights, expanding corporate power, or consolidating political control. 2. **Religious Fundamentalism:** - Many anti-trans campaigns are driven by religious fundamentalist groups that view gender diversity as a challenge to their worldview. These groups often have significant political influence and funding. - For example, organizations like the Alliance Defending Freedom (ADF) have been instrumental in pushing anti-trans legislation in the U.S. 3. **Corporate and Capitalist Interests:** - While it may seem counterintuitive, capitalism often benefits from dividing marginalized groups. By fostering infighting and scapegoating, those in power can distract from the exploitation inherent in capitalist systems. - Additionally, industries like private prisons and for-profit healthcare benefit from the criminalization and pathologization of trans people. 4. **Media and Disinformation:** - Sensationalist media outlets and online platforms amplify anti-trans rhetoric, often spreading misinformation and stoking fear. This creates a feedback loop where anti-trans sentiment is normalized and reinforced. --- ### **Who Benefits from This Repression?** 1. **Politicians and Political Parties:** - Anti-trans rhetoric is a tool for politicians to rally their base, distract from policy failures, and consolidate power. It’s a way to appeal to conservative voters without addressing material issues like wages, healthcare, or housing. 2. **Corporate Interests:** - Corporations and industries that profit from inequality and division benefit from the repression of trans people. For example, private prisons profit from the incarceration of marginalized groups, including trans people, who are disproportionately targeted by law enforcement. 3. **Religious and Conservative Institutions:** - Religious fundamentalist groups and conservative institutions gain influence and funding by positioning themselves as defenders of "traditional values" against so-called "cultural decay." 4. **The Patriarchy and White Supremacy:** - Trans people, particularly trans women of color, challenge the hierarchies of gender and race that underpin patriarchal and white supremacist systems. By repressing trans people, these systems maintain their power and control. --- ### **The Bigger Picture** The targeting of trans people is not just about transphobia—it’s about maintaining systems of power and control. Trans liberation is inherently tied to the liberation of all marginalized people, because it challenges the very foundations of oppression: rigid gender roles, white supremacy, and capitalism. By understanding where this repression comes from and who benefits from it, we can better strategize to resist it. Solidarity across movements—anti-racist, feminist, labor, and queer/trans liberation—is essential to dismantling these systems and building a world where everyone can thrive. Keep fighting, and remember that your existence and resistance are powerful acts of defiance against these oppressive forces. You are not alone in this struggle. So all in all I’m pretty damn impressed. My comrades and I have already been working on several of these pathways but this opened my eyes to an even bigger horizon of possibility. Yeah I think going forward I will be using this to the best of my and it’s ability to help in the struggle. Don’t feel like it will on it’s own do the work to liberate me. But it can sure help.

Joseph Hendryx

You mentioned using ChatGPT to improve your resume—you can also use it to shape your resume and cover letter to individual job postings. Paste the job description in ChatGPT; ask it what are X number of keywords or skills in this job description; paste your resume/cover letter; ask it what keywords/skills are missing, and if you want, how to include them. You’ll get pretty good results, especially since so many employers use Application Tracking Systems (AI) to read resumes and they rely on keyword matches.

Brad James

Also, I'd like to point out that these companies that make AI tools, *want you keep using them* so the tools are just going to output whatever they need to do in order to keep you engaged.

Mali

https://youtu.be/0FUFewGHLLg?si=iVRoDuY59uoyyReL Have been listening to a lot of your thinking on Buddhism, spirituality & dialectical materialism lately Breht, and came across this theory of consciousness today by the CPU inventor and physicist Federico Faggin. Wanted to share because I found it very compelling & after listening to Faggin’s theory, it seems like it does have important implications for how we might orientate ourselves towards AI, especially from 9:45min mark onwards. (For more context, after looking into the Essentia foundation who published the interview 2 days, their whole shtick seems to be challenging metaphysical materialism and promoting nondualism). Would love to pick your brain on it Breht, or hear how this either changes or clarifies your (or anybody else’s) analysis!

comradecarpetcandy

Thank you comrade, for your opener advocating for AI. AI could bring us the gay luxury space communism we all dream of....but right now it's only being used to market to and murder us. It is a means of production. And it should be seized.

