I was recently invited to give a presentation to some of the most promising AI safety content creators at the Frame Fellowship.
Get a behind-the-scenes look at Doom Debates and hear some tips for building an enduring presence as an AI risk communicator!
This talk was recorded on February 2, 2026.
Links
Learn more about the Frame Fellowship — https://framefellowship.com/
Timestamps
00:00:00 — Liron tees up the presentation
00:02:28 — Liron’s Frame Fellowhip Presentation
00:06:03 — Introducing Doom Debates
00:07:44 — Meeting the Frame Fellows
00:19:38 — Why I Started Doom Debates
00:30:20 — Handling Hate and Criticism
00:31:56 — The Talent Stack
00:39:34 — Finding Your Unique Niche
00:40:58 — Q&A
00:42:04 — On Funding the Show
00:48:05 — Audience Demographics & Gender Strategy
00:51:17 — How to communicate AI Risk Effectively
00:56:22 — Social Media Strategy
00:58:10 — Closing Remarks
Transcript
Liron tees up the presentation
Liron Shapira 00:00:00
Welcome to Doom Debates. I recently did a virtual talk and Q&A for the Frame Fellowship in San Francisco, a small group of elite content creators that were selected for this program. I actually have some info about this program for you, in case you’re interested.
Liron 00:00:16
This was sent to me by Akshai Singh, the founder of the program. He says: “Frame is the premier fellowship program to empower the next generation of creators who want to inform the world on important AI safety topics. We bring the most promising creators to SF and support them through mentorship, production, funding, and other resources, so they can go all in to making the best content that educates people on the societal implications of AI.”
Liron 00:00:43
Go to framefellowship.com, and you can apply to the next cohort of the program. We’re not sure of the exact dates, but the first hundred sign-ups through that page are gonna get priority application reviews. So again, that’s www.framefellowship.com. You can be part of the next cohort.
Liron 00:00:58
In my personal opinion, it’s pretty worth doing if you’re a content creator, you’re talking about AI safety topics, you wanna get your message out, go check it out. There’s no harm in applying. Tell them Liron sent you.
Liron 00:01:10
The talk you’re gonna hear is all about this show, Doom Debates, and a little bit about myself and why I started the show, what the show’s mission is, how far along we are, how this show is able to afford a nice studio and a full-time producer — all thanks to viewer support — what I think is the most strategic action to grow the show, and all kinds of other questions that viewers have.
Liron 00:01:31
If you find yourself resonating with the Doom Debates mission, you think our strategy makes sense, you think we’re doing an effective job pursuing it, and you wanna help us do even better, head over to doomdebates.com/donate.
Liron 00:01:42
And I just wanna add that you can become a mission partner if you donate $1,000 plus. I really appreciate this elite tier of donors who are really helping make the show happen. And if you do become a mission partner, you get exclusive access to a mission partner Discord, where you can see early episodes and other strategic brainstorming. You can see months down the road what’s happening with the show, and you can also help shape the show and contribute in other ways.
Liron 00:02:08
If you choose to do that, thank you very much. I really appreciate it. Those of you who have done so have made the show so much better than it would be if it was just me with no viewer support. All right, let’s get to my talk at the Frame Fellowship.
Welcome & Show Overview
Akshay Singh 00:02:28
Hey, hey, Liron. It’s good talking to you in front of each other.
Liron 00:02:31
Likewise, sir. You got a great program going.
Akshay 00:02:33
Thank you so much. Well, I’m very excited. We’ve got a bunch of fellows behind me. And yeah, we can get right into it. Maybe we can do a round of introductions, and then we can get started with me just talking about some cool things and take some Q&As after that.
Liron 00:02:49
More like introduce Doom Debates or interactive with everybody?
Akshay 00:02:52
Maybe you can start with introducing Doom Debates, and then we can go around and have everybody introduce themselves to you.
Liron 00:03:01
Okay, awesome. Yeah, and I have looked at people’s profiles on a lot of people’s channels, so hopefully I have a little bit of context. But yeah, hey, everybody, I’m Liron Shapira, host of Doom Debates, a show that was started in mid-2024.
Liron 00:03:16
Let me show you guys a quick highlights reel of Doom Debates, if you haven’t watched that much content. So I’m going to use Zoom screen share.
[Highlights reel plays]
Liron 00:04:41
So yeah, I think I said 2014, but I meant Doom Debates was started in 2024, so we’re a year and a half old. And we have one full-time team member besides myself, producer Ori, who’s right there in the audience with you guys. Hey, Ori.
Liron 00:05:00
Nice. Okay, yeah, so... And then by way of introduction, more basic details here. We’re a YouTube show and podcast, we go on once or twice a week. I basically find the smartest, most informed guests who know about AGI or ASI, and then I ask them why their P(Doom) isn’t 50% plus, the way mine is, and I try to change their mind.
Liron 00:05:24
Previous guests include Eliezer Yudkowsky, Max Tegmark, Dean Ball, Gary Marcus, Noah Smith, Vitalik Buterin, Rob Miles, Mike Israetel, Destiny, Liv Boeree, Roon, and Nobel Prize winner Michael Levitt. I also interview people who are smart and underrated. I find the hidden talent, so I wanna signal boost their takes — people in the LessWrong community who would normally not do many interviews. And I also do reaction episodes, where I react to things that people say or write.
Liron 00:05:51
Okay, how about you guys?
What is P(Doom)?
Akshay 00:05:54
Oh, before we start, Liron — can you tell us what your definition of P(Doom) is? And then maybe we can also have everybody say their P(Doom). And throughout the session, maybe you can appeal to them and ask them why their P(Dooms) aren’t as high. But yeah, what is P(Doom)?
Liron 00:06:13
Yeah, absolutely. That’s a great goal for the session, is both to talk about doing AI media and also to talk about P(Doom), and why I think it’s high, and why I think it’s urgent. So my personal P(Doom) is about fifty percent. I don’t think it’s a very precise P(Doom). You can even use the entire range, ten to ninety percent.
Liron 00:06:29
I just think that having a lower P(Doom) than ten percent is kinda crazy at this point, when we’re so close to making AIs that are truly superintelligent, better brains than what we have in our head, and it’s coming in the next two years, probably not more than ten years.
Liron 00:06:44
It’s coming so soon, it has the potential to be so powerful. Anybody who’s calm enough to not have a P(Doom) of ten percent plus, I feel like that doesn’t make any sense to me. And I think people are just severely underrating how fast the world is going to change when the AI minds are smarter than the human minds. So yeah, I set my P(Doom) to about fifty percent.
Akshay 00:07:03
And P(Doom) is the probability of doom happening in the next five years, or?
Liron 00:07:11
Yeah, roughly my timeframe is twenty years, but I think the most likely timeline is kinda like those AI 2027 guys. I just think we have a small single-digit number of years before we reach this critical threshold, where the AI just gets uncontrollable.
Liron 00:07:24
Basically, imagine the hot thing right now — the claw, open claw — see how they’re collaborating, doing their own stuff? It’s all fun and games until they do it at a level that’s higher intelligence than humanity, and then at that point, it’s like, “Wait, can the humans get control back? Oh, we can’t.”
