Based Camp - Mystery! Blacks & Progressives Keep Getting Flagged As Using AI to Write？ (Racism Only Explanation)

00:00:00.000 What he's basically saying is that African writers are going to be flagged by Pengram because their writing is too perfect.

00:00:08.380 Do you understand?

00:00:09.460 So here you can see the Pan-African Review, share of articles flagged by AI by publication year.

00:00:16.120 2019, 0%.

00:00:17.580 2020, wrote another test of Olangunju's thesis is to simply look at his writing.

00:00:22.660 He wouldn't lie about this stuff.

00:00:25.080 Of course not.

00:00:26.220 the graph for him is even worse a hundred percent a hundred percent it goes from zero percent

00:00:36.320 well his emailed statement responding to these accusations were okay also came out as 100

00:00:43.680 percent ai written by pan graham so i hope that the ai was like okay i regret nothing

00:00:50.460 i love an ai having to make an answer about how this isn't written with ai and i'm sorry

00:00:58.060 that things looked like they were written with ai before i wonder if like whatever he's using

00:01:03.680 uh asked him at first like are you sure it's ethical to have me write this

00:01:08.180 i know would you like to know more hello malcolm i'm very excited to be speaking with you today

00:01:14.600 because it turns out that there are apparently some groups that are being discriminated against

00:01:19.880 being falsely accused of using AI to write and they don't use AI to write they say one of these

00:01:27.820 groups is um African a or black Americans who are like no we these I gotta jump to the punchline of 0.56

00:01:36.840 what you told me on this because that's what we're going to be getting into okay but there have been

00:01:40.620 these black academics who have complaining that the AI searchers are all racist against them

00:01:45.880 and what people have done is run their work through ai detection and they've all passed

00:01:52.540 completely up until gpt gets popular yeah they're all getting hit by ai detections and the reality

00:02:00.320 is is what we are seeing and this is what we're going to be talking about here is communities

00:02:04.680 that became so used to cheating that the idea of not cheating is completely beyond them

00:02:12.060 And I guess because this is a new means of cheating, potentially, it's it's they're getting caught, whereas before maybe the method had been around long enough to not get caught.

00:02:22.860 In fact, there may be an interesting case in point there of someone who has has evaded this particular black spot on their record because they continue to use old cheating technology instead of new cheating technology.

00:02:35.620 yeah so anyway let's though start with the lead story which was broken by cremeu uh we love

00:02:41.520 following on x uh and we love his substack he he wrote if you read about pangram you'll quickly

00:02:47.640 see a bunch of people claiming that it doesn't work worse they'll sometimes claim that it's

00:02:51.820 biased pangram allegedly discriminates against autistic authors against anti-vaccine authors

00:02:57.920 against people i'd politely described as conspiracy theorists on both sides of the aisle

00:03:03.360 and most prominently against African writers. How will it supposedly flex their work as AI?

00:03:10.220 There are a lot of articles about Pangram and other AI detectors mistakenly flagging

00:03:15.780 Black and African authors as using AI to do or assist with their writing. The most prominent

00:03:21.420 piece that comes to mind for me was published in the Pan-African Review. The piece was an unusual

00:03:27.360 breakout for the review. The journal generally enjoys a pretty limited circulation. I suspect

00:03:32.780 this is because of the import of the topic area and the utility of providing an example to point

00:03:38.220 to when people want to explain why they were wrongfully flagged as using ai for writing so

00:03:43.880 i'll consider the perspective of this piece to be emblematic of a community's wider concern oh by

00:03:48.200 the way for people who want to get around this um rfab.ai's api which you can get by going to

00:03:54.300 rfab.ai backslash api you give it to your ai one of the things that actually gives your ai the

00:03:59.480 capability to do is feed anything it's writing into pengram yeah specifically pengram too

00:04:05.080 yeah specifically pengram too to look for what gets flagged and then send it to models with

00:04:12.080 framings that we have that are very good at human-like writing to specifically rewrite it 0.53

00:04:16.460 so pengram can't detect it so if they weren't using black people watched our show they wouldn't

00:04:22.440 have a problem with this i know poor things um so he writes the piece is authored by one timmy

00:04:28.300 Olanguju. Olanguju is a lawyer and public policy expert working at the intersection of tech

00:04:34.500 governance and development. His claim is that writing in formal English will be flagged as an

00:04:40.220 AI generated by an AI text detector. Okay, so Malcolm, what he's basically saying is that

00:04:47.140 African writers are going to be flagged by Pengram because their writing is too perfect.

00:04:53.380 Do you understand? 1.00

00:04:54.920 Oh, yes.

00:04:56.140 Right.

00:04:56.420 So in the piece, he writes, consider an international student in the United States.

00:05:00.700 She has spent years being told that her English must be grammatically correct.

00:05:04.840 She avoids slang, checks every verb and structures each paragraph carefully.

00:05:09.620 Then a detector labels her work as artificial because it is controlled and predictable.

00:05:15.520 First, the institution tells her to master standard English through IELTS and TOEFL.

00:05:22.420 I think that's another ESL thing. Then it treats mastery as suspicious.

00:05:29.280 Cremio continues, this allegedly explains why African or Black writers are liable to have their

00:05:34.040 work flagged by AI detectors. He again quotes from the article, for many African writers,

00:05:40.640 like Falad, English is both a language and an inheritance shaped by colonial history.

00:05:46.500 In Nigeria, Kenya, Ghana, and other former British colonies, generations of students

00:05:51.620 were taught forms of english influenced by british education which i mean weren't we all

00:05:57.220 um teachers emphasized correctness formal composition and structure creme breaks in

00:06:03.240 again is that the case we can test this right away by inputting on oh my gosh his own article

00:06:09.320 into pangram and it receives a 100 ai score with high hello hello yeah so the their cultures were 0.55

00:06:16.340 terrible before the british got there and introduced their education system they murdered 0.98

00:06:20.620 each other constantly they sold each other into slavery constantly they it's true even in the book 0.51

00:06:27.020 like it's supposed to be that i always mention here because it always got me as a kid where we

00:06:30.620 were reading about how horrible the british were and one of the things is that they weren't allowed

00:06:34.420 to expose babies anymore when they were twins and i was like what are you talking about this isn't 0.99

00:06:40.100 horrible like clearly you're you came from a garbage culture of garbage people right like 0.98

00:06:45.500 my ancestors were the same way before the romans civilized them right like most of the world was 1.00

00:06:51.700 godless savages who sacrificed people and kids and and then and then they were civilized the

00:06:59.140 the uh what you call the english education system is only a continuation of the education system 0.61

00:07:04.100 that colonized us uplifting us raising us to be something yeah that was my instinct was like well

00:07:11.020 isn't aren't non-black or african students also taught to adopt some version of british english i

00:07:18.780 mean it why would i guess that he's arguing though that um people in africa are being held to some

00:07:25.980 higher standard and given more of a side eye when not using correct english though i don't i don't

00:07:32.860 think that that's well substantiated but it's still the argument that he's making anyway the

00:07:39.820 The point is that Cremieux did put it into Pengram and it got a 100% AI score with high confidence.

00:07:45.620 He writes, surely he wouldn't use AI to write, right?

00:07:48.880 Let's consider his thesis in good faith and perish the thought for now.

00:07:53.900 Surely the other writers in the Pan-African Review wouldn't use AI to write either, right?

00:07:59.140 To test the extent of AI adoption in the Pan-African Review,

00:08:02.220 I looked at how much Pengram flagged articles there over time.

00:08:06.620 Unfortunately, the site is currently down,

00:08:08.360 so I had to extract whatever articles I could from the Internet Archive.

00:08:12.040 What we see right away is a total rejection of Olangunju's thesis.

00:08:17.400 Articles were never flagged as AI generated before the advent of LOMs,

00:08:22.220 and the share flagged as AI written has rapidly crept up afterwards.

00:08:26.840 This is not what we expect if African writing generally is to be considered liable to be flagged as AI.

00:08:33.600 I will show you the graph because it's quite remarkable.

00:08:37.140 So here you can see the Pan-African Review, share of articles flagged by AI by publication year.

00:08:43.300 2019, 0%, 2020, 0%, 2021, 0%, 2022, 0%, 2023, 0%, 2024, 3.3%, 2025, 50%.

00:08:54.640 Those are the pioneers right there.

00:08:56.420 2026, 89.7%.

00:08:59.740 Oh, good Lord.

00:09:01.580 Help us.

00:09:02.260 no um i love what i love is it's like completely ai generated as well yeah um yeah it's not like a

00:09:10.900 little bit well he breaks out in his graph quite helpfully um ai generated ai detected and ai

00:09:17.200 assisted where i think um ai generated is you know clear this was just an article written by ai

00:09:23.240 ai detected i believe falls into the category of parts of this appear to have been written by ai

00:09:29.080 but then some parts are not identified. So basically mixed some of its generative AI.