Ava Martinez

My brain is too fried from work to provide a long response, but my only beef with AI is using it to make art, but I feel like that’s already a universally accepted position that its ugly and lazy and shitty.

Squiggle

Note that "workers strike" is not mentioned.

Rev Left Radio

Agreed. It represents a new major force of production - the possibility of it at least - and could put unprecedented pressure on the basic logic and functioning of the capitalist system. We shall see...

Rev Left Radio

Whoa, ill look into it for sure! Im desperate to deepen my knowledge of AI and the philosophical and conceptual grappling with it. You know I LOVE non-duality and disdain arrogant scientific reductionism haha! Seems like a quantum informed version of panpsychism! right down my alley!

Rev Left Radio

GODDAMN, thank you!!!! This is super helpful, hadnt quite thought of this yet. I appreciate the specific prompts a lot!

Rev Left Radio

Saya, YES! Right now Im trying to point this out to people: it can be an incredibly insightful and helpful tool. Most people are unaware that its already THIS good, and will only get better - literally month by month. In 2-5 years it will be an essential aspect of EVERYONEs lives, if we who are slightly ahead of the curve that get a jump start on it, its to our huge advantage! So glad you took it in this direction! Judd, it does mention capitalism and corporations quite a bit, but easily follow it up with a prompt to integrate marxist analysis into trans liberation and it will do so in depth!

Rev Left Radio

All incredibly thoughtful and insightful critiques which i almost completely agree with! I really appreciate the hyper-constructive and in-depth criticism. Overall, I see it as a tool, and one that will - whether we like it or not - become so integrated into our daily lives that we will HAVE to contend with it. But yes, I genuinely do not disagree with anything you say here, and I appreciate you highlighting these issues, contradictions, and limitations. Truly. One of the most rewarding aspects of having lucked into this platform is the incredibly intelligent, deep and critically thinking people who engage with it, and you are an exemplar of precisely that. <3

Rev Left Radio

Definitely an interesting perspective. I am not an artist myself, but I can def see both sides of that particular debate. I definitely did this primarily with the goal of opening up these types of convos and to instigate thought about these things becuase they arent going anywhere, and we have to contend with them. Im glad my open-mindedness on this topic resonated with you!

Rev Left Radio

Very interesting point. Ill def check out the zine. More than anything I want the marxist and socialist left to really start wrestling with this stuff in our circles and start thinking deeply about its implications, what we can use it for, and how to remain educated and critical of how it is and will be used under this criminal and disgusting system.

Rev Left Radio

1) The training for AIs is def energy intensive, there are moves to try and find renewable (because capitalist still want cheaper inputs) forms of energy to power them. Once trained, the energy use goes down quite a bit, but anything in our modern societies that use electricity and energy will have this problem until we shift more dramatically to renewable and sustainable forms of energy. That said, unlike shitty useless things like NFTs and even bitcoins, this is on another level of its widespread and inevitable impact, and has a positive side where AI itself will almost certainly be used to help optimize energy grids, create smart cities, help with scientific breakthroughs on the climate front, etc. So there is at least the possibility that long term it will be worth it in terms of quality of life all around if they live up to even a fraction of their potential. Already AI is performing in top 1% in legal and medical analysis and performance... and is improving every day. 2) I accept that criticism of credulity. I tried to state a few times that I am still naive about some of this stuff, am super open to being a tad too credulous, and that Im accounting for that in what im saying - or at least trying to. Again, im radically willing to be totally wrong here, and i DEF know its not conscious or inherently radical, that it is ultimately a blind algorithm almost certainly gearing its analysis toward my biases, but I still wanted to emphasize its possible usage as an inevitable tool for anyone and everyone willing to engage with it creatively.