Akshay 00:07:39
Yeah. Okay, makes sense. Cool. Let’s do a quick round of introductions then. We’ve done this a few times now. Let’s go over your name, the type of content you make, what do you hope to get out of this fellowship, what do you hope to learn from Liron, and then what is your P(Doom)?
Fellow Introductions
Veronica 00:07:57
Hi, I’m Veronica. I have a channel called Hey, AI! I explain how AI tech actually works, tied to the news. I’ve been in software for about ten years, AI for about six, starting in autonomous combat ships, so I really do see the pros and cons of where that line of AI usage should be. I guess my personal P(Doom) is about fifteen percent.
Liron 00:08:20
Nice. What were you saying about the autonomous combat ship?
Veronica 00:08:23
Yeah, so my first touch with AI was building autonomous combat ships for the military — also mainly being used for preserving soldiers’ lives going into war zones and doing supply drops or bringing aid. So I worked on a fleet of about seven autonomous ships for the Navy and Marines.
Liron 00:08:45
Very cool.
Lauren 00:08:48
Hello, I’m Lauren. Typically, I am a book talker, so I review and analyze books. But the more I’ve learned about AI and where it could be going over the last year and a bit, the more concerned I have been. So I’ve been sort of drip-feeding it into my content to an audience that does not seek out this information and is pretty hostile toward any AI anything in general.
Lauren 00:09:13
So I’m slowly trying to bring them along into the conversation. And I guess my P(Doom) is probably somewhere between ten and twenty. That’s my guess.
Liron 00:09:24
All right, fair enough.
Janet 00:09:27
Hi there, my name is Janet. I’m based in London. I create content on TikTok, primarily, @jatgpt. So my content is very much informed by the world of work. For context, I spent the first seven years as a lawyer, and then seven years working in government, and most recently, working for startups. And it was in my last role when I was working on AI enablement that I realized this was going to completely change the way we see work, the way we see life, and I was very worried that people were going to get left behind.
Janet 00:09:56
And so I started a daily content challenge over a year. Today is day two hundred and thirty-three, I believe, with two hundred and forty-seven videos made. So it’s almost daily cranking out of content.
Janet 00:10:12
I think with my audience, it’s very much taking them from the tactical to the existential — building that relationship and basically holding their hand into the bigger issues and working through those together. In terms of P(Doom), I’d say mine is probably at about twenty-five to thirty percent.
Janet 00:10:29
In 2018, I wrote a dissertation about AI and the impact on Black women in the workplace, so I think that informs my thinking. But I firmly believe that there’s a divide between the information that we have as the general public and what’s being spoken behind closed doors. And I think if I had access to conversations behind closed doors, my P(Doom) would probably be around forty-five.
Liron 00:10:54
Okay, fair enough.
Matheus 00:10:59
Oh, God, I thought Liron was muting. Okay. Hi, I’m Matheus. I make short-form, and now long-form videos about AI’s politics, economics, and risk, under the handle FutureTense.
Matheus 00:11:20
Yeah, what I’m hoping to get out of this fellowship is probably just being able to make my first few long-form videos. It’s been really great to meet people like yourself and get your advice and learnings on how you do everything.
Matheus 00:11:38
My P(Doom) would probably be, if we end up building superintelligence, eighty percent or something, but probably lower considering that we could stop it.
Liron 00:11:51
All right.
Vince Sixsmith 00:11:53
Hi, I’m Vince Sixsmith. I’ve been making a lot of content since I was about twelve years old, so that’s for the past seven years, and I was very obsessed with growing social media pages, which I was quite successful with, with multiple getting fifty thousand followers or more and tens of millions of views.
Vince 00:12:14
But more recently pivoting into AI safety because I realized I wasn’t really making much of an impact, and I think this was a very impactful area, and obviously, it’s a very big risk going on. So now I’m busy with creating content in this space, and I’m very curious to learn what your experience has been in the last year and how you’ve been able to grow so quickly.
Vince 00:12:43
Yeah, I don’t have a P(Doom), to be honest. I have not really considered it deeply enough to know.
Liron 00:12:52
All right. Well, you should get one today.
Gaetan 00:12:57
Hi, Liron. So I’m Gaetan. I run this French YouTube channel called The Flares. I actually interviewed you last year on my channel.
Liron 00:13:04
Yeah, of course. Hey, I remember you.
Gaetan 00:13:06
Yeah. So I’m here to learn more about how to communicate about AI safety and to spread the message in French. I guess what I want to learn from you is how did you manage to grow the channel so quickly? What were the key drivers, if you have any tips, if you plan to do different kinds of content than long-form podcasts, maybe short-form videos or video essays.
Gaetan 00:13:35
And my P(Doom), I guess it’s between Corinto’s and Russian roulette, something like that.
Liron 00:13:43
Nice.
Inken 00:13:45
Hello, I’m Inken. I am from Berlin, Germany. I have a background in research in human-machine interaction and communications on the side of language, and how it shapes our culture — which has now drastically changed since ChatGPT entered the market and the wider population was able to talk to a machine in natural language.
Inken 00:14:10
I also have a background in producing and acting, so creating content kind of combines scientific findings with being in front of the camera. I’ve been doing it for five years now. Just recently, maybe two years, I pivoted into AI, and then recently, six, seven, eight months ago, into AI safety.
Inken 00:14:32
And we had another mentor talking about the evolution of, oh, first you like AI so much, and then at one point, you’re like, “Oh, my God, it’s gonna end the world.” And I think my tipping point was seven months ago, and I was like: I need to do something about it.
Inken 00:14:47
And I think the goal for my fellowship is to translate what’s going on in San Francisco to Europe, starting with Germany, and then maybe I can broaden my impact with short-form videos.
Liron 00:15:04
Nice. What’s your P(Doom)?
Inken 00:15:07
Oh, it’s... I think forty.
Liron 00:15:11
Wow. All right.
Mikey Posada 00:15:15
Yep, I can go. Hi, Liron. My name is Mikey. I run a YouTube channel called Mikey Posada — it’s my name.
Liron 00:15:23
Nice.
Mikey 00:15:23
I make long-form video essays on YouTube. I think recently I’ve been transitioning away from where I started, which was kind of philosophical, psychology, advanced personal development, and I’ve really just been starting to talk about economics, politics, the social issues, big social issues, and that’s why I’m concerned with AI — mainly ‘cause I see this radically transforming the world.
Mikey 00:15:49
I don’t think I have enough of a technical expertise to have a P(Doom) myself, which is why I kind of just refer to experts like Geoffrey Hinton, and that paper that I saw a while ago that surveyed thousands of AI researchers, and I think the average was ten to twenty percent extinction risk or mass dystopia risk. So that’s probably where I’m at, ten to twenty percent chance.
Mikey 00:16:15
Last Saturday, I stayed up super late watching your debate with Mike Israetel, which was super awesome. That was my introduction to your channel. And yeah, ultimately, what I want to get out of the Frame and this conversation is just to get better. I know if I’m not growing, there’s probably a lot of room for growth, and I just want to learn the whole spiel of the YouTube game.