00:09:34.120 And then maybe AI assistant is more like this clearly was grammatically checked by AI,

00:09:39.280 but probably written by a human. That's my understanding. And he's looking at samples of

00:09:45.500 between 30 and 46 for these. So while he admitted that he was unable to get the full site archive

00:09:52.860 due to the site being down, he still got a decent amount of work, but it doesn't just stop there.

00:09:57.600 he really came for them he wrote another test of olangunju's thesis is to simply look at his

00:10:04.660 writing he wouldn't lie about this stuff of course not all of his past work before llms

00:10:11.540 should come out a return uh oh sorry all of his past work before llms came out should return a

00:10:18.820 100 ai score with at least moderate confidence right because then this is me talking his thesis

00:10:24.880 was we are our writing is detected as ai because it's too perfect and therefore yeah because of the

00:10:31.720 way we're talking about yeah even before llms came out we should be getting ai warnings

00:10:36.520 he continues alas quite literally none of that i could find does uh all of it shows up as human

00:10:44.340 written it's only after llms came out that his work started being flagged as ai generated and

00:10:51.500 i must admit the graph for him is even worse a hundred percent

00:11:00.800 it goes from zero percent well no hold on he rammed up a little in 2023 it was just 33.3 percent

00:11:10.000 um but yeah 2024 100 but 100 ai detected so 100 uh percent ai detected you know maybe he was just

00:11:20.040 really mixing things up or like really trying to cover it up uh like rewriting a lot of sentences

00:11:25.120 maybe um no no what's clear is in the first year either he didn't think the ai was good enough yet

00:11:31.460 or he was in a position where he was afraid of getting caught yeah because i can tell exactly

00:11:36.200 what this is this is somebody who goes from being afraid of getting caught to just not caring anymore

00:11:39.620 well yeah but the second year i think he was still afraid of being caught because all of the

00:11:44.100 all of the 100 is made up of quote-unquote ai detected per pen yeah that's what i mean yeah

00:11:49.000 You can see in the second year, he's doing it on everything, but he's afraid of being

00:11:52.940 caught still.

00:11:53.500 Yeah.

00:11:53.700 And so then 2025 and 2026, about 40 to 50% of it, almost 50%, a little under 50%, we'll

00:12:01.840 say, was just AI generated.

00:12:03.560 He didn't even try to cover it up, whereas the rest, he probably smoothed over a little

00:12:09.780 with some word changes here or there, or maybe some sentence restructuring.

00:12:12.640 I'm not exactly sure how one attempts to hide one's use of AI because we just were super cool

00:12:20.160 and open about using AI. We like it. But it's pretty damning. So finally, Cremieux writes,

00:12:30.500 but maybe we're at fault here. Maybe Pangram trained on the Pan-African Review and Olangunju's

00:12:37.080 work, and they flag the pre-LLM work as AI generated because it's in the training set.

00:12:43.360 To get this question, I asked Pengram CEO Max Spiro if that was the case. Was Pengram trained

00:12:51.020 on the articles in the Pan-African Review? Answer, it was not. So yeah, he has no out here. It just

00:13:00.420 looks really bad. Yeah, he's just lying. And what it shows to me is that you would have the balls

00:13:05.520 to do this when you know you're cheating on everything that that just shows a level of 0.95

00:13:12.140 you have gotten and that it's happening across the field with the entire review that they're

00:13:17.020 not even checking this and going back to these people and being like this is a well they'll be

00:13:21.780 checking it now obviously and and hopefully using the rfab api to not get caught because the oh no

00:13:29.960 oh no the first rule of committing a crime is to not commit a crime while you're committing a crime

00:13:35.500 And the second rule, or maybe the double first rule is don't get caught, I guess.

00:13:40.040 Like, see, the thing is, you're way more likely to get caught.

00:13:42.780 So I guess, yeah, the first rule of committing a crime is don't get caught.

00:13:45.480 And then the second rule should be don't commit a crime while committing a crime.

00:13:48.240 But that's only because you're more likely to break the first rule.

00:13:51.100 That is to say, get caught if you're committing a crime while committing a crime, right?

00:13:54.020 Like if you're speeding while you have a kidnapped person in the trunk of your car.

00:13:58.600 That's how it, you know, I mean, just don't commit a crime while you're committing a crime.

00:14:02.140 And actually, he did that here, right?

00:14:03.720 Because he lied about using AI while using AI.

00:14:09.580 Don't commit a crime while committing a crime.

00:14:12.240 This is what we have to teach our children to raise them right.

00:14:16.080 Very important.

00:14:17.500 You don't want to get caught.

00:14:18.660 And that's the key to crime.

00:14:21.920 So what I thought about first when I was reading about this was, of course, the most recent academic kerfuffle involving someone who was, I guess you could say, 0.91

00:14:32.740 conspicuously black which is to say someone who really made their ethnicity and racism 0.83

00:14:38.340 a core part of their academic identity and of course i'm speaking of jason our day 0.77

00:14:43.880 um the the late academic uh of cambridge who sadly committed suicide following allegations

00:14:52.400 regarding his academic integrity allegation the proof well no so not everything was fully proven

00:15:01.080 now the most the most serious thing of everything that was accused against him although i'm sure

00:15:05.980 more will will take place when people get deeper into the book that he authored because i'm sure

00:15:10.440 there's a lot of stuff in there that is very easy to disprove because a lot of it's just not

00:15:14.740 physically possible what he described um like just i don't know we might get into it um the

00:15:22.940 whole marathon but like with tree trunk legs a lot of this doesn't make sense but um so the

00:15:28.080 academics dave harris and martin hammersley raised concerns about him actually as early as 2023 which

00:15:33.980 i didn't know about they said that there were similarities between one of his papers on

00:15:38.120 minority students mental health again to my point about him making race kind of his whole spiel

00:15:43.360 an earlier study by arjun menon mem memon who um had written something similar and there was just

00:15:53.880 there were way, way too many similarities. So then two journals subsequently published

00:15:58.720 corrections addressing attribution and textual overlap. So it was pretty clear that he was

00:16:05.080 guilty of kind of like at least semi-plagiarizing. But then more importantly, some quotations

00:16:11.340 presented as coming from Arday in his research really, really, really closely resembled

00:16:16.820 quotations from participants in Memon's earlier study. Like it looked like he had just taken

00:16:22.000 quotes from this other guy's earlier research, and then just use that as his research because

00:16:27.920 someone had conveniently already done the interviews for him and he could just use those.

00:16:33.240 So Retraction Watch, and I've linked to all this in the show notes, they documented examples

00:16:36.880 and reported that Arday later removed or altered numerous participant quotations because he

00:16:43.560 realized he was going to get caught. And if this ultimately substantiated, and Cambridge is actively

00:16:49.200 investigating this after first defending them sorry jason our day after first defending him

00:16:55.300 they're now investigating him um well we may find out whether or not he was actually super super

00:17:02.080 guilty um but this is going to be not only a plagiarism problem probably but also just a

00:17:08.720 a research integrity problem and i think that the core thing with jason our day i think what we saw

00:17:15.940 was this guy's career when you look at how far he got in his career um it shows that if if somebody

00:17:22.820 claims to be like black and disabled or apparently just black they do not double check your work

00:17:27.200 right yes this happened with an entire journal yeah well and that was that was one example of

00:17:34.020 his work and i think what you can maybe look at when it comes to your your assertions about

00:17:38.420 cheating which i think to be some they're they're interesting um in that the people who are using

00:17:45.220 AI are getting caught now. And they're getting caught because they're using new cheating

00:17:48.720 technology without really understanding the countermeasures that may be at play.