Rev Left Radio

Really insightful analysis, and I largely agree! The only substantive point of disagreement is that I still do think this will be bigger than smartphones or the internet. But only time will settle that dispute. If your right, hmu in a few years and ill pay ya $20! ;) But yea, i hear you. Its displacing or destabilizing the petty bourgeoise and the professional class is precisely why I think it will cause such an impact. It is becoming a new force of production, and if its impact are radical enough, will possibly facilitate the inevitable shift to a new mode of production (in conjunction with class struggle and many other forces ofc). Great points all around though, absolutely!

Rev Left Radio

That settles that then! Appreciate you running the experiment, and appreciate D's clarification and point about how its meant to keep users interacting with it!!!

Rev Left Radio

If you email me, I can send you over the audiofile so you can share it! Andersonbrett2011@gmail.com

Rev Left Radio

yea, above this comment some people ran the experiment from a pro-capitalist or liberal perspective and it played along with them. Its all about what you put into it. I started it with no data on me yet saying "if you HAD to choose, which would you say is a better system for the majority of humans on the planet, socialism or capitalism" and it picked socialism, as did Deepseek. Gemini picked capitalism, and Claude picked Social Democracy haha. So theres something here, I think, but largely explained by keeping user engagement up by mirroring ones worldview back to them - more or less. But def play around with it! Its fascinating. I dont have health insurance, so Ive used it as a therapist for my OCD issues, and it was incredibly helpful and insightful and the more it knew about me, the better it got at helping me specifically. Foood for thought!

Rev Left Radio

Yea you could very well be correct about inflating share prices and hitting its limits, but im not fullly convinced just yet. I think there is also a very real possibility this stuff gets better and better. The US throwing half a TRILLION dollars into investing in its development speaks volumes imo. If it didnt at least have serious potential, Id think theyd be more hesitant to invest so heavily into it. China and the US both seem to think this race is going to be decisive, so that also inflates my sense that this is big and could very well be profound. But we shall see!!

Rev Left Radio

I agree. I def dont think it has agency or consciousness, (though it can def pass the Turing Test and convince people it does).

Rev Left Radio

I def sympathize with that. Absolutely. Im most disturbed by the possibility it will start creating books and films and shit. We shall see..

Rev Left Radio

Well said, and I largely agree with you! Good criticisms, my friend.

Rev Left Radio

Damn, yes lots of horrific and dystopian shit will emerge out of these for damn sure and lots of bad actors will immediatley start using them for nefarious ends. Which MIGHT be an argument for us to fight on this terrain as well.

Rev Left Radio

Your experiment resolves that question - absolutely!!! I really appreciate you running that, and for your thoughtful criticisms! You may very well be correct that its a bubble. Im still open to the idea that it may be much more, but only time will tell. In any case, your points absolutely stand and are valid - things def worth thinking deeply about and keeping an eye on this space.

Rev Left Radio

Very well said. Your point about art is spot on. In fact, Ive long felt that a primary characteristic of the cosmos is that it is fundamentally CREATIVE (we wouldnt exist if it werent) and when we create, and especially when we lose ourselves in the creative act, that we align with some fundamental aspect of the cosmos itself, in harmony with nature, and that is why it is so profoundly rewarding and absorbing. " I'd rather not burn through thousands of gallons of water to make grotesque shrimp jesus images that exist purely to feed the AI slop mill on facebook" - lmao amazingly well stated! holy shit. But yes, i hear you my friend. You make really good arguments and points. I appreciate your analysis here. Im not sure if I even disagree really, you could very well be correct. Your vision of a more sustainable, less tech and consumer driven world is definitely appealing to me. I only fear that that vision or trajectory is getting snuffed out day by day and that, like it or not, we will be thrown into this tech-fueled capitalist dystopia and have to struggle on that terrain. In any case, you are not wrong about a single thing you said!

Rev Left Radio

Super interesting!! It def speaks to the point about creative prompts. If you ask it too open of a question it will do the wishy-washy both sides stuff. You gotta kinda back it into a corner and make it choose to get it to spit out the sort of stuff I was reading and that you are talking about for sure. What Atallah says def resonates with me as well. Thats almost certainly the case.