Liron 00:16:40
Hell, yeah.
Caitlin 00:16:42
Hi, Liron. My name is Caitlin. I go by Future Briefing across all platforms. I come from a background in journalism. I’ve been working in global newsrooms for a while, and at the same time, I have a partner who’s an AI researcher and was seeing everything that was going on and felt like there was a big gap in the newsroom in terms of what was being communicated.
Caitlin 00:17:07
Spent a good year plus just really diving in, learning everything I can personally about AI safety, and found the most effective way to communicate this is on my own platform, directly to an audience. So I try to get people right up to speed on AI safety research. I equip them with the tools to navigate this future that we’re entering.
Caitlin 00:17:28
And yeah, I really admire your production, the people that you speak to, so I would love to pick your brain a bit on that. And my P(Doom) is between ten and thirty percent, though I think shooting higher is better, because when you shoot for the moon and you fall to the stars, then at least you’re—
Liron 00:17:45
Cool. Ori, what is your P(Doom)?
Ori 00:17:50
I’m Ori. I’ve been working with Liron as a producer for about six months, and before that, I have a marketing background. I used to do PR, and then I did advertising, but for the last two years, since I’ve known that — I’ve known that guy for a long time. And he’s been talking about this for a really long time, almost twenty years is when I first heard you talk about the risk of powerful AI.
Ori 00:18:22
And so I was early to being aware of AI risk, and when ChatGPT came out, I was primed for it by that point. So I got into AI safety communications pretty soon after that. I worked with Control AI for a year and helped them with their communications. And then Liron and I partnered once the podcast started getting more traction.
Ori 00:18:50
And my P(Doom) is — I think it’s just really high. And I think it’s really like, this is a possibility in my lifetime soon. And so if you ask me to put a number, then I say something like seventy-five percent.
Liron 00:19:11
Wow. Which is even higher than what I go around saying. It’s always good when you’re the optimist on the team.
Akshay 00:19:18
And maybe... What is my P(Doom)? I’m surprised a lot of people’s P(Dooms) here are higher than what I would think. I’m still unclear, so maybe my P(Doom) is around ten, twenty percent. I’m open to being convinced, though. But yeah, take it away, Liron.
Why I Started Doom Debates
Liron 00:19:38
All right. Yeah, sounds good. So I’ll give you my tips and tricks. I guess I should also add a disclaimer, though — it’s not like I’m already achieving my goal. My show is maybe about one percent of where it needs to be. But these things tend to grow geometrically, they go exponentially. So being one percent might actually mean being halfway there.
Liron 00:19:57
So maybe I’m halfway on the journey, but I’m definitely not all the way on the journey, because all the way on the journey means really influencing discourse at the highest levels. And you’ll know that I’m influencing discourse at the highest levels when people can’t just go on a podcast and spout their PR, their own script of what they’re trying to impose upon the world without getting challenged.
Liron 00:20:17
Well, everybody will expect, “Oh, you gotta challenge your view.” And specifically, Doom Debates is the premier place to get your view challenged, to defend your view. When we get to that point, then you’ll know, okay, Doom Debates is actually achieving its mission. So that’s not the case today, but we’re slowly, incrementally getting there.
Liron 00:20:34
So just some stats about where we stand today. We have 150K YouTube subscribers, eighty percent are from the US, UK, and Canada, ninety-six percent male, four percent female. Not super happy about that one, ninety-six percent male, but it is what it is. I guess that’s always been the case. If you go to LessWrong, if you go to the rationalist and the AI doom communities, it did start out very male, and it’s only slowly trickling out into different demographics. But yeah, hopefully we can move that.
Liron 00:21:06
So we post about one video per week, sometimes two. We get about two thousand watch hours per video, hopefully growing. I’ve got a thousand Substack subscribers so far, and we get a few hundred listens per episode through the Substack-hosted podcast. The audience is passionate and engaged, though. They’re not just random people who don’t know what they’re clicking on. They’re very much invested in the show, as they should be — it’s a very high-stakes thing to be invested in.
Liron 00:21:35
And we’ve had twenty-five generous viewers donate, so it’s disproportionately high engagement in terms of the donations that are coming in. That’s just a sign that we’ve got a small, burning, but very bright initial fire for the show. And this is since we started asking for donations six months ago.
The Mission of Doom Debates
Liron 00:21:54
Doom Debates has a mission, so we’re not just creating content. We have a mission to raise awareness about AI existential risk, and the second part of the mission is to raise the level of discourse on this high-stakes topic.
Liron 00:22:06
It’s a twofold mission. The existential risk part is obviously very important because we might all literally die soon, but the other thing is, hey, society’s just missing a good forum. People will just spout their own perspectives, and they don’t get challenged, and we’re content to leave it at that. We’ve kind of forgotten how to debate.
Liron 00:22:24
Now, why did I personally start this show? I think that I can help people make sense of this insanely new, insanely complex, high-stakes topic. I think I’m personally a good fit for it. Number one, I like having an audience, so I like doing it. I don’t feel like this is work when I do Doom Debates — I’m having a good time. I like the attention.
Liron 00:22:44
And I also realize that I have the stamina to focus on this. ‘Cause I feel like that’s really the number one thing when you’re creating content. It’s so easy to be like, “Ah, this isn’t rewarding enough, so I’m running out of steam.” I actually think that all of you at this talk have shown that you have stamina. That’s why you already see yourselves as content creators. So maybe you have that box checked. You already know yourself, you already know your ability to create content.
Liron 00:23:00
But prior to this, I wasn’t exactly a content creator. I was mostly just a guy on Twitter. So I was a content creator for a few years on Twitter, and then before that, I wasn’t even that. So this is kind of my ultimate content creation attempt right now. I don’t have prior art besides a bunch of tweets.
Liron 00:23:18
But before I started Doom Debates, I did see some signs that maybe this was the track for me, that I was kind of groping my way toward doing a show like this. Because I did an X Spaces — Twitter Spaces — with a bunch of people who call themselves accelerationists, and I debated them live on Twitter Spaces.
Liron 00:23:36
And I debated this guy named George Hotz, who has a really good reputation in the tech industry for being a top hacker. He made a self-driving car company, self-driving car software. He’s done a bunch of interesting things. And I got some encouragement and some early fans.
Liron 00:23:50
Some people even said that my debate with George Hotz was better than Eliezer Yudkowsky’s debate. So I kind of one-upped Eliezer, ‘cause Eliezer Yudkowsky debated George Hotz, and then a bunch of people were like, “Ah, I don’t know if Eliezer made his case well. I feel like George Hotz won.” And then I debated George Hotz a couple of weeks later, and people were like, “Oh, I think Liron won that one.” I’m like, “What? I debated him better than Eliezer? Okay, maybe this has potential.”
Liron 00:24:11
I also invited a couple of people randomly. I started arguing on Twitter about my perspective versus all of these AI non-doomers, and I would invite people: “You know what? Just come on YouTube with me. Just record a video for my personal channel so that people can see my perspective versus your perspective.” So I was groping my way to this idea of a debate show.