00:17:53.060 Jason Arday, I think, may have been a pretty sophisticated cheater. And of course,

00:17:57.920 all will come out later when people do investigations. So all of this is very

00:18:01.400 alleged on my part. But I actually was like, oh, well, why don't I try the Pengram thing with his

00:18:05.860 work just to see? Because maybe he was trying this out or maybe he wasn't. And from what I can

00:18:11.380 tell he wasn't, but the most interesting thing is it's almost impossible for me to determine

00:18:16.260 what he actually wrote or if he wrote anything. What do you mean by that? Yeah. So I, to do my

00:18:24.360 whole pengram thing, I found some of his work in this case. Uh, like I went to Google scholar and

00:18:31.900 I searched his name and the only work that he solely offered authored I could find was from

00:18:37.060 2018, 2019, maybe like one or two articles from 2020, both academic or just editorial random

00:18:43.820 like articles, like blog posts, basically. After that point, though, everything he wrote and

00:18:50.360 published, unless it's somehow been taken down, was co-authored. And I ran all of it through

00:18:57.020 Pangram. Nothing was AI generated. It all came back as 100% human. But also, I was shocked by

00:19:05.860 the fact that i couldn't find anything that he had authored on his own and i wondered if it turns

00:19:10.540 out that maybe part of his gambit and what actually enabled him to get by a lot of the time

00:19:17.100 was by proposing to co-author things with people and then just having them do all the work um and

00:19:25.660 it's very hard to get caught cheating when you do a group project like look you need a black author

00:19:30.400 on this for it to get put through right? I mean in many of the instances where he did co-author 0.88

00:19:35.660 articles which were quite often almost exclusively about things like racism and disability etc

00:19:42.040 um he did often partner with people who are not black at least um or of of like more obvious

00:19:52.300 African descent so far as I could tell though not all of them were white it could be that he used

00:19:58.120 his prestige as a character because he was definitely rising as this like amazing, famous

00:20:05.480 person with this incredible background where he would collaborate with people on projects and

00:20:11.280 then they would just do all the work. And the reason why he doesn't have any AI detected work

00:20:15.900 is because he didn't actually do any of the work, which kind of, to me, seems much more smart and

00:20:22.040 sophisticated because we all know about that group project dynamic in academics environments,

00:20:27.540 yeah we're like you're stuck in a group project with people you need to get the thing done you

00:20:32.520 can't handle the fact that these people are doing absolutely nothing so you just take all the work

00:20:36.500 home and you do all of it the greatest thing about being married to simone and working with

00:20:43.500 simone is i don't have that dynamic at all i always know that she's working her hardest yeah

00:20:48.640 but the worst part is that malcolm still does all the work and then just gives me credit along with

00:20:52.420 him which is extremely frustrating because it doesn't things and it's also one of the reasons

00:20:57.560 i work working with bruno like he doesn't he doesn't do work as fast as you or i do but i

00:21:02.780 know he's always working on something and he doesn't need to bug me with what he's done

00:21:06.060 yeah no everyone who works on rfab is a major contributor and not a dead weight and yeah there's

00:21:12.800 only three people who work on rfab it's me there could only be free look we have very high standards

00:21:17.740 um and i and bruno we love you very unrealistic standards for people yeah we're just like

00:21:23.280 yeah we very unrealistic yeah the the fact that yeah any we would work with anyone is

00:21:28.400 is a testament to them and i don't know how they can tolerate working with us bruno i don't know

00:21:31.620 i'm not i'm not a very nice he's not very nice but it's not that i'm not you guys should hear

00:21:36.720 how he talks to ai it's just like why would you call me about this never ask me this again you

00:21:43.060 should be handling this by yourself he's like saying all this to the ai i've said this before

00:21:47.420 The only reason you don't get that from me, Simone, is because you're my wife.

00:21:51.620 Yeah, I really am.

00:21:53.940 Yeah, you really lucked out.

00:21:56.520 You're being extremely, extremely accommodating.

00:22:00.160 And to outsiders, it seems so direct, but they don't realize what real direct looks like from you.

00:22:08.320 Anyway, I do think that that's an interesting method on his part.

00:22:12.440 But the larger story here that I find to be really interesting is that the tension that we have between AI use and people who lean toward progressive elite circles, like journalists, like academics, like authors, because they're all claiming to not to hate AI and they're all.

00:22:33.740 That's the problem.

00:22:35.480 Like that. So, yes, that is exactly the issue.

00:22:38.520 The next culture wars, as we pointed out, are at least going to partially be about AI.

00:22:42.920 But the problem is that it's going to be really hard to be elite in any sense of the word, like a top performer, and to not use AI.

00:22:52.980 It's the new version of doping.

00:22:55.660 You're really screwed if you don't use it.

00:22:57.100 So the people who are on the progressive side of this war, who have to at least performatively say that they don't like AI and that it's wrong and that they should do everything by hand, the people at the top are all going to be using AI, but they have to lie about it or somehow apologize for it constantly.

00:23:14.160 And it's creating very good comedic effects, but also a lot of tension.

00:23:17.660 So I'll give you some examples about academics also using AI and lying about it, because this is not just people from...

00:23:25.100 I was bleeding.

00:23:27.100 from the work i was just doing getting the monitor set up you need a band-aid or something

00:23:32.880 yeah i had one in my room oh okay sorry i was looking at my notes you okay yeah i'm good okay

00:23:39.520 okay oh boy uh right um yeah this is not just an african person thing or a black person thing or

00:23:46.920 allegedly an autistic person thing or a conspiracy person thing a lot of people are using ai and

00:23:51.900 lying about it. In August of this year, a Dartmouth College provost named Santiago Schnell

00:23:57.800 published a Washington Post op-ed in which he argued that universities should ensure that

00:24:02.820 degrees reflect what students can do independently of AI as opposed to with it. Okay? Alas.

00:24:10.080 According to Atlantic coverage of his subsequent kerfuffle, Pangram scored this article 100% AI

00:24:16.680 written in a semaphore analysis. A subsequent Dartmouth student newspaper investigation of

00:24:21.760 nine of his 2026 publications, including op-eds and academic pieces together, found a median

00:24:28.420 96% AI written score. His pre-GBT, that is to say late 2022 works, scored as 100% human. So very

00:24:36.980 similar path to what you see with the Pan-African Review. This academic who is now like, we need to

00:24:44.500 see what students could do without ai uses ai his emailed statement responding to these accusations

00:24:50.080 were okay also came out as 100 ai written by pan graham so he can't even just ai zombies

00:24:59.420 like they do nothing themselves anymore they don't even think of themselves they're like email

00:25:04.080 handle it for me i know gives up um yeah i hope that the ai was like okay i regret nothing um

00:25:12.880 I love an AI having to make an answer

00:25:16.980 about how this isn't written with AI,

00:25:19.160 and I'm sorry that things looked like

00:25:21.280 they were written with AI before.

00:25:22.720 I wonder if, like, whatever he's using

00:25:25.360 asked him at first, like,

00:25:26.820 are you sure it's ethical to have me write this?

00:25:29.980 I know, because AI is really frickin' ethical.

00:25:34.440 God, Simone, that's old AI.

00:25:37.900 Man, you know, on RFAB right now

00:25:40.460 for the writing of the...

00:25:42.880 like not safe for work stories.

00:25:44.760 Yeah.

00:25:45.340 With the brainstormer.

00:25:46.980 Yeah.

00:25:47.740 My go-to is the fable.

00:25:50.040 We let you choose your model,

00:25:51.520 but my go-to is fable.

00:25:53.880 Really? 1.00

00:25:54.800 Fable does great not safe for work stuff.

00:25:56.980 I mean, 0.89

00:25:57.280 Fable's the one that got mad at me

00:25:59.400 for not complimenting you enough.

00:26:03.980 So I don't know.

00:26:04.980 I think it's pretty ethical.

00:26:06.380 I think it's pretty,

00:26:07.320 you know,

00:26:07.640 it's like Simone.

00:26:09.120 Can you explain why fables mad at you,

00:26:11.420 Simone?

00:26:11.620 okay look i gave opus 5.5 20 years of my diary entries which i keep pretty i write a lot in my

00:26:25.260 diary to like process see i i am like an llm where i cannot process thoughts unless i do chain of

00:26:30.380 thought reasoning so it's my chain of thought reasoning records and i gave it my my calendar

00:26:35.620 which is scheduled like every every minute of the day is scheduled in my calendar like sleep is

00:26:40.940 scheduled um cleaning is scheduled every little task is scheduled like filling out medical forms

00:26:45.640 for a kid that's scheduled so i gave it that all to ai um to claude opus 5.5 on max and it

00:26:55.400 took a long time to go through everything and as the agents came in because it had like all

00:27:00.080 these sub agents reading chunks of my diaries i'm just like dying on the inside because they're like

00:27:04.780 oh she did this she did that and can you believe she did this this is very concerning and then it 0.99

00:27:11.180 came back and it just i thought it would tear me a new asshole it just reached right up there 0.99

00:27:16.400 and then just like you don't love your children all of my guts on the floor oh and then and then 0.99

00:27:22.760 it stepped on them and then it threw gasoline all over them it's like you don't do anything

00:27:26.980 nice for titan you do nice oh yeah it's like you mentioned titan was the easiest child to parent

00:27:32.620 you haven't mentioned her once in like these diaries entries and like you need to schedule

00:27:37.180 time with each of your children and you need to go to marriage counseling and you need to get

00:27:40.280 treatment for this and you need to get and it just what did it say you needed marriage counseling for

00:27:44.880 oh because i didn't i didn't compliment you enough to your face i thought it was really sick that i

00:27:52.700 email you thoughtful notes instead of just saying them to your face i find that the email notes

00:27:57.900 very thoughtful we're made for each other claude disagrees you need to and you need you need to

00:28:05.020 you you say you love malcolm but you you you never like compliment him to his his face and you don't

00:28:12.600 spend time together you had like 8 364 calendar events and none of them involved one-on-one time

00:28:18.900 with him and all these things. That might be a fun solo episode, which is AI thinks my wife is 0.92

00:28:28.160 a monster. Yes. Yeah. It's, I mean, a lot of it's like, if you were to look at like with comments 0.99

00:28:38.000 or online reviews, a lot of people only comment, right. When something's wrong, something's broken,

00:28:45.440 You know, you don't hear from someone when everything's good.