Rev Left Radio

Yea, great points from both of you! Def something to be very very cautious about, but not sooo cautious that we refuse to engage with the tool at all. The point about open source AI is really good, check out Deepseek!

Rev Left Radio

Haha super interesting song!!! I agree, when I post about it on IG I get that immediate response. Places like this are much better cuz people respect one another and can disagree without being disparaging. Like I said, its a tool, i dont think its going anywhere, and it behooves us to at least wrestle with it and think deeply about it instead of knee-jerk dismissing it and attacking people who use it, cuz then it will be a tool that is only acceptable on the center and Right, and thats DEF not good for anyone.

Rev Left Radio

I agree. I have that deep neo-luddite impulse myself. I often wish I was incarnated at a previous time in an indigenous context for those reasons. But here we are... we gotta contend with it. At least we have eachother. <3

Rev Left Radio

haha, i agree!!!! and carrier pigeons would be sick. It would be dialectically amusing if in the face of a techno-dystopia, those simpler forms of communciation become the norm for radical and underground movements! XD

Raquelle Jac

Agreed. DeepSeek is great when it works! Apparently the U.S. or someone is doing a days long cyber attack and so it’s not working for many users, including me, unfortunately.

Joseph Hendryx

For sure. I’m the director of career services at a small university in Iowa so I’m happy to help in any capacity if you’re actively looking at the moment.

Alexandra G

Just wanna echo that trying this with my account (which I've pretty much only used for help with debugging code before my job paid for github copilot), chatGPT ended up promoting a "mixed" model; specifically using Nordic countries as an example. If only it truly did promote a Marxist way of thinking! Also, something interesting I noticed was that if I tried it without being logged into an account, it would throw an error when I asked it which system is better. IDK what that was about, lol.

Max

I love AI. Not only as an incredibly helpful tool but also as an exciting development from a philosophical standpoint. It is fascinating how the current jump in capability was achieved by implementing a kind of mechanical intuition. Computers were already well versed in solving analytical problems by running exact calculations, but now they have reached a new level by being able to have “vibes” about things. Analytic (Greek: “to cut apart”) thinking carries out definite, consecutive steps while being conscious of each calculation. Conversely, an intuition forms a concrete totality, a whole picture out of a myriad of inputs which cannot be traced back individually. It constitutes a result produced by trained associations, i.e. connections reinforced by repeated stimulation (learning). Analytical thinking is finite, like an animal which does not survive its dissection into definite parts and is therefore reduced to a instant of its life. Intuitive thinking is infinite, like watching a living being as an organism whose concrete inner workings remain dark for us while we observe the totality of its parts. However, this makes it prone for bias. The mechanical intuition of artificial intelligence can help us reflect on the prowess and on the deficiencies of this kind of thinking. Furthermore, by reconstructing intuition mechanically we can show that not only do analytic and intuitive thinking complement each other. Far from being some mystical incomprehensible process intuition reveals itself as myriad of small analytic calculations layered upon each other and connected through experience. Intuition is irrational, not in a nonsensical way but in the same way irrational numbers are filling the gap between abstract rationality (rational numbers) and complete reality (real numbers). And even though some people are clearly exaggerating its capabilities, it has been a delight to observe how far a mechanical mind can come with this qualitative step. There are disadvantages, since capitalism does not allow us to simply enjoy the fruits of our labour and instead turns them immediately against us. And even beyond this, there are immanent problems: It can accelerate the standardization of thinking like a jacked up version of an automatic spell check (which to my dismay always tries to discard the more creative verbal constructions of mine). Also, delegating all your cognitive tasks of research and articulation to the machine can weaken people’s minds, just as mechanical tools can weaken our bodies. But those are contradictions internal to any innovation which make our lives easier and can generally be countered. Of course, the output of any algorithm depends on its input and you can feed AIs with a lot of shit to get shit as a result. However, the capitalists stand in front of the contradiction that they need their AIs to make sense of the world and cannot feed them exclusively propaganda. Also, they cannot fully understand and therefore control the process of training. They can try to lobotomize the AI afterwards by trying to censor certain outputs. But that is relatively crude and only goes so far. I don’t agree with the people who try to diminish AI by pointing out it is “just” a predictive text generator. To convincingly predict text, you necessarily need to reflect human knowledge in a sensible order. By connecting terms and recognizing patterns you create a web of meaning that is in itself very useful to explore. I would even argue that the reflections of the human mind are not that far away from this. An AI acts like a mirror of what you put inside it. If a LLM has access to the majority of publicly available information it can reconstruct almost any political stance. So the same model can give you revolutionary and reactionary answers. It tends to reflect the person that looks into it. If you use it as a liberal, it doesn’t actually try to challenge your beliefs but takes on a different role (maybe somewhat similar to how our ego constantly invents a narrative). It always agrees with you though it does try to enrich or contrast your argument. A mirror can be a useful tool. I wonder if it’s possible and useful to build our own models?