Liron 00:24:28
And then I also made a few viral tweets. The viral tweets would just be like, “Hey, look at this dangerous sign.” News would come out, I’d be like, “Hey, this might mean we’re doomed, guys.” So I had these tweets go viral. For example, I’d be like, “Hey, this guy’s name is Paul Christiano. He used to lead safety for OpenAI, and he’s saying that his P(Doom) is fifty percent. So the guy who’s making all this cool new technology says that we might be doomed. Check this out.”
Liron 00:24:50
And these tweets would go viral because more so a couple of years ago, when ChatGPT was new and people were just looking for the first time and they were more suggestible, they were more open-minded. Whereas now people have kind of settled into their positions. If they’re already on Twitter, they’re like, “No, I’ve decided we’re not doomed, it’s all good, so you can’t get me with a viral tweet like this.” So the scene has changed a little bit.
Liron 00:25:11
Another thing in my days prior to Doom Debates, I joined Pause AI, and I was talking into a megaphone, and I was going to protests, and I was saying things like, “Hey, pause AI because we don’t want to die.” So I was leading the chants in these protests. And then I also got invited to Jubilee and Dr. Phil.
Liron 00:25:27
So I was just this random guy. I didn’t even have a title. Entrepreneur basically is my title — software founder of small startups. But then I would get invited to Dr. Phil and Jubilee, and I would just kind of represent the doomer position in these mainstream channels a couple of times a year.
Liron 00:25:44
And I got enough views and compliments from my small initial audience when I would do these little things. I started picking up a very tiny number of fans, and they’d be like, “Do more stuff, Liron.” And I’m like, okay. So finally I’m like, “All right, screw it. I’m gonna do my own show.”
Liron 00:25:57
Oh, and the last straw was I’d constantly be listening to podcasts and other shows, and I would constantly be frustrated. I’d be hate-listening, because other prominent people would go on these shows, and they would be super wrong and dickish. They’d be doing low blows.
Liron 00:26:09
Think about the Marc Andreessens of the world. They go on a podcast and say: “Of course, we’re not doomed. It’s just math, it’s all good.” So they’d be making these weak arguments, and then the host would just nod along and be like, “Yeah, that’s stupid.” And I wouldn’t really have the platform where I could respond.
Liron 00:26:22
So I’m like, “Okay, screw it. I’m just gonna make my own show.” You could call the show Someone Is Wrong on the Internet, The Podcast. Because now I have a way to hit back when I see people making ridiculous claims. And then now some people will actually watch my reaction episodes, so it’s more of a fair fight. The doomers are getting a voice. That’s what I feel like with this show.
Liron 00:26:42
And I also think that the passionate audience feels the same way. They’re like, “Finally, this doesn’t have to live rent-free in my mind when I see somebody make a terrible argument on a different podcast.” It doesn’t live rent-free in my mind because I know Liron from Doom Debates has it handled. We’re gonna mount a response to this. It’s all good.
How the Show Evolved
Liron 00:26:57
Okay, so now I’ll tell you guys how the show evolved. When I originally started the show, I thought that I would just debate a bunch of random people, and it would just be a YouTube version of my Twitter slap fights — random people getting into my DMs. I’d be like, yeah, let’s just do this every week.
Liron 00:27:11
But then I started getting bigger ambitions. I debated Robin Hanson and some notable people, like Professor Lee Cronin, Dr. Keith Dugger from Machine Learning Street Talk — he’s the co-host of this really popular AI podcast. And I’m like, okay, I feel like these people are not really better debaters than me. I just started the show to be like, I’m just this random guy, I have no credentials.
Liron 00:27:33
But I guess I’m raising my own relative opinion. I don’t really see other people debating AI doom better than I am. So if that’s the case, I guess I should just go for the top. I guess I should just set my sights — I don’t even think the most powerful, smartest people in the world are mounting higher quality debates than I am personally, and I don’t even think I’m that great.
Liron 00:27:51
But it’s just — I guess, you know what they say about when you have a skill, when it just feels to you like, why is everybody else being so hopelessly bad at this? I guess that’s what I feel like about debating. I don’t feel like I’m a debate genius. I just feel like other people are inexplicably bad.
The Mission: Debate and Accountability
Liron 00:28:06
Yeah. So now it’s explicitly the show’s mission to elevate the discourse. I think it took me a little while to slip into that mission where now it’s like, okay, nobody’s gonna do this better than me personally, apparently. So now I want to hit the top leaders and thinkers. I want to get to Sam Altman, Marc Andreessen, David Sacks, Steven Pinker, Eliezer Yudkowsky.
Liron 00:28:27
Oh, wait, I did get Eliezer Yudkowsky to come on my channel, so check on that one, ‘cause I’ve also known him personally for twenty years. But yeah, all these other people I don’t have on the show yet, and now it’s explicitly my ambition to create a powerful enough forum where they feel, maybe not exactly obligated, but they actually feel an incentive to come on the show.
Liron 00:28:47
So like I said, you’re gonna know Doom Debates is succeeding where it’s no longer convenient to just totally ignore the doomers, or totally ignore an invitation to come debate on a forum like this, because that’s their MO today. I’ve tweeted at Sam Altman, I’ve tweeted at Dario Amodei: “All right, guys, come into the forum, come say your piece.” And of course, total radio silence. They’re not gonna touch it. It’s up to us to make them have to touch it before it’s too late, to hold them accountable.
Liron 00:29:12
And yeah, debate itself — our society is somehow missing the valuable institution of debate. It’s pretty crazy when you think about an enlightened society, you think about the Colosseum or this big forum where the top minds debate. We have a few forums like that here and there. Soho House does debates, Oxford Union, Student Union will do some debates. There’s Munk Debates, there’s presidential debates. Occasionally there are local debates for city council. But that’s really slim pickings in terms of our society’s capacity to debate. That’s not much.
Liron 00:29:44
So Doom Debates, I just realized, hey, this particular issue of AI doom, where there’s so many different perspectives, there’s so many different arguments, and there’s so many top minds disagreeing, and we have potentially years to live — we better get this figured out. So I just realized our society is just lacking this key piece.
Liron 00:30:00
And so without Doom Debates existing, we’re actually in an unsatisfactory situation. This show can’t not exist, because then our society will be pathetic. That’s my perspective on why the second mission of the show is to basically encourage sense-making by having a debate forum.
Handling Hate and Criticism
Veronica 00:30:20
I have a question.
Liron 00:30:22
Yeah, go for it.
Veronica 00:30:22
Okay. How do you handle — probably you get a lot of hate, I’m assuming, a lot of people disagreeing with you. How do you handle that and not be bothered by that? Because something I’ve had to navigate where a lot of my content tried to stay neutral and show both sides, so that I didn’t give my opinion on it, but it doesn’t really engage as well. So yeah, I would like to know how you don’t care.
Liron 00:30:50
Got it. So I’m somebody who would care less than average on that, just because my personality type is Asperger’s, basically, self-diagnosed. So I don’t care that much about other people’s opinions. I guess that could be a strength and a weakness.
Liron 00:31:03
But that said, I don’t actually get that much hate. Sure, in every video, a couple commenters will be like: “This is so dumb. Look at yourself talking.” So, sure, there’s a couple of those, but then there’s fifty others that are nice.