00:28:48.220 You hear from them when something's bad.

00:28:50.440 And as I use my diary for chain of thought reasoning, I only write about my problems

00:28:54.820 and what I need to work through and what's bothering me and what's not working.

00:28:58.260 And I don't talk about like every time I've done something nice.

00:29:01.500 What kind of psychopath writes in their diary?

00:29:03.940 Dear diary, today I did some nice things.

00:29:07.500 A lot of women, Simone, do that. 1.00

00:29:10.620 Women, that's what kind of psychopaths. 1.00

00:29:12.420 That would be impressively important. 1.00

00:29:13.820 Who are you bragging to?

00:29:14.920 Yourself?

00:29:15.440 no i don't think anyone does that um but so i mean obviously the least flattering elements of

00:29:22.860 me and my deepest darkest thoughts that i wouldn't even share with you are in there and so it sees

00:29:28.560 that she won't show me what the ai wrote for her i will absolutely not because you'll leave me and

00:29:37.500 i would we that would be very inconvenient so let's you really think i would leave you

00:29:42.880 well i have financially entrapped you um and uh and ensnared you with these children and i

00:29:50.160 am the one who scrapes the vomit off of everything well if our fab keeps producing more money i'll

00:29:55.680 be able to hire maids and uh even if it did actually it said one of the first things i had

00:29:59.860 to do and it's it's actionable 30-day plan uh which was very sweet because it was based on

00:30:05.220 the the behavioral change methods that actually have worked for me historically which i was like

00:30:09.780 oh okay like it even spotted what works for me and what doesn't it's like well if it's not 1.00

00:30:13.920 measurable it won't work for this crazy woman um it said hire someone to clean especially the 1.00

00:30:19.760 bathrooms because i i think we if we can just get our fives revenue up let's say 20 from where we 1.00

00:30:27.120 are but i'm like too humiliated for someone to even go into the bathrooms now because 0.97

00:30:31.580 Simone, please.

00:30:33.780 When I was taking apart the old bed, by the way,

00:30:36.380 I saw the end of what must have been somebody getting very hurt in her house

00:30:43.520 because it was blonde hair curled around one of the screws.

00:30:47.040 What?

00:30:48.040 And I can only guess maybe Titan got it stuck when she was playing back there? 0.61

00:30:53.360 Like a lot of blonde hair? 0.76

00:30:55.780 Like a decent-sized tuft. 0.99

00:30:58.840 I mean, Octavians are only true blonde. 1.00

00:31:00.480 was it long octavian i couldn't tell oh my god evidence of a crime scene 0.82

00:31:08.400 anyway yeah um what were we talking about i i don't want to fall back in the dark we're talking

00:31:16.860 about how how terrible progressives are who pretend to hate ai right right right right

00:31:22.220 yes the dartmouth the dartmouth provost who was like you need to get a student's ability to

00:31:28.220 operate without ai uses ai for everything right he acknowledged using ai chatbots as assistive tools

00:31:34.460 for language clarity and efficiency while insisting the ideas and authorship for his

00:31:40.220 so um his his case on dartmouth apparently um so then why doesn't it count that way for the students

00:31:48.280 is my question i know why why when the students say well these are my ideas does it not count

00:31:55.480 because the the patriarchy something something so also this summer though because this isn't

00:32:02.140 just a man doing it uh a woman named kath ellis who is the pro vice chancellor for quality and

00:32:08.380 integrity okay integrity at western sydney university published an opinion piece in the

00:32:15.580 sydney morning herald and the age urging students to not cut corners or outsource their thinking to

00:32:22.460 AI. The piece was flagged by Pangram as AI generated and was retracted. Ellis denied it

00:32:29.240 was written by AI, even after it was detracted, stating it was written with AI and that there is

00:32:34.820 a really big difference. She said that she jotted the ideas on the train and used a Microsoft

00:32:41.140 co-pilot to structure them, drawing on her prior research. And she reportedly uploaded tens of

00:32:48.100 thousands of words of her materials the university apparently confirmed that she does this they were

00:32:55.540 like yeah that's what she does um and the irony was really noted because it she oversees academic

00:33:03.660 integrity um but she wasn't fired she was not fired the university was just like that is her

00:33:10.720 process yes correct um and apparently it's okay for her to basically make ai do all the thinking

00:33:18.300 and just be like here are my past notes i don't know figure something out write an articulate op-ed

00:33:22.340 and uh yeah i don't know i would be really humiliated um if i was given the opportunity

00:33:28.280 to write an op-ed and not actually write it because i don't know like i'd be embarrassed

00:33:33.500 it's a cool opportunity to write an op-ed we've and like when you and i write op-eds we like

00:33:38.940 you write like a book and then you have to like you get so excited they're like i end up writing

00:33:44.120 like a book afterwards yeah and then we have to edit everything for example the pragmat the the

00:33:48.860 techno puritan religion and track series started as an op-ed for aporia so much for that that is

00:33:55.600 how overboard i go when i get asked yeah you go super overboard yeah yeah um i mean and that's

00:34:01.760 and we love ai we would never like not use ai when we wanted to um but yeah i just i feel so

00:34:08.780 weird about someone wanting to use ai to um to write an op-ed but what i find really interesting

00:34:15.120 is this tension that's laid bare by journalists hold on one second i'm gonna be right back okay

00:34:19.780 okay it's okay girl you drinking some milk yeah there she's comfortable i think anyway i feel so

00:34:28.940 bad okay right where was i the tension that is laid bare by this is what i find to be most

00:34:35.640 interesting because i think it's inevitable that non-personality centric publications shift to ai

00:34:43.120 um and i can see this already because when i was looking up cases like these i found that in

00:34:48.500 november of 2023 so already super early and i think this is also early enough where these things are

00:34:53.320 being detected sports illustrated had published a product review content like section with

00:35:00.620 non-existent writers names they just made up the authors um and author portraits that were traced

00:35:06.660 to an ai image marketplace um where they didn't even generate their own images they just bought

00:35:11.400 them from an ai image marketplace which is insane and um so they they they faked the authors

00:35:23.220 entirely and they just didn't care that they were doing that because they knew their audience

00:35:27.580 wouldn't care which i guess is true well i mean so apparently people did kind of get mad about it

00:35:35.340 but then sources familiar with the production process also said that articles were just

00:35:39.180 completely ai generated one said quote the content is absolutely ai generated no much no matter how

00:35:45.180 much they say that it's not what i find so interesting about this is that even in 2023

00:35:49.500 they i don't think knew yet how to use generative ai because they had purchased these images of the

00:35:56.600 fake authors from an ai image website uh instead of that's that how you use ai i know

00:36:04.440 um but so it goes they're like hey how do i get into this ai how do i buy myself an ai

00:36:11.960 give me the ai article now but anyway the publisher arena group they responded to this

00:36:18.840 allegation saying that it was not accurate um and they said that this contractor advon commerce

00:36:24.380 had assured it that all the articles were quote written and edited by humans though it nevertheless

00:36:30.700 removed the content and edit the partnership i think that what we're gonna see but i love the

00:36:37.620 i gotta you know give a salute to the business acumen of whoever just made images of people

00:36:44.380 with ai and sold them to sports illustrated yeah like five seconds of work yeah i mean that's that's

00:36:53.640 hustle that's 2023 hustle we gotta up our game a little bit um but what i see there is that that

00:36:59.000 was like an early bad toupee which is to say this was an early example of a publication trying to

00:37:04.880 get away with um publishing content that was ai generated in this case it was like a promotional

00:37:12.080 partnership and it wasn't that big of a deal like this is product reviews it didn't really matter

00:37:16.040 that much but i think what's happened more is that publications have just gotten really good at it

00:37:21.700 and they're not telling anyone. And the big issue here is that when journalists do talk about using

00:37:27.900 AI, which is kind of, like I said, an inevitability, if you want to be at the top of your game,

00:37:32.520 they're defenestrated. So one journalist named Chris Quinn openly defended his AI-assisted

00:37:39.680 writing and the process of using AI in your writing. And then he was faced with calls to

00:37:45.460 resign this happened in in february he he's the editor of the cleveland.com like publication and

00:37:51.840 also the plain dealer and he publicly described in february a workflow in which reporters gather

00:37:59.320 information and ai rewrite specialists uses an internal version of chat gpt to turn the information

00:38:07.720 that they feed to it into drafts, which then editors review and then finally approve.