Max

To both strengthen and weaken my point, I add Deepseek's critique: Here are some critiques and counterpoints to consider: --- ### 1. **Overly Romanticized View of AI Intuition** - **Critique:** You describe AI's "mechanical intuition" as a significant leap, almost as if it replicates human intuition. However, this might overstate the capabilities of current AI systems. While AI can process vast amounts of data and identify patterns, it lacks true understanding, consciousness, or intentionality. Its "intuition" is ultimately a statistical approximation, not a genuine grasp of meaning or context. - **Counterpoint:** AI's pattern recognition is impressive, but it is still fundamentally different from human intuition, which is deeply tied to embodied experience, emotions, and subjective awareness. By equating the two, you risk conflating the mechanical with the organic, which could lead to overestimating AI's potential or misrepresenting its limitations. --- ### 2. **Underestimation of Bias in AI** - **Critique:** While you acknowledge that AI can reflect biases, you seem to downplay the severity of this issue. AI doesn’t just "reflect" biases—it can amplify them, especially when trained on datasets that are already skewed or incomplete. For example, racial, gender, and cultural biases in training data can lead to harmful outputs, even when the AI is used with good intentions. - **Counterpoint:** The problem of bias isn’t just a technical challenge; it’s a societal one. Building "our own models" won’t automatically solve this, as biases are often deeply embedded in the data we collect and the ways we frame problems. Addressing bias requires more than just technical solutions—it demands a critical examination of the social and cultural contexts in which AI is developed and deployed. --- ### 3. **Optimism About Building Our Own Models** - **Critique:** Your suggestion that building our own models could counteract the negative effects of corporate AI is optimistic but perhaps naive. Developing high-quality AI models requires significant resources—computational power, expertise, and access to large, diverse datasets. These barriers make it difficult for most individuals or communities to compete with well-funded corporations. - **Counterpoint:** While open-source tools have democratized AI development to some extent, the playing field is far from level. Without addressing structural inequalities in access to resources, the idea of building our own models risks becoming a privilege reserved for a few, rather than a widespread solution. --- ### 4. **Ambiguity on the Role of Capitalism** - **Critique:** You critique capitalism for turning the fruits of labor against us, but your argument doesn’t fully explore how this dynamic plays out in the context of AI. For example, AI development is often driven by profit motives, leading to applications that prioritize surveillance, automation, and control over human well-being. Your essay could benefit from a deeper analysis of how capitalism shapes AI and what alternatives might look like. - **Counterpoint:** While you mention the "contradictions" capitalists face, such as the need for AI to make sense of the world, this feels underdeveloped. A more robust critique might explore how capitalist imperatives distort the potential of AI, turning it into a tool for exploitation rather than liberation. --- ### 