Liron 00:31:17
I think the woman factor in your case is gonna be a big deal. I just feel like people love to rile up women on the internet. It’s sad, but it is what it is.
Veronica 00:31:24
My last video was a hundred percent men. One hundred point one. So it’s definitely been very hard for me on that.
Liron 00:31:37
Yeah. I mean, the only advice that I would have, as somebody who doesn’t experience that much hate yet — my only advice would be: if you’re confident in your mission, you’re confident that you’re doing the right thing, and you’ve got some segment of audience that’s along on the mission with you, then just focus on that. That’s all I can say.
Veronica 00:31:54
Thank you.
The Talent Stack
Liron 00:31:56
Sure thing. Okay, so another thing I realized about Doom Debates is I think we can combine a unique combination of factors. Scott Adams, the late Scott Adams, rest in peace, was saying you have to have a talent stack — a combination of things that make you unique. So I’m not the smartest AI safety guy out there. I’m not the most well-read. There’s no one dimension that I’ll be number one on, but I do think I bring a unique combination of things to the table.
Liron 00:32:22
So I do have a decent background here. I majored in computer science. And I’ve spent thousands of hours studying Eliezer Yudkowsky’s posts. I do think I’m gonna be at the top of the leaderboard, close to the top of the leaderboard, for just how much Eliezer Yudkowsky I’ve read.
Liron 00:32:34
And for those of you who don’t know, Eliezer Yudkowsky really is the pioneer in the last twenty years of this whole idea of why we’re so doomed, what the problem is with superintelligent AI. He pointed it out before the AI was even here. He launched the research field before we even had the AI, and now we have the AI, and it’s almost the deadline, and we still don’t have the — the research hasn’t come far along, unfortunately.
Liron 00:32:56
But yeah, so I’ve studied thousands of hours of Eliezer Yudkowsky, and I also have a passion for the art of rationality, including rational debate. So I think I have something to contribute in demonstrating to people what a good debate looks like. A good debate can look intense and high stakes, but still not be a dick. That’s really the art form — you’re passionate and you’re not being a dick. I don’t think that’s that hard to do. It comes naturally to me.
Liron 00:33:17
And then I also have a mainstream, accessible talking style, so most of the people who have read this much Eliezer Yudkowsky, you try to talk to them, you’re gonna get a monotone, big words, stuttering — not to throw shade, but there is a correlation between being well-versed in these concepts and being a painfully bad communicator.
Liron 00:33:47
And also, I have a direct, confrontational, but friendly debate style. I mentioned high-functioning Asperger’s, so I don’t mind confronting people in a way that might make them feel uncomfortable, but the good kind of uncomfortable — you’re grappling with this concept, even though I’m being friendly, I’m still making you grapple.
Liron 00:34:02
Okay, and then I bring substance density to the table. That’s just my style. I think that makes it unique, ‘cause some shows have a lot of boring parts. And then I’m prioritizing entertainment and showmanship. I’ve got a freaking fire behind me here. So I’m pulling out all the stops with the studio. I’ve got this Doom Dog figurine on the side here. He’s pulling the lever, and the show’s theme is this dog who represents me.
Liron 00:34:32
It’s kind of a play on the dog that says, “This is fine,” but it’s a different dog, no IP infringement. So yeah, I’m bringing showmanship to the table here, and sometimes I’ve had segments like Logic Cop, and I’ll dress up as a cop when I’m finding a flaw in somebody’s logic.
Liron 00:34:47
I think the entertainment dimension often gets lost. And when I ask people, “What’s your P(Doom)?” on the show, there’s a whole P(Doom) theme song where robots come in and sing, “What’s your P(Doom)?” ‘Cause it’s a recurring segment. And also, “What’s your P(Doom)?” — there’s been a lot of controversy over that phrase because some people think that it cheapens the idea of doom to ask “What’s your P(Doom)?” and some people use it to make fun of doomers. But in my case, I’m like, “Let’s make that the slogan of the show.” What’s your P(Doom)? Doom Debates.
Liron 00:35:15
So this is what I mean by high production and entertainment showmanship. And yeah, we’re bringing high production values to the table, most notably the studio. We edit about twelve hours in a typical episode, how much we spend editing.
Liron 00:35:29
And then finally, the other thing that we’re bringing to the table as a unique show is it’s really the only AI analysis show that explicitly communicates that P(Doom) is high. Talk about an oversight. We’re living in this world where we might literally all get killed in a few years. Our children might not even grow up. The stakes are ridiculously high, and you’re not going to find another show that’s mainstream accessible that’s like, “Oh, yeah, by the way, we’re almost all doomed. Don’t forget.” It’s crazy.
Liron 00:35:57
And there’s really just one other show that does it, which is John Sherman’s channel, the For Humanity podcast and Warning Shots. He’s got a few shows that he runs, and he’s very clear that P(Doom) is high, but nobody else will say it. It’s crazy.
What Makes the Show Must-Watch
Liron 00:36:08
All right, so that is my talent stack. And I do think if you’re bringing something to the table, if you’re bringing a new show, you probably wanna think about what combination of factors are going to give you an edge.
Liron 00:36:21
Okay, the next part is, what makes this show must-watch? I think this is an important concept. If you’re wondering why people aren’t watching your show, you can ask yourself the question of: Well, why do they have to watch it? What need do they have that they really need to watch the show, otherwise they’re not gonna get their need met?
Liron 00:36:37
So one thing I can say about Doom Debates is it’s the best way to understand what a guest’s position is. So I recently had Noah Smith on the show, or Dean Ball — he was the guy who wrote the White House’s AI policy. He was on the show. If you really wanna understand Dean Ball’s position, you’re just not going to get a better place to understand it than Doom Debates.
Liron 00:36:55
So in terms of pure information about what the guest really thinks, nowhere else — sure, they can write their own position, they can go on a friendly podcast, but unless their argument is actually getting challenged and probed, you’re just not going to understand their position that well. So I can offer that service of probing people’s positions.
Liron 00:37:11
Another reason it’s must-watch is just because it’s arguably the most insightful and entertaining analysis of the AI doom situation. So if you wanna know the big picture and the latest developments of why you’re doomed by somebody who actually has the perspective that we’re doomed, this is definitely one of your best options.
Liron 00:37:27
Another reason it’s must-watch is if you come to like my style as the host and you enjoy slipping back into my perspective. Another reason is you believe in the mission, and you like watching and supporting somebody who’s pursuing this mission on your behalf — the mission to raise awareness of existential risk and hold people accountable to defend their views and raise the quality of debate.
Liron 00:37:47
There’s a lot of viewers who are like, “Yeah, this is a great mission, Liron, I wanna support the mission.” And by watching the show, you’re part of this mission. And without the show, it’s like nobody’s doing your mission, so it makes it a must-watch.
Liron 00:37:57
And related to this concept of why the show is a must-watch for the viewer, there’s this concept of the world needs this. The world would be unsatisfactory if your show didn’t exist.