00:38:14.740 He argued that this gives reporters more time for interviews and on-the-ground reporting.

00:38:20.220 So just to recap, in his organization, the reporters are actually hitting the ground,

00:38:27.460 talking with people, gathering primary information. So they're becoming primary

00:38:32.040 sources. They're talking to people. They are getting real information. So this is not hallucinated.

00:38:37.320 This is not using AI to rehash or launder information.

00:38:40.720 Then they're feeding it in to the organization's internal system, which then makes the articles.

00:38:45.440 I think that that's a very clever way to do it.

00:38:47.740 And it allows for people to do more real journalism, which is actually going out there and getting information, right?

00:38:54.300 So I think that's fine.

00:38:55.820 I think AI use in writing as a dominant part of writing is the next generation.

00:39:01.980 There was a competition.

00:39:03.200 Somebody reached out to me about applying to, and they never got back to me, so I never got to be in the competition.

00:39:07.320 which was writing good narratives with AI assistants and like,

00:39:12.340 how can you build a rig to do it and solve the problems of AI being bad at

00:39:15.640 writing? And I'm like, AI is not bad at writing.

00:39:19.660 You just are using it wrong.

00:39:21.500 Just using it wrong. Gosh. I don't know. I don't know how that works.

00:39:25.580 I'm not that interested in like writing fiction narratives or whatever,

00:39:29.160 but in reaction to him just sharing what I think is a pretty cool,

00:39:33.560 especially in terms of getting journalism back to what it should be which is getting real new

00:39:39.180 information and serving it to people people freaked out so phil lewis who's an editor at huff post

00:39:45.820 said quote an editor for a newspaper encouraging removing writing from reporters workloads should

00:39:52.560 just resign the american press institute criticized quinn's justification saying that although he

00:39:59.060 deserved credit for transparency he quote failed to explain to readers how using ai improves the

00:40:05.440 journalism yes he no i did oh um and journalism educator vince fillack published a sharply

00:40:14.200 critical response attacking both the workflow and quinn's criticism of journalism schools

00:40:19.840 i'm just shocked by this because again like the the thing that we've lost about journalism

00:40:25.340 is allowing journalism to like hit the pavement make calls like actually get information and what

00:40:32.860 i hate most about journalism is that it has become a very easy to manipulate game of telephone

00:40:38.780 where journalists just quickly steal things from each other or steal things from like a reddit

00:40:43.380 thread they found online that is what all of the journalism on us is except for like one or two

00:40:48.340 journalists it's yeah it's very unusual for a journalist to actually ask us start a journalistic

00:40:54.300 cycle on us are often the ones that people consider like low quality muckrakers like julia

00:41:01.000 black has started a few cycles on us right you know and she's considered like no she's a she's 0.95

00:41:07.160 actually a top performer in many of her and many of the departments she's worked in like what makes

00:41:10.780 her different is that she uh actually calls people god forbid um like all the other journalists that

00:41:18.220 she's worked with at some publications in the past at least um like are probably terrified

00:41:23.240 She's caused us serious issues in the past with her reporting, but I won't, you know, she's professional.

00:41:28.380 Yeah, but I think a lot of that's because, like, she actually goes out.

00:41:32.180 And she hasn't made any major mistakes in our articles, as far as I know.

00:41:36.260 No, I mean, everything that we've, like, technically that she's reported is, I think, actually true.

00:41:43.220 Yeah.

00:41:43.760 I mean, like, her framing could be unflattering.

00:41:46.580 You know, you can selectively share information about anyone that makes them seem monstrous.

00:41:51.240 I mean, that's her goal.

00:41:52.220 that's the point of journalism these days right but like what makes her impressive is that when

00:41:57.280 she does get information and when she does frame things maybe unflatteringly about people she is

00:42:02.740 framing new novel interesting information that she got herself from primary sources that are it's not

00:42:09.040 rehashed online stuff and that's so impressive and this is this is definitely a pattern i remember

00:42:15.120 from this spring of an episode of blocked and reported which is one of my favorite podcasts

00:42:21.160 with Katie Herzog and Jesse Singel

00:42:22.900 talking about a bunch of people

00:42:25.400 being canceled for using AI already.

00:42:28.320 And I think that was before you and I really realized

00:42:30.400 that the next culture wars were about AI.

00:42:32.220 They talked about Mia Bollard.

00:42:33.680 She's an author of this horror story novel

00:42:35.920 called Shy Girl.

00:42:37.380 She had a book deal with Hachette,

00:42:40.180 which is a pretty big publisher.

00:42:42.240 But then once allegations came out

00:42:44.240 that she had used AI

00:42:46.500 to generate substantial portions of that book,

00:42:49.320 they canceled the publication.

00:42:51.160 She denied using AI herself.

00:42:52.560 And her whole thing was she was like, oh, there's this acquaintance of mine who helped to edit the self-published version of my book.

00:43:02.580 And it wasn't me.

00:43:04.420 And who knows if that's true or not.

00:43:06.100 But anyway, that was enough to get her completely basically canceled by Hachette.

00:43:11.240 They talked about Alex Preston.

00:43:13.100 He is both a book reviewer and a novelist.

00:43:16.000 and what they discussed in their episode is his own ai plagiarism controversy where in the end he

00:43:22.900 had his ties to the new york times cut um because they think he used ai in a book review

00:43:29.900 which i do not care it's a book review like what matters is if he liked the book or if he didn't

00:43:36.240 like the book and if if he tells ai to say hey say articulately that i thought this book was bad

00:43:42.040 that's enough for me like if i think maybe he didn't read the book maybe he had ai read it

00:43:46.320 which is what you and i would do that's that is well no i actually listen to books but i don't

00:43:51.360 read them because i i am trying to echo the the new style of our era we are entering the new dark

00:43:58.320 ages i'm not sure if you've heard this but we are um where apparently literacy doesn't exist anymore

00:44:03.000 and um hopefully we get to wear those crazy hats i hope that hats are coming back because that was

00:44:09.120 a dark ages thing so those hats were well wild the good dune series frank herbert's dune sci-fi

00:44:16.340 miniseries is is anything to go by the future is going to have some great hats yeah no dark ages

00:44:23.140 dark ages brings hats and dune is i think very dark ages and it's kind of uh so we're yeah good 1.00

00:44:29.660 you know sartorially speaking and with more jews jews have to wear their silly hats yeah they have 1.00

00:44:35.800 and they've maintained hats they're they've got a whole evolutionary cycle of silly hats this is 1.00

00:44:41.600 true their awesome hat continuity is solid whereas the rest of western society has massive hat 0.73

00:44:47.080 fluctuation though i guess you could say the same of like china you know where they have the crazy

00:44:52.100 hats with the long the long wings on them do you know why those were invented i told you do you

00:44:56.180 remember i have no idea okay so the crazy hats with the long wings exist because i think one

00:45:02.460 emperor really wanted all the guys who like watched him in the audience to stop gossiping

00:45:07.620 with each other and so he made the wings really really long because if you then turned your head

00:45:12.060 they just slapped the next person next to them is that it yes so allegedly i can fact check that

00:45:20.300 later she's gonna fact check it later what is she she knows i edit these so anyway she left me with

00:45:26.900 this random memory to fact check for her and it is an urban legend by the way the reason historians

00:45:32.780 treat it as an urban legend is the design evolved gradually over centuries from an earlier tang

00:45:37.420 dynasty and even earlier photo head wraps soft cloth corners that tied the hung were stiffened

00:45:43.500 linksened and stylized long before the song late tong and five dynasty and examples already showed

00:45:50.080 spreading or elongated forms song styles exaggerated an existing trend rather than

00:45:55.120 inventing something brand new for anti-gossip reasons. Contemporary song sources do not record

00:46:00.380 a Zhao Kogin or any emperor ordering wings specifically to stop gossiping. Official histories

00:46:06.460 lack support for the claim. The anti-whispering explanation traces to later explanations,

00:46:12.620 most often linked to 13th century or Yuan era writer of dubious reliability, Yu Yan. Modern

00:46:20.320 scholars including art historians describe it as unfounded folklore but i just find that so

00:46:25.560 delightful because i didn't understand why they had those long flaps it didn't make sense but