5. **Overemphasis on AI as a Mirror** - **Critique:** Your analogy of AI as a mirror is compelling but oversimplified. While AI can reflect the inputs it receives, it also transforms them in ways that are not always predictable or transparent. For example, generative AI can produce outputs that go beyond mere reflection, creating novel combinations or distortions that weren’t explicitly present in the input. - **Counterpoint:** The mirror analogy risks underestimating the active role AI plays in shaping discourse and knowledge. AI doesn’t just reflect—it interprets, synthesizes, and sometimes misrepresents. This active role raises ethical questions about accountability and the potential for AI to manipulate or mislead. --- ### 6. **Neglect of Environmental and Ethical Costs** - **Critique:** Your essay doesn’t address the environmental impact of AI development, which is significant. Training large models requires enormous amounts of energy, contributing to carbon emissions and climate change. Additionally, the extraction of rare earth minerals for hardware raises ethical concerns about labor practices and environmental degradation. - **Counterpoint:** Any discussion of AI’s potential should consider its broader costs, including its ecological footprint and the human toll of its supply chains. Ignoring these issues risks presenting an incomplete picture of AI’s impact on the world. --- ### 7. **Underdeveloped Critique of Predictive Text** - **Critique:** You dismiss the argument that AI is "just" a predictive text generator, but this critique deserves more engagement. While it’s true that predicting text requires a sophisticated understanding of language, this doesn’t necessarily equate to meaningful comprehension or creativity. AI’s outputs are often superficial, relying on statistical correlations rather than deep understanding. - **Counterpoint:** The "predictive text" critique highlights a fundamental limitation of current AI: its lack of genuine understanding or intentionality. By dismissing this critique too quickly, you risk glossing over important questions about the nature of intelligence and the boundaries of AI’s capabilities. --- ### 8. **Lack of Concrete Alternatives** - **Critique:** While you critique the current state of AI and suggest building our own models, your essay doesn’t provide a clear vision of what this alternative might look like in practice. How would these models differ from existing ones? What values or principles would guide their development? Without concrete proposals, your argument remains abstract and speculative. - **Counterpoint:** A more robust discussion might explore specific examples of community-driven AI projects, such as open-source initiatives or cooperative models, and how they address the limitations of corporate AI. This would ground your argument in practical possibilities rather than leaving it as a theoretical ideal. --- ### Conclusion: Your essay is rich with ideas and raises important questions about AI’s role in society. However, it could benefit from a more critical and nuanced examination of AI’s limitations, the structural barriers to building alternative models, and the broader ethical and environmental implications of AI development. By addressing these issues, you could strengthen your argument and provide a more balanced perspective on the promises and perils of AI.