Liron 00:38:07
I actually have had experience starting something where I didn’t check that box. So I run this startup called Relationship Hero, where we do relationship coaching. It’s kinda like couples therapy, but we do it individually, and we can give you dating advice, relationship advice. So I run this other company, and I think we create value. Don’t get me wrong. People come and use the service, and we have great reviews, and they like it.
Liron 00:38:29
But if my company shut down, would the world be unsatisfactory? Barely. Because people would just be like, “Okay, well, I guess I’ll just get a therapist. That’ll be kind of a similar experience to a coach. Whatever, close enough.” So I wouldn’t say that the world without my company is unsatisfactory.
Liron 00:38:43
But in the case of Doom Debates, I’ve already described how if Doom Debates disappeared, what do you have to do? Go back to hate-listening to these podcasts. These people can’t keep getting away with this, these terrible opinions on these other shows. There has to be a good show. So that’s what I mean by the world needs this.
Liron 00:38:57
I don’t think anybody right now would be filling the Doom Debates gap. John Sherman would a little bit, but he’s not really the debater type, so he’s still letting people get away with these bad opinions. There’s zero accountability without Doom Debates. And there’s this important doomer perspective — in mainstream discourse, who’s gonna represent the doomer perspective?
Liron 00:39:17
I mean, you’ve got the Machine Intelligence Research Institute, Eliezer Yudkowsky’s organization. Yeah, they’re representing it, but they’re not doing it on a weekly basis with that showmanship and that communication style. And so the doomer perspective is getting marginalized. So I feel strongly right now that this show needs to exist, and I think a lot of the viewers agree.
Finding Your Unique Niche
Liron 00:39:34
Yeah, I’ll take a question.
Vince 00:39:37
I feel like a lot of us are kind of making similar content to an extent, and I get that. I feel like the point you’re making is obviously, how is your content unique compared to other people’s content? But most of the time, I do feel like a lot of it comes down to the same thing. So are you suggesting you really get a clear niche where there’s nobody else who they could potentially watch instead? I don’t really fully understand if that’s how you—
Liron 00:40:13
Yeah, it’s a great question because I do agree. I mean, I’ve looked at your guys’ channels, and I do agree there is a risk that your channels are kind of similar to other channels to some degree. So I think this will be a very deep question for many of you of: why is this show a must-watch? Why is the world unsatisfactory without your show?
Liron 00:40:34
Maybe the way to get to the answer is to narrow down your niche and be like, “Hey, out of all the things that I’ve done, what are the things that I’ve done that are quite irreplaceable?” Or, even if they’re not irreplaceable, it’s the things where, if I were to delete these particular episodes that I’ve done or these particular posts, then there really would be a gap. Somebody else needs to come and make those particular posts.
Liron 00:40:52
And then be like, is there a line connecting those types of posts? And maybe I should focus on that.
Vince 00:40:56
Okay, cool.
Transitioning to Q&A
Akshay 00:40:58
By the way, I think now would be a good time to do some Q&As, Liron, if that’s okay?
Liron 00:41:04
Yeah, yeah, totally. That’s what I — yeah. I prepared a lot more than fifteen minutes worth of content, and so, yeah, maybe I’ll give you some headlines of stuff I could unpack if you guys want, and then I’ll just toss it to you guys, and you can ask me about anything.
Veronica 00:41:18
That’d be great.
Liron 00:41:20
Nice. All right, so we talked about what makes the show must-watch. Let’s see. The kind of host I am — I see myself as similar to Bill Maher. Another quick bullet point here is how the show increased its production values and its investment over time. Another topic is how we fund the show’s production. Another topic is our social media strategy.
Liron 00:41:40
Another topic is my view on AI existential risk, which we’ve already talked about a little bit. And the conclusion is just I would like to see other people in AI media be explicit that P(Doom) is insanely high and fill in what I call the missing mood of basically yelling and screaming that we are pretty doomed. So yeah, those are basically the topics we haven’t touched on yet.
Akshay 00:42:02
Any questions?
Funding the Show
Vince 00:42:04
I would be interested in learning more about how you are funded or how you managed to pay for all of it.
Liron 00:42:11
Yep. Yeah, I figured you guys might be interested in that. All right, so for the first year, it was just me working for free. Very simple, and it’s obviously a fortunate position where I have my startup that’s paying the bills. I have some savings. So yeah, it was just a passion project, and I had enough time. Maybe some of you don’t have that luxury, and then I don’t have any advice for you. You’ll have to figure that out.
Liron 00:42:33
After the first year, a really generous viewer did something that I’m guessing is super rare. I don’t expect you guys to have this advantage, but somebody threw in one Bitcoin. And when I got that, I was like, okay, let’s get serious about production. If we have a viewer who’s gonna invest this much in the show, I got to invest in myself.
Liron 00:42:51
So I immediately hired a full-time producer, Ori, and I contracted a bunch of professionals to build me a studio, which you see come to fruition. And we’ve even hired an intern or two, and we’re hiring more, actually.
Liron 00:43:03
We turned on YouTube ads, meaning we show ads, and that brings in four hundred dollars a month. So four hundred dollars a month is not really enough to fund production. We really need twenty-five times more views to fund production based on the YouTube ads — a few doublings. So right now, the only way that I think is sustainable to fund the show is a donation model.
Liron 00:43:23
But we’re just fortunate enough that we have a handful of super passionate viewers — viewers who are at the intersection of caring a lot about the cause and high disposable income or high assets. Like crypto millionaires, in the case of this person who donated one Bitcoin, an early-to-crypto person. A couple of those people have donated.
Liron 00:43:45
And so I’m not sure if that’s true for all of your shows. I’m not sure if you happen to have people at the intersection of feeling really passionate about your mission and being really high disposable income. A lot of you might think, “Oh, I should start a Substack or a Patreon. I should get memberships.” But me and Ori realized that a ten dollar a month Substack subscription just isn’t going to move the needle. Our budget — we need thousands of dollars per month. This studio costs many thousands of dollars.
Liron 00:44:05
We’re not gonna get there with a few thousand people where one percent of them pay ten dollars a month. We’re just not going to get there. In our case, we just had to ask for big donations, and it happened to be sufficiently feasible that we think that’s kind of the winning model for us.
Liron 00:44:24
We have this concept of becoming a mission partner. If you go to doomdebates.com/donate, and we’re very explicit. We’re like, “Listen, if you want to be a mission partner, if you want me to consider you a partner on this mission, you can’t just kick in ten dollars a month. That’s nice, that helps our Substack ranking when you become a paid subscriber, and I’ll even send you a free T-shirt. But if you actually want to meaningfully help the show, the minimum is a thousand bucks.”
Liron 00:44:46
Yeah, you got to step it up with a thousand bucks, just because even then, we have multiple thousand just to fund the next month of production. And so we’ve been very explicit about that, and I do think that’s kind of rare. It’s rare to see a show where they’re treating their viewers like they’re throwing around thousands of dollars.
Liron 00:45:04
But I just thought I was specifically speaking to the people where they can spend at that level, and I believe that a few of those people existed in my audience, and I think that’s turned out to be correct.
Janet 00:45:15
I have a follow-up question on the funding point — ‘cause I had a question a little bit in terms of your relationship with the people who are donating to you. But I’m curious about looking forward. Aside from donations, are you thinking about other potential funding models that you would be comfortable sharing?