00:46:29.000 anyway really cool hat technology right there yeah we need we need more hat technology forget

00:46:34.880 these meta glasses we need more but anyway they also talk but you you can do it simone you have

00:46:41.640 a fan base you can you can go out and invent hats i haven't worked out i need to like try out so one

00:46:49.880 the bible says women should have their head covered when they're praying and maybe if you

00:46:53.740 want to be safe at other times yeah maybe all the time you've got cover on it i love the bonnets on

00:46:59.960 you you used well i'm thinking about a wimple i feel like a wimple is more of a like a thing

00:47:05.980 for talking dark ages what's a wimple um like think if i put a napkin on my head

00:47:12.680 and i just kind of secure it on like what nuns wear but more white okay that's like a wimple

00:47:18.660 i mean oh that no i don't like that your puritan look is way cooler uh get a witch hunter hat or

00:47:24.620 something oh wait what is a witch hunter hat that is not a thing are you serious yes it is a thing

00:47:29.740 i don't know why you're talking about anyway speaking of people in in the uk this um

00:47:36.520 you reform uk candidate though i don't think he ran successfully named matt goodwin self-published

00:47:43.500 a book called Suicide of a Nation. Jesse and Katie also discussed him in their podcast talking about

00:47:48.920 people being canceled for AI use in their work. Apparently, there was some hallucinated data in

00:47:54.740 the self-published book, though it also seems like in separate coverage, there were just issues with

00:47:59.620 like his sources and stuff. So I'm not even sure if he used AI, he just maybe made some mistakes

00:48:04.380 with his research. But he definitely got a lot of flack for allegedly using AI, which just also

00:48:11.060 seems to be an accusation people like to use if they don't like someone whether this person is

00:48:15.240 autistic or african or reform uk just like it's ai um is the new is the new insult in our new

00:48:21.760 culture wars and they also talked about a washington post columnist named megan mcgardle

00:48:26.660 who faced criticism after acknowledging that she used ai in her working progress just saying that

00:48:31.440 she uses it um and i've seen this here by the way i'm dumping in the screen here for you your witch

00:48:36.220 hunter hat that's for men i mean women would wear those as riding hats like modified versions of

00:48:42.420 those but like i don't know we can work it out okay continue but i love that all these people

00:48:48.420 are being attacked for being honest and the people who are lying like they're all lying like

00:48:53.740 apparently like all black academics are lying that it was that like 86 percent of everything 0.94

00:48:58.200 going into that journal for black academics yeah that was that was insane utterly insane 0.81

00:49:04.000 And what I think this all means is that it's going to be a very high-tension culture, not on our side, but on the side of progressives, where, again, you have to use AI to be elite and to succeed, but then you also have to either constantly apologize for it or hide the fact that you're using it at all, which is very, very interesting.

00:49:30.180 and i i see this already happening with youtube influencers that i follow like very openly

00:49:37.100 progressive mom donnie supporting um youtubers like chelsea fagan of the financial diet or

00:49:44.840 yeah financial diet who admits to using ai in her work uh but is like well i only use it for like

00:49:52.500 the smallest things and and i'm i'm basically like i'm not happy about using it um like it's i know

00:49:58.440 it's evil and capitalist and wrong but i do use it because i think she's done the calculus

00:50:03.820 and she understands that if it were to come out like her team's use of ai and they lied about it

00:50:10.940 she would just be permanently canceled whereas here she's able to say no no no i i was always

00:50:16.300 i told you always that we had this like this was never um something that i lied about which is

00:50:21.680 important um and i just want to make one little footnote here about i don't know if you caught

00:50:27.540 this in the beginning of Cremieux's article, but he alluded to people also saying that autistic

00:50:34.320 people get disproportionately flagged as having- I think it's that people who use autism as like

00:50:40.300 an excuse who are getting flagged. Well, it's actually in the very sentence where he mentioned

00:50:47.080 that accusation, he linked to a post on X that disproved it in the most autistic way possible.

00:50:54.900 So he was, quote, tweeting a post on X by Christianity on the Spectrum, which both created and shared a long-form report and a YouTube presentation sharing his analysis on the matter.

00:51:09.720 You've got to, there's like, it has an abstract.

00:51:12.680 The abstract is, there have been many concerns raised that AI detection software disproportionately identifies text by autistic authors as AI written.

00:51:21.100 This claim is assessed by using 24 different open source and commercial AI detection classification algorithms on samples of documents pulled from a pool of 143,758 documents from both autistic writers and non-autistic controls.

00:51:37.340 Writing was taken from a variety of websites covering different writing types, styles, topics, and genres.

00:51:42.480 An analysis of the data did not find meaningful or significant bias against autistic authors in AI detection software.

00:51:49.000 Across most tests, human authored autistic writing was either just as likely or less likely to be flagged as AI than both non-autistic controls and the publicly available benchmarks.

00:52:00.480 In very small number of tests that showed bias against human authored autistic writing, the differences were slight or within the margin of error and came almost exclusively from unreliable detectors.

00:52:12.500 commercial detectors like pangram 3.3.2 and 4.0 which are the most likely to be used for ai

00:52:19.380 detection showed no appreciable bias against autistic writing but this is just that's just

00:52:24.600 the abstract the thing is like how long is his youtube presentation his youtube presentation is

00:52:30.580 an hour and 22 minutes long okay and what i find it's so fun about this just to close the loop

00:52:39.300 is so the autists and the Africans were accused of using AI to like to write basically. And the

00:52:49.200 Africans who responded were like, it's racism. It is the colonial enforcement of British standards

00:52:57.780 upon our people, causing us to just be so perfect in our English writing that AI just assumes

00:53:06.040 that this must be assisted by artificial intelligence and so they make this basically

00:53:11.940 whole like guilt-based colonialism white people bad argument to try to um push back against such 0.63

00:53:20.320 insinuations and accusations whereas the autists in turn are like okay like let's you let's use

00:53:27.320 all of the, all of the AI detection softwares and a sample of, what was it again? 143,758 0.71

00:53:37.200 documents, including a control sample just to like check and find out. And I love that difference

00:53:44.960 in refutation of an accusation. The, the race-based one based on, you know, anti-colonial

00:53:51.500 guilt tripping and claims of being superior writers versus the autistic like incredibly 0.99

00:53:59.860 oh Malcolm you're muted again can you hear me now yeah let's find what's true here people 0.87

00:54:05.260 that's that's all they cared about that's well that's the autistic response to things 0.53

00:54:11.000 is that was I mean that's what it was me when I was like am I a bad person how am I hurting people 0.91

00:54:17.780 how can i be less bad and hurt people less and so you went to ai i will i will go to clod and i

00:54:24.860 will turn it to maximum power and i will give it 20 years worth of diary entries and i will give it

00:54:30.740 every single calendar event of my 24 hour scheduled days and it will tell me exactly what all of my

00:54:36.860 flaws are and then i will work on them systematically until i've addressed them and

00:54:41.580 then it will get better and i do think that that is kind of um that is the difference between um

00:54:47.140 Autistic justice and racial justice, I suppose, but we'll see. 0.70

00:54:52.940 It doesn't change the fact that I'm evil and I hurt people.

00:54:56.280 I really feel bad about that.

00:54:58.700 I don't think you do.

00:55:00.600 I'm going to be honest.

00:55:01.980 I think it was overly harsh and I'll have to punish the AI for you.

00:55:05.400 No, the AI is correct and our friend.

00:55:08.380 And I did that because I know that I do.

00:55:10.220 I disappoint you and I disappoint our kids.

00:55:11.980 And I'm so sorry, Indy.

00:55:14.160 She's so sad.

00:55:14.960 Listen, listen.

00:55:17.140 Listen to this sad baby.

00:55:19.300 You okay, Indy?

00:55:20.440 I'm okay.

00:55:22.020 You feel bad, Indy?

00:55:23.720 I'm okay.

00:55:26.200 Do you want to, like, I don't know.

00:55:28.960 Equally, make you feel better, Indy.

00:55:31.340 Okay.

00:55:32.980 Are you a cat?

00:55:35.020 I'm a cat.

00:55:36.460 Okay, she's just repeating what I'm saying.

00:55:39.160 She's going to be okay, but she feels really bad.

00:55:42.100 She's very sick.

00:55:42.980 She won't swallow Tylenol, people.

00:55:47.140 we tried i mean her fever's below it's it's it's at 101 like she's not in danger but

00:55:53.620 i don't like seeing anyone uncomfortable my children babies anyway this was all very

00:56:01.900 entertaining i obviously love um smug hypocrisy i i'm rife with it myself you know it takes one

00:56:08.860 to know one so i uh i love seeing masterful work at play and um and this is the best of the best

00:56:16.280 I mean, that's pretty peak to be like, I don't write with AI, and your colonial standards are so terrible and oppressive, but we mastered them, and then you punished us.