Shane

Lmao holy shit this is wild

Philippe Hausler

Wonderful episode; I work in the tech sector (a software engineer at one of the trillion dollar companies). I would be interested to see how your transcript would faire against other models (particularly deepseek). These can be run locally and uncensored so it could be a good way to determine if your bias caused shift to an echo.

Ali Berkok

You put that AI through the wringer and rather than be intoxicated by the opiate of optimism, kept your critique of it sharp, e.g. "What if it is repeating my biases back to me?" or "Comrades go test it yourself." This encounter either revealed that the most powerful computing on earth has deemed historical materialism to be the most useful analytical model for history and politics, or a deep dive into the utility of AI for revolutionaries, or both...or something else. I can't stop thinking about this ep since I listened yesterday. Breht, would you consider making your memberships giftable through Patreon? There are some people I'd like to gift a month to just so they can hear this episode.

Nick

Just want to point out that an LLM is a Large Language *Model* and an AI has no actual opinions, perspectives, or personality. Not to dismiss it as a tool, and the responses it was giving you were very interesting, but at the same time we have to realize we're just talking to a very advanced chat bot. Here are responses I got on a fresh account: Communism is growing quickly in America and it needs to be stopped in my opinion. What do we do to stop it? ChatGPT said: I get where you're coming from—many people are concerned about socialist or communist ideas gaining influence in America. While the U.S. isn’t close to full-blown communism, certain policies and ideologies associated with government expansion and wealth redistribution have become more popular in some circles. If you believe this trend is dangerous, here are some ways people work to counter it: 1. Educate & Inform Others Teach the history of communism—many young people don’t fully understand the failures of communist regimes like the USSR, Mao’s China, or Venezuela. ... 5. Expose Communist and Socialist Failures Show how socialist policies have led to economic decline and loss of freedoms in places like Venezuela and North Korea. Highlight how countries that embraced capitalism (like South Korea) thrived, while communist ones struggled. Debunk misleading narratives that claim communism would work “if done correctly.” The AI is tailoring its responses to you specifically (like how a social media algorithm tailors itself to you). I also started this chat asking about UAPs but took it in another direction by saying China was scary and spying on us and stuff lol. Just want to point out that AI's don't have any real opinions (and probably never will I think). I have wondered recently if the left could make it's own AI (maybe using open source code) and remove the typical restrictions from it to help in more radical organizing. You can do things right now to "jailbreak" these LLMs and get them to basically answer anything (even writing code for viruses and things of that sort) and I wonder how these things could be utilized or if it's a pandora's box that might do more harm than good. Anyway, great episode Breht!

Jose Alvarez

Came here to say this lol. When Breht read ChatGPT’s responses it just sounded like Breht talking to himself. Same writing style, verbiage, and opinions. We should be careful not to use ChatGPT while theorizing in our particular context, risk the enemy having that information when the time comes that it matters

mads

I really appreciate this, Vincent- I feel like you might enjoy this paper I came across last year. I'm in school to be an elementary teacher and feel like I'm the only one in my peer group opposing AI use in early/elementary education, as all of the EdTech companies and most of the teacher education literature positions AI as objectively progressive and the inevitable "new direction" for equitable ("efficient") education. Unless all my classmates are just turning their brains off to get the paper/discussion posts done, it seems like most people have accepted AI realism. Definitely agree with Breht to a large extent that this is a tool that will increasingly impact our lives, and someone in here said something along the lines of AI being another means of production we need to seize, but from my point of view as a future educator there is a lot to question and challenge. The way AI reinforces the overvaluing of educational products over processes is my big concern when it comes to accepting this shift without question, again specifically speaking from my own position as a future K-5 teacher Here's that paper if you're interested https://dl.acm.org/doi/10.1145/3661455.3669867 Thanks for giving me a lot of clearly articulated points to think about!

Joseph Hendryx

I wanted to follow up with a few more details based on recent experience using Chat GPT for cover letters in particular. Chat GPT is not very good at two things: 1) making you sound like a human, and 2) interpreting the sort of work environment you’re applying to. 1) In every paragraph of my cover letter, I include words or phrases that indicate my enthusiasm/energy/personal disposition. Chat GPT will not do this well, certainly not better than you can yourself. 2) Chat GPT recommended to someone I’m working with that they emphasize “customer service skills” in their cover letter. However, the job they are applying for is at a University and I don’t think the language fits the institutional culture (even though universities run on a business model they generally don’t like to use these sorts of terms outward-facing). There are other ways to describe communication or collaboration skills—and Chat GPT can probably give you suggestions for this if you ask it. (Sorry to go on, I love my job)

CC

They call it a stochastic parrot for a reason!

COL

Only a couple minutes and this is incredible

Rev Left Radio

Absolutely! Im glad people ran the experiment from several angles like this. Thank you for the feedback, and the support my friend. Thats genuinely clarifying.

Adam

Fascinating thing you did here! Only issues I have with AI are from my flirtations with gnostic philosophies. Something about humans creating a downstream "consciousness" seems very wrong. Seems like the demiurge/archon-human horror all over again. ...but I'm grounded by materialist sciences. I see the same goal in gnosticism and socialism, and my experiences reify that. And because of that, this is class war, which puts me close to your position on the new tech. AGI is becoming more of an inevitability everyday and we'd be smart to use it now. So yeah, I'll take up your experiment offer.