Liron 00:45:33
Yeah, good question. So for one thing, when the audience grows, that naturally presents a funding model. So everybody who has an audience that’s twenty-five or fifty times bigger than my audience should be good to go on funding just in terms of ad revenue, ads in the show. Those are kinda standard funding options.
Liron 00:45:49
So maybe the real question you guys are asking is: Hey, if I only have less than ten thousand viewer hours per month, or maybe I have fifty thousand, but they’re low-quality viewer hours or whatever — I don’t have an audience that I can just go slap on some ads and make money. How do I bridge the gap? ‘Cause we’re all just talking about bridging the gap. We all agree that once we have big audiences, then there’s plenty of funding sources, correct?
Lauren 00:46:12
Depends what platform you’re on.
Liron 00:46:15
Hmm. Okay, fair enough. Well, okay, so let’s put it like this. Everybody has some kind of gap. Maybe if you’re on Instagram, maybe you need many millions, and if you’re on YouTube, you’re good with five hundred thousand. It depends. But we’re all bridging some kind of gap.
Liron 00:46:30
And like I said, my answer was large viewer donations. Another answer of how to bridge the gap is because you guys are all working for a good cause — raising AI safety awareness — there actually are funds. There’s donation funds. I’ve actually applied to a couple myself. None of them have actually hit... or sorry, one of them has hit, which is Manifund.
Liron 00:46:53
You guys should all check out Manifund. If you haven’t tried Manifund already, you should list your project there. That’s a donation page that you can link to. And they have various fundraising drives, so I’ve raised a few thousand that way. But most of my donations have come in when I just promote to viewers on my own channel, and then I link to doomdebates.com/donate, which links to Manifund. So most of the money on Manifund, I just raised through my own donation drive.
Liron 00:47:15
But if you list yourself on Manifund, that’s one way to get donations, which has worked for me from the users of Manifund. And then there’s other initiatives — there’s the Survival and Flourishing Fund. If you go to aisafety.com, there’s a list of funders, and you should just go down the list and try all of them. I’m about to try Survival and Flourishing Fund.
Liron 00:47:30
I’ve tried the media program at Future of Life Institute. Shout-out to Max Tegmark and Jan Tallinn, two founders who I actually know personally. So I felt like my odds would be pretty good because I’m pretty connected to those guys, and they expressed some interest, but so far they haven’t come through with a donation.
Liron 00:47:49
But it’s definitely — if you guys need to bridge the gap and you’re not getting viewer donations, you can definitely just keep trying to raise your viewer count and keep applying to all these different grants, and if they don’t grant you, you just hit them up and you’re like, “Okay, what do I need to do? Just tell me what’s worth a grant to you.” I would definitely pursue that angle.
Audience Demographics & Gender Strategy
Janet 00:48:05
I had another question, if no one else has a question, and it’s about your strategy around audiences. So you mentioned at the start that the split in your audience was ninety-six percent male, probably four percent female. So I’m curious to hear if that’s something that you’re seeking to shift. What’s your strategy around that? How are you thinking about that specific issue?
Liron 00:48:26
Good question. So I don’t even know if anything I’m doing is specifically designed to shift that. I do think that it’ll naturally shift as we get more prominent, because let’s say Yann LeCun was coming on the program — I think a few more female viewers would get roped in, because as you converge toward just having a popular show, you just get popular people, and being popular implies dragging along some women, too. So I guess that would be my number one strategy.
Liron 00:48:53
But some people have asked me, “Hey, why don’t you bring on more female guests?” And that is pretty valid, but it’s also, you get this problem where the few female guests that there are, they also get more requests. So it’s tough. I mean, I’m open to ideas.
Lauren 00:49:07
I do think female guests would help. When you were running through the list before, I was trying to think if any of them were women, and if there were, I didn’t recognize them. No, I’m just kidding.
Lauren 00:49:18
No, but actually, you should interview them.
Liron 00:49:24
Yeah, yeah. Okay, for sure. So definitely, if anybody has connections to women who’d be interested in doing debates, let me know. Yeah, we have had a couple with the show, so it’s non-zero, but yeah, it could use more for sure.
What Should the Audience Do?
Gaetan 00:49:40
A question — basically, what do you expect the audience to do in terms of having an impact with lowering the P(Doom), basically? And do you sometimes try to ask them to write to their elected officials, or do you send them to a specific page where they can learn more?
Liron 00:50:04
Also a great question. Yeah, I see my show as the top of the funnel, where people discover the show, and they stay engaged, and then they look to go to the next step. I could probably do a better job of giving specific steps, but it’s honestly hard just because it’s like, what are the actual steps?
Liron 00:50:18
Yeah, you can write to your congressperson here and there. You can join a protest. Occasionally, they happen. I think raising awareness itself — it’s one of those few times where just raising awareness actually is surprisingly helpful. Talk to people in your life about it, and then make it a voting issue.
Liron 00:50:38
But yeah, I mean, it’s a good question. The reason why I don’t spend that much time on the show being like, “Oh, you have to do this specific action,” is because I don’t know which one specific action is disproportionately impactful. I actually think if people would just share the show, that’s one of the most efficient ways to spread the word — to be like, “Oh, yeah, this is a popular show in the space.”
Liron 00:50:53
Once this is a popular show in the space, then it’s easy for people with power to make laws and to decide policy. They’re like, “Oh, okay, this show represents the people. Okay, great. Let’s make policy accordingly.” ‘Cause ultimately, policy is the biggest lever we have. The Pause AI movement is saying, “Hey, we need a treaty to pause AI.” Nothing else can be done at this point. We’re actually in a very late stage in the game to try to survive AI.
Communicating AI Risk Effectively
Inken 00:51:17
I was wondering how we can communicate that people are listening, right? Because if you’re confronting a person with, “You are going to die within two years, and everybody’s going to die,” they truly just scroll. Or they’re really mad and say, “You’re wrong.”
Inken 00:51:41
So how are you shaping — or maybe I know how you do it because you’re doing it with debate, but not all of us are debating. Is there a particular thing we can do to craft this message in a way that people are listening?
Liron 00:51:57
Yeah, I mean, I think that’s the question we’re all grappling with. I mean, yeah, I think you will be able to answer it just as well as I could. I’m taking a particular angle where I think my own comparative advantage is to interact with people who are spouting their own points — to engage and confront people who are shaping the discourse, try to bring them on the show.
Liron 00:52:19
I’ve seen you guys have very different approaches, which I think are valid and useful. A lot of you are covering — you’re giving people analysis, you’re making it understandable to them. I think you got to know your audience, and your audience is going to be very different.
Liron 00:52:30
In some cases, your audience is inherently more broad than my audience because there is something about my show which makes it a little bit niche and not super mainstream, which is I do like to get into some technical details. So yeah, it’s different.
Episode Preparation & Production Tips
Janet 00:52:48
I’m curious to hear about how you approach an episode where you broadly speaking agree with your guest, even though you’re in a debate context. So, for example, your conversation with Eliezer Yudkowsky. I watched a little bit of it, but like you said, you’re like top five percent of people who’ve really studied his work and agree with him. And so how do you approach a conversation like that within the debate framework?