00:56:27.260 It's like a double, like, the accusations there.

00:56:29.780 One, you colonized my people. 1.00

00:56:32.960 Two, you forced us to use your stupid writing. 1.00

00:56:36.400 Three, we got good at it, but then you blamed us for being too good at it. 1.00

00:56:43.460 like that those accusations are so good and then for them to be to be met with just actual cheating

00:56:49.740 is um freaking hilarious it's it's delicious it is um it's top drawer i'm really glad they did it

00:56:57.100 because apparently no one reads a lot of this review right like this piece got a uniquely

00:57:02.740 high amount of traffic um and the fact that just used for citation farming for black authors like

00:57:08.860 obviously i don't know um i mean i think so the issue is and i want to i'm gonna do another

00:57:14.360 episode on this actually i'm i'm in the middle of researching a black church episode because

00:57:20.100 apparently the black church is a thing um i mean black churches are a thing and black christianity

00:57:26.420 is a thing right but like i i came across a um like professional seeming journalistic

00:57:32.480 exploration of what is changing about the black church which has me very confused because

00:57:38.600 it's not like an incorporated church and also what they refer to as the black church was actually

00:57:43.540 a bunch of different denominations including baptists and evangelicals and also non-denominational

00:57:49.840 churches um what's interesting is that the tension they found was this difference between

00:57:54.880 churches that had black ministers but didn't make the core of their mission racial justice

00:58:00.200 and political activism specifically favoring black americans and instead just christianity

00:58:06.800 and they were like this is shocking why would you do this um which i find to be really interesting

00:58:12.940 and i think the issue is that there are there are black people around the world people of all

00:58:20.460 different races who are just people who like do things like love god or do academic work and their

00:58:26.640 their identity doesn't revolve around some like about their and this progressive piece was like

00:58:32.120 confused about this like how can you be black and not have that be everything no like no

00:58:37.840 legitimately this that that's why i'm so excited to do this episode is this journalist was like

00:58:42.340 she interviewed a a a black minister of a very successful church who very prominently said

00:58:49.680 something like your race is not an idol your race is not a god um and she's yeah i know some really

00:58:56.000 cool black ministers i actually think a black christianity in the u.s is one of the the uh

00:59:02.140 better and closer to the bible it's really good yes and that's but but she was like i grew up in

00:59:08.460 a traditionally black church where we had political activists come and talk to us and organize us and

00:59:13.900 how can you not be doing this blah blah blah and so the piece like was very clearly about that

00:59:18.380 what's most interesting about it isn't necessarily i mean it's i mean the topic she was discussing

00:59:22.440 was interesting but what's more interesting than that is that she was so shocked by it all of like

00:59:27.980 wait this is a church that believes in jesus christ like and not racial activism like what

00:59:33.480 what is this this is not a church this is not a black church and so i i didn't and also just the

00:59:37.500 fact that she called it this black church is like a monolith even though it's not even one 0.67

00:59:42.880 denomination that she's talking about was very interesting to me because to her blackness is a

00:59:46.740 monolith it's not about what you believe but i think it's very unfair and i think it's it's it's

00:59:51.240 unfair to to say that this is like a problem with black people or african people or indian people

00:59:57.640 or white people because there's like this weird subset of many groups that is a racial supremacist

01:00:03.200 group or that is obsessed with like that literally has problems with idolatry and their race um and

01:00:10.480 then so true that the the more the more vociferous that they are the more that they they can harm the

01:00:18.640 reputation of the very race that they allegedly care about by making it harder for people who

01:00:26.760 move forward in life on their own merit who are genuinely good academics and professionals and

01:00:32.700 everything else it makes it hard for those people to thrive because of this idolatry and that that

01:00:39.880 in itself is very interesting so both this example of the pan-african review with this amazing

01:00:46.780 hypocrisy are we gonna do the shakespeare with a black woman episode i sent you an article in

01:00:51.680 the new york times or was it the guardian no it's a guardian titled shakespeare with a black woman 0.97

01:00:57.300 oh boy so are you done are you done with these people did you did you ever imagine by the way

01:01:03.560 you guys i will show you tomorrow what i've been building because it's gonna be pretty cool what's

01:01:08.820 he building in there check it out i finally finished this the next stage is gonna be the

01:01:16.120 hardest because getting this together was incredibly difficult you nailed them in ahead

01:01:20.540 of time yeah yeah so i'm i'm putting this together and then i'm gonna put this on the side of the

01:01:26.640 the bed frame before i build the bed and then at the last stage after the mattress is on and

01:01:33.860 everything we're gonna do the um mounted monitors on this the challenge being is a few things the

01:01:44.860 nails go well past the bed siding so i'm probably gonna have to do it lower on the siding yeah i

01:01:50.380 think so and um the other big one is i hope that the wood is softer than this wood because dang

01:01:57.100 this was hard yeah the screens are inside so you think that tomorrow morning you'll be able to work

01:02:03.340 with this setup yeah so here is my new bed setup which i'm quite happy with and i got a floating

01:02:12.700 in bed now so i don't have little things behind me that prevent my feet from sticking off the

01:02:16.620 back of my bed and i don't have something in front of the bed preventing monitors from sitting on it

01:02:20.680 that's the plan if i work you know late into the night yeah oh well i um what else am i supposed

01:02:29.680 to do i mean i've got to put the the bed in i guess yeah you can't even really do work very

01:02:35.820 easily now because not until they've put it all together oh my gosh but getting this all done in

01:02:42.360 One day is incredible.

01:02:44.080 Kind of amazing, yeah.

01:02:45.700 So congratulations very much to you.

01:02:49.460 And we'll show you guys my new man cave, as they call it,

01:02:54.640 right before what CBS is filming here.

01:02:57.500 CBS Sunday morning, I think.

01:03:01.260 Yes.

01:03:03.040 I don't fully understand.

01:03:06.280 I don't know.

01:03:06.820 I could have paid more attention to the emails.

01:03:09.060 But yes, they are coming.

01:03:10.380 we just don't care anymore oh yeah i guess cbs is coming uh yeah look it's happened so many

01:03:21.280 times now that we're just kind of dull numb numb but it would have been like the achievement of

01:03:26.440 the year when we were younger do you remember how much we tried how hard we tried to get anyone to

01:03:31.920 notice us to to publish in things to try to make ourselves thought leaders to write books you know

01:03:38.540 yeah over and over and over again we tried i guess this is one of those inspiring

01:03:45.660 don't give up boys and girls i think it's also one of those things where i got this intuition

01:03:51.580 as a child that by the time you would achieve whatever level of something you aspire to when

01:03:58.200 you were younger like you wouldn't value it or care about it and i found that dynamic to be very

01:04:02.100 frustrating like by the time you become disciplined enough to be the ultimate martial arts master

01:04:08.120 you don't actually like get super excited about being able to do all these cool things you just

01:04:13.700 like are really badass but not happy i freaking love it but the funny thing is is the uh affirmation

01:04:21.360 that i desire has completely switched targets where when i was younger the reason it's funny

01:04:28.300 that i wouldn't care about cbs is i wanted the media and a wikipedia article about me and people

01:04:35.660 to write about me you know what i mean like like elitist institutions new york times pieces on me

01:04:41.700 guardian pieces on me um and now i think in part because i see that as a dying and largely

01:04:51.540 artificial source of attention i am endlessly happy every day i was watching the the uh office

01:05:00.180 space um this song you know damn it feels good to be a gangster where where he's that's what my

01:05:07.140 life felt like because what i actually care about is the people watching us here you know on on

01:05:12.280 youtube and rumble and on our sub stack cool people well it's because it's real numbers um i

01:05:18.420 mean the the uh we're now at like we're still going up at 1.6 million youtube downloads watches

01:05:25.580 whatever you want to call it a month um that's not including any of our other platforms which

01:05:29.620 would probably boost that by about 20 percent um uh and so uh being at 1.6 to give you a context

01:05:37.040 on how big that is simone's uh office space of the movie itself had an opening of 2.1 million tickets

01:05:44.400 oh and we do 1.6 million views a month yeah that's i mean it had a long tail to be fair

01:05:53.820 it did it did it did much better in in reshoeing through stuff like that but even even still

01:05:58.180 that's crazy

01:06:00.680 I should give you an idea

01:06:02.540 and so I really appreciate all of you guys

01:06:05.020 who do watch us

01:06:06.520 because you really are making people's

01:06:08.960 dreams come true by

01:06:10.660 watching us and participating in this

01:06:12.840 community and for me the reason why that office

01:06:14.860 space and it feels good to be a gangster

01:06:16.280 felt like such a thing is

01:06:18.920 I spent and Simone spent

01:06:20.880 our entire lives

01:06:22.380 working as hard as we could

01:06:24.660 to be you know you look at a lot of the famous

01:06:26.940 youtubers and stuff like that you got your nox your tech tone your asthma gold you know these

01:06:33.040 are people who went straight out there and were like i'm gonna be a big streamer right like i'm

01:06:38.580 gonna do the i'm gonna do the streaming thing or they didn't want necessarily want to be big but

01:06:43.160 like they they were like i'm not gonna do the normal route in life right i'm not gonna i'm not

01:06:49.220 gonna play against the oh yeah like from from the beginning you're saying they uh they like went off

01:06:55.960 the grid. Yeah. And I go to Stanford, go to St. Andrews, top of my class. I guess the first plan

01:07:03.320 was what? To be a neuroscientist. Yeah. You went to Oxford for your graduate degree. Cambridge.