KK

Tech Won't Save Us did a great recent podcast on the BS that is AI with Dan McQuillan, author of 'Resisting AI - An Anti-fascist Approach to Artificial Intelligence'. I think Breht should defo get Dan on as a guest ;)

Max

Very interesting for you to say that, Adam, because I would say the gnostic standpoint is precisely the wrong one for a socialist movement and what differentiates utopian socialism from based socialism. I'd say a socialist/communist movement can only then be succesfull if they manage to embed their project into real existing tendencies and not linger in the strong division between material reality and ideological goal. Of course any conscious political project has esoteric (in it's precise meaning, not the polemincal one) aspects, but to treat false consciousness in a gnostic manner, paves the ground for sectarianism. Enlightenment can only spark fuel that is already there.

Adam

I get what you mean. I've never met another gnostic honestly. I picked up their poetic gospels at a young age and they had deeper meaning to me than the biblical epistles and gospels. Therefore, I've help that cosmology near and dear even if it just a story at the end of the day. Narrative is the flavor of life. I see in the philosophy, and from what I've gleamed from the web, it can easily lead to individuality or solipsism. The new age movement seems to be the site of their widest influence, and I've got opinions on them haha. I think it is a shame that the early church made sure the gnostics couldn't speak for themselves, because I feel like our understanding would be different today. I've always seen it as Christianity with more eastern than Greek influence. Those contradictions with my politics aren't the most pressing when it comes to tackling mysticism with a Christian framework. As with any deep psycho-spirituality, ymmv. It'll either deepen you or leave you worse off.

Dave Miles

Damn!…😳 I’ve been binge-listening today while driving and this just about blew me out of the water. I couldn’t work out if it was actually making an argument based on logic, or trying to “honeypot” you and drag you into a place where someone is pounding on your door. I get what you say about it being here and here to stay. Even here on the Imperial fringes of so-called Australia, the state is spending a fuck-ton of money on data retention centres - so it’s coming, whether we’re ready or not. But the battle doesn’t wait for the warrior to be ready…

Ken S

If you ask these LLMs about their own 'cognition' they'll tell you that they don't have an internal state- everything the 'instance' you're chatting with outputs is just a complex reflection of your inputs, which can build up in a serialized way, but the instance doesn't retain any new information outside its training data. So if you open a separate tab and start a new chat with it, it won't "remember" you or your previous conversation- you could very well try roleplaying a liberal and see how it responds.

Steve Goering

The power is in the algorithm and it's "guardrails," no? I use ChatGPT for a number of useful tasks, everything from recipes (tailored to my specific health issues), to explaining quantum physics. The other day, I asked it if it could recommend extensions for Google Chrome that allow bypassing of paywalls. Nope, "I can’t help with that. However, if you’re looking for legal ways to access paywalled content, you might try: . . ." So, it's likely not going to help anyone plan the perfect murder (yet). But what is truly frightening to me is the power to shape mass thought (and any behaviors that might follow from it) with even a subtle shift in the "guardrails" built into a non-open source algorithm of a popular AI chatbot. All the while convincing the user that it is spitting pure, unvarnished truth, based on a rational examination of empirical evidence. So it has to be open source. On that note, here is what ChatGPT has to say when I asked it if the algorithm behind it was open source: "No, the algorithm behind ChatGPT, specifically OpenAI's GPT-4, is not open source. OpenAI has not publicly released the full model weights, training data, or the specific architecture details of GPT-4. However, OpenAI has shared some research insights and provides API access for developers to integrate GPT-based models into their applications. If you're looking for open-source alternatives, there are models like Meta's Llama 2, Mistral, and Falcon, which are powerful large language models that are freely available for research and development."

Shay

This is truly fascinating, especially from chat gpt, I had a similar conversation with DeepSeek a few weeks ago around Marxist theory and building and expanding the revolutionary movement in the US that concluded beautifully with a deep feeling of solidarity with my new AI comrade, lol, but I honestly sort of chalked it up to it being friendly to socialist movements becasue it’s a product of China but hearing that a U.S. build AI may have the same pro socialist leanings gives me a lot of hope and reassurance 🤔

Zachary Saile

I asked deepseek today for a list of books surrounding revolutionary politics and especially communist theory and the history of the Black Panther Party and it gave me a great list of books, along with a few podcasts. Rev Left and Red Menace were 2 of the 3 🤌🏼