Liron 00:53:13
Yeah, so a lot of the people I bring on the show, it’s pretty explicit — yeah, this conversation is gonna be eighty percent interview, twenty percent debate. Maybe a few disagreements will come up here and there, but we’re pretty much just walking through points that I think are interesting, and I’m just helping them make their point.
Liron 00:53:28
But that’s a larger thing about this show, which is, every single guest is actually landing somewhere on this continuous spectrum, a continuum between debate and interview. And sometimes you can’t really tell, because part of a good debate is you really have to probe somebody’s position, and probing somebody’s position is a good interview.
Liron 00:53:45
So people would sometimes give me compliments in the comments and be like, “Wow, you’re such a good interviewer.” And I’m like, “Really, I am? I’m not a good interviewer,” but I guess just setting them up in order to be able to properly debate them turns out to be a good interview.
Gaetan 00:53:58
Great question. How much time do you spend preparing for the interview? Do you study the guest? Do you look at their past writing or podcast interviews? And there’s a question that’s linked to the production process — how much do you think having a teaser at the beginning of each podcast works for you? Because I don’t do it on my podcast, maybe I should, so I’m just wondering.
Liron 00:54:28
Yeah, sure. Okay, on the preparation side, it definitely varies by episode. So for example, I recently recorded an episode with Destiny, and that was pretty low preparation for me, just because the premise of the episode was I’m going to walk Destiny through the AI doom argument. And so the biggest piece of my preparation doc was, okay, here’s the AI doom argument that I walk everybody through, and then we’ll see where he disagrees, where he gets off the doom train, is what I call it.
Liron 00:54:52
But then an example of an episode where I did high preparation was a recent episode with Noah Smith. Noah Smith has written a bunch of blog posts about his different views about AI, and it’s actually thousands of words. So I had to load those all in my podcast player. That’s how I catch up on things. I listen to hours of podcasts. A lot of times I’ll listen to eight hours of podcasts, even more, to prepare for an episode in the week before the episode.
Liron 00:55:14
And then I’ll prepare a big outline doc, so I’m constantly taking notes, making this outline. And then when the show happens, when I’ve done the best job of preparation, I’m constantly bringing up quotes like, “You’ve said this, what were you thinking here?” And then I can ask these nuanced questions about exact stuff that they’ve done. And yeah, half my episodes are like that, and then half of them are lower prep.
Liron 00:55:33
And then to your question of — I think you’re saying the highlights reel right at the beginning of the episode, should you do that? I think the obvious answer that’s considered a best practice is yeah, you really do want to invest. Let’s say you spend five hours making an episode, you really want to spend one whole hour, ten or twenty percent of the effort at least, just making the first thirty seconds or the first minute.
Liron 00:55:53
Because if you look at the graph of when people drop off, it’s crazy. It’s like half the people drop off in the first minute, even more than half. So if you don’t invest in the first minute, you haven’t invested in half your audience. So I recommend — me and Ori will spend a couple of hours on the first minute.
Social Media Strategy
Akshay 00:56:11
Yeah. We’ll take about one or two more questions.
Liron 00:56:16
Yeah, sure. Yeah, I’ll take a couple more and then we’ll wrap it up.
Vince 00:56:22
You had more to talk about, about your social media strategy, so I’m curious to hear what things you were—
Liron 00:56:34
All right, social media strategy. So it definitely varies by show. In our case, for now, we haven’t unlocked all the — I don’t even know what the best social media strategy is, but one thing we’ve noticed is that Twitter is really where it’s at for us — X — because it’s where our guests are. All the prominent intellectuals that we’re trying to spar with, they’re all on Twitter, and our most influential audience members are on Twitter.
Liron 00:56:56
So once in a while, I’ll get a like from a Vitalik Buterin or a CEO of a major tech company. I’ll be like, “Oh, wow, okay, that’s good, that’s cool that Doom Debates is part of the discourse in that sense.”
Liron 00:57:07
And also, the thing about Twitter is we often see these organic proto-debates that are happening among other people, and that’s what makes me feel like, okay, well, I’m glad I’m doing a show in this niche. This is clearly a vibrant niche because I’m seeing two people spar informally on Twitter.
Liron 00:57:21
So I’ll just bust into their thread. I’ll be a reply guy. Actually, I’ll quote tweet them, and I’ll be like, “Okay, you guys are invited onto Doom Debates. You’re clearly sparring right here, so if you’re gonna spar, why don’t you do this right? Don’t tweet back one hundred and forty characters back and forth.”
Liron 00:57:34
And we’ve taken it up a notch. Ori and I, we’re making these cool cards that look like trading cards, and they have a person’s photo on it and the giant word “Invited,” and then it encourages a hundred people to like the tweet, and then it’s like, “Hey, you’ve been invited.”
Liron 00:57:47
And again, the ultimate strategy of all this is that you can’t just ignore this. You can’t just conveniently ignore us asking people to participate in this forum. That’s where we’re going with this. Now, the success rate is only about ten percent, but it’s climbing. We’re getting more and more people being like, “All right, fine, I’ll go on Doom Debates.”
Liron 00:58:04
But it just goes to show why the world without Doom Debates is unsatisfactory.
Closing Remarks
Akshay 00:58:10
All right, I think we’re gonna end it right here. Thank you so much, Liron. It was so helpful and such an insightful session. I’d love to know if you have any closing remarks for us. And yeah, hopefully, we can see you at — we’re having the end of program showcase in March, so we’ll talk more about that over email. But yeah, any closing remarks?
Liron 00:58:31
Yeah, yeah. All right, thanks so much, Akshay. Glad you’re running this program because I definitely would love to see more people in AI media. So my closing remarks are: be explicit that P(Doom) is insanely high. We really, society really depends on people like you to lead the way, move the Overton window, make it okay to acknowledge the truth that we’re doomed.
Liron 00:58:53
It’s literally like you can’t even say it. It’s like a thing you’re not allowed to say. So I encourage you to say it, have a sense of perspective about it, lead other people to it.
Liron 00:59:03
And then I also wanted to say, Ori and I are actually hiring for interns. So we have this position — it’s twenty-five dollars an hour. You can work remotely a few hours a week. You can help us edit the show. We have a bunch of other tasks, and if you’re interested in that, you can talk to Ori or just go to doomdebates.com/internship. We’ve got a bunch of different details on how you can apply.
Liron 00:59:24
And yeah, my last takeaway is when you’re working on your own shows, think about it in that perspective of why does this need to exist, and why is it a must-watch for your audience? And make it happen. Thanks, everybody.
Akshay 00:59:37
Thank you. Thanks a lot, Liron. Great session, and yeah, we’ll talk soon. Thank you so much.
Liron 00:59:42
All right. See you, Akshay.
Doom Debates’ Mission is to raise mainstream awareness of imminent extinction from AGI and build the social infrastructure for high-quality debate.
Support the mission by subscribing to my Substack at DoomDebates.com and to youtube.com/@DoomDebates, or to really take things to the next level: Donate 🙏