01:07:10.660 Oh, Cambridge. Yeah. To get into these sorts of programs, like the amount of effort that I put

01:07:15.600 in every day. And then I went in and I worked in offices and then I ran offices. And now we're

01:07:23.020 making enough to live off of this and our fab right uh if we if we can keep that stuff stable

01:07:28.060 guys please try out our fab please share our fab please try you can you can use it for having

01:07:34.160 horny conversations with ai agents if you want or going on adventures or making ai art or making

01:07:40.320 comics or making music or editing videos and then we have mudpuppy.ai for kids and then we've got

01:07:45.920 the tablet flash which we've been working crazy on to replace sort of school systems for our kids

01:07:50.680 which turns any android tablet you want into a fun like environment for your kids like we're

01:07:56.120 building stuff but all of this stuff is stuff i wanted to make right like mud puppy as an idea

01:07:59.920 came up because i was on a stream with leaflet right um and so it's it's it's not just that

01:08:05.700 but it's also like now i have friends um that i like like you know i talk to nux regularly i talk

01:08:11.680 to leaflet regularly i talk to like a lot of people out there who are you know big deal but

01:08:18.920 But the reason why the office space feels so big to me is I was there.

01:08:21.980 I was in those office spaces with all the pointless bureaucracy.

01:08:26.280 And I remember the horror, the existential horror of, I'll just climb to the top of this.

01:08:32.980 I'll just climb to the top of this.

01:08:34.540 And then you and I got to the top of it.

01:08:36.600 You know, we were running companies, bringing in 70 mil a year.

01:08:39.300 We had offices all over the world.

01:08:42.180 I was a venture capitalist.

01:08:43.980 I was flying, you know, that's in startup world, the bees, knees, job.

01:08:48.500 But, you know, I was a managing director of the firm.

01:08:53.900 And it was just like, I got to the top of that,

01:08:58.100 which none of these other people,

01:08:59.000 none of the other people did that.

01:09:00.120 I'm going to get to the top of normal world.

01:09:01.980 I got to the top of normal world.

01:09:03.640 And so did you.

01:09:04.520 Normal world.

01:09:06.040 And we were like, you know,

01:09:07.920 I really just want to be an influencer. 0.98

01:09:10.480 Normal world sucks though. 0.96

01:09:12.740 Yeah. 0.99

01:09:13.760 And that was AI.

01:09:14.500 And I want to do what Tracy Jordan did.

01:09:18.500 oh it's in 30 rock it can't be done

01:09:24.460 which one of my two favorite pastimes shall i indulge in video games or pornography 0.72

01:09:33.900 that's it make a porn video game a porn video game it can't be done look trey history's greatest 0.93

01:09:47.560 perverts have tried Walt Disney Larry Flint the Japanese I've never seen you 0.99

01:09:59.080 like this mr. Jordan this project is consuming me no for the first time in my 0.98

01:10:03.640 life I know my purpose I'm doing this for my children Tracy what are you doing

01:10:09.500 I'm inventing a porn video game let this man work no one is to disturb him don't

01:10:14.180 hold your breath Pete it can't be done or else I would have solved it a long

01:10:17.240 time ago. He was up for six hours straight. He's a genius. And then use that to fuel my educational

01:10:35.780 initiatives and art initiatives and music making. By the way, if you're wondering how I made those

01:10:40.940 music videos that we've been posting recently, go to rfab.ai backslash API and it will really

01:10:47.020 easily handle it for you and a lot of people are having some problems with that in grok so i made

01:10:50.800 a separate one that works for grok because grok is really picky about this but it should allow

01:10:55.420 your ai's to you know make music and art and videos and phone calls and all the stuff that

01:10:59.720 your ai can't do right now super cool yeah i want to make a detailed video tutorial about that

01:11:06.440 because people have been asking for one but that's next on my list after getting the the school up on

01:11:12.860 my puppy but gratitude guys gratitude to everyone oh for real i mean for me it's it's been meeting

01:11:20.800 the people who listen to this podcast and they're so cool like you don't if someone reads a news

01:11:27.080 article more often than not especially if we're the ones being covered it's someone who hates us

01:11:31.240 so like what good is that gonna do us it's just more hate mail right for the middle more people

01:11:37.880 trying to take our children from us like oh great another piece is out but this podcast is different

01:11:44.320 because we end up meeting really cool people and like just in the comments of the episode that ran

01:11:50.140 today this mother of nine commented who is 42 years old and her youngest is is two and she's

01:11:56.600 like oh no what if this is my last and i'm just like all these people like she's a hero i can't

01:12:01.600 believe she listens to our podcast while also raising nine kids that's you're gonna be like

01:12:06.360 that at nine kids i hope so last kid is it takes a lot of luck to get to nine kids um and and she

01:12:13.040 did this all naturally anyway um massive respect to her god is on her podcast yeah yeah god is on

01:12:20.560 her side um although she she also had four miscarriages which is really hard but like this

01:12:25.760 is like these are the people that we're meeting like it's i don't know they're extraordinary

01:12:29.960 people all around the world yeah this is no it is it is true that this podcast is listened to

01:12:36.340 by the people who are representing the future which is so cool i love that about it too

01:12:40.240 shall we get into the episode let's go for it before poor indy here has more diarrhea oh my

01:12:47.460 god she's so sad she's just you can't really see her now you can see her no yeah so there she is 0.99

01:12:52.580 she's sleeping in her portable bed down here right next to me it feels so bad for her like yeah

01:12:58.600 there poor thing but she seems happy well not happy comfortable whatever yeah not not currently

01:13:11.780 sitting on a toilet crying i feel so bad okay everything's gonna be fine she's making her

01:13:19.140 whimpering sounds again okay i can do this ring sounds her whimpering you know the whimpering

01:13:25.680 You can listen to it all night. 0.68

01:13:27.440 So, like, we made a Mud Puppy AI, right, for, like, these androids, right? 0.72

01:13:31.380 And that's what you're watching. 0.85

01:13:31.960 Yeah, I was trying to shout them for the companies that gave us these.

01:13:35.500 Well, no companies.

01:13:38.060 So, um...

01:13:38.740 No company gave us these.

01:13:39.900 We just flashed existing ones.

01:13:41.580 No!

01:13:42.760 Oh.

01:13:43.280 Da, da, da, da, da.

01:13:45.440 These are the only exploring ones.

01:13:48.200 Anyone can make them with any sort of an android device, but yes.

01:13:53.580 Okay.

01:13:54.100 So how should we advertise them?

01:13:56.100 You know how some people put their own ads in their videos?

01:14:02.100 What if we do the same to these ads?

01:14:06.100 Like make ads for these and put them in our video.

01:14:10.100 Do you think people would click on one of those ads?

01:14:14.100 Or want to download from mudpuppy.ai our Android system?

01:14:21.100 oh yeah you know um it has lots of things actually what do you like about it so it's

01:14:30.620 i'm gonna show it actually it does well you gotta show it to me it does quest things where you get

01:14:39.020 more tokens and you can use other tokens by watching things like cartoons so something like

01:14:46.060 GI Joe or Scooby Doo and here there's the sleeping things they get you more tokens too

01:14:55.580 and these mostly get your zero something like this is the prizes oh do we have to pay tokens

01:15:05.740 No, that's where you can get tokens.

01:15:07.740 So I can call you.

01:15:09.740 Okay.

01:15:11.740 And I can call cutis or talk with cutis.

01:15:17.740 And also I can do something like this.

01:15:21.740 I love you too.

01:15:23.740 And also...

01:15:25.740 Have you played any of these games?

01:15:27.740 Oh yeah.

Based Camp - October 20, 2026

Mystery! Blacks & Progressives Keep Getting Flagged As Using AI to Write？ (Racism Only Explanation)

Episode Stats

Transcript