DarkHorse Podcast Summary
Episode: Where has all the science gone? – The 290th Evolutionary Lens
Hosts: Bret Weinstein & Heather Heying
Date: August 21, 2025
Episode Overview
In this episode, Bret and Heather use their evolutionary biology background to examine three pressing societal topics:
- The cultural and evolutionary implications of AI-generated fashion models (inspired by a tweet from Elon Musk).
- The controversy surrounding Michael Tracey’s stance on moral panics in the context of the Epstein case and a related legal incident involving an Israeli official.
- The alarming rise of scientific fraud, as revealed by a recent paper in the Proceedings of the National Academy of Sciences (PNAS), which exposes the scale and mechanisms of systematic defection in science.
Throughout the discussion, they explore the breakdown of trust in institutions, the consequences of technological substitution in human relationships, and the collapse of scientific norms, highlighting the general theme of society "losing its footing."
Key Discussion Points and Insights
1. AI Fashion Shows and the "Obsolescence" of Sexes
(Starts ~14:04)
- Elon Musk's Tweet: The episode opens with Bret introducing a tweet from Elon Musk, which showcased AI-generated images of female models from various Large Language Models (LLMs) like ChatGPT, Gemini, Grok, and more.
- Bret: “What we are beginning to see is effectively women are having men replaced by AI and men are having women replaced by AI. In a sense that matters deeply.” (15:29)
- Human Relationships vs. AI Surrogates:
- Bret and Heather discuss the evolutionary importance of partnership, specifically how sexual relationships historically forced negotiation and understanding between the sexes.
- Bret: “The key to the human species is partnership around sex and marriage and the partnering in the raising of offspring that results from the navigation of a space that is... adaptive in the fact that you don't get exactly what you want, that you have to navigate with a real person.” (24:16)
- Societal Risks:
- The hosts warn that widespread technological surrogacy could lead to a "bifurcation of the species," referencing incel and MGTOW movements as early signs, and mapping the replacement-of-human-functions trend onto robots in domestic and industrial roles and AI in intellectual roles.
- Heather: “Right at the top when you start talking. I objected to the use of the word women. Like, that's not women. We didn't see women there. Just like trans women aren't women. And part of how the trans madness took over..." (20:33)
- Loss of Skills and Specialization:
- They trace a parallel between this technological obsolescence and economic specialization/hiring out of domestic roles, suggesting both hasten the erosion of self-sufficiency, family ties, and intergenerational learning.
- Heather: “The kinds of work that men have historically done within a partnership have been outsourced earlier than the kinds of work that... women do. We hear a lot about how feminism has failed the American family... but by this analysis, it's actually being able to hire men to do the kinds of work that men do within the family unit...that actually helps slide us into the expectation that... the family unit isn't going to be self-sufficient.” (25:34)
Noteworthy Segment:
- The Danger of Losing Cultural Knowledge:
- Bret: “You can lose important culture in one generation. You just can, right? You can be a seafaring people who depends on ancient wisdom about how to make kayaks. And you can make landfall and you can go inland far enough that nobody needs a kayak and you can forget in one generation how to do it.” (32:02)
2. Michael Tracey, Moral Panics, and the Epstein Case
(Starts ~65:08)
- Michael Tracey’s Argument:
- Michael Tracey, a journalist, claims that concern about the Epstein case, and related scandals, may amount to a "moral panic." He argues against hysteria and media-driven panics, preferring skepticism towards widely accepted scandal narratives.
- Quote from Bret, referencing Tracey: “He is effectively portraying this as a moral panic while the question about whether or not it might be a moral panic is still completely open, and there's a whole lot of evidence that says it isn't.” (66:51)
- Epstein, Kompromat, and Governance:
- The hosts argue that the real question is not about panic but about whether systematic child exploitation and blackmail (kompromat) have compromised the governance of democratic institutions.
- Bret: “Do we live in a republic that is based on the idea of consent of the governed, or has somebody hijacked it using Kompromat? We need to know the answer to that question.” (71:02)
- Type I vs. Type II Errors in Public Outcry:
- Heather notes the danger in committing to a single type of epistemic error (either ignoring potential crises or endlessly panicking), arguing for flexibility and a willingness to change one’s stance as evidence emerges.
- Heather: “At some level it comes down to actually you can't have a static line in the sand and say, I will never acknowledge that something is possible if a bunch of people are suddenly thinking that it is... you have to not have a static position or else you're going to be extremely wrong some of the times.” (73:11)
- Owning the Downside:
- The hosts stress that if one claims it’s all a “moral panic" and is later proven wrong, the consequences must be accepted.
3. The State of Scientific Fraud
(Starts ~78:42)
a. Review of the PNAS Paper on Systematic Fraud
- Main Findings:
- The paper (“The entities enabling scientific fraud at scale are Large, Resilient and growing rapidly”) argues that scientific fraud is not rare, nor solely the product of ‘bad apples,’ but is systemic, organized, and escalating faster than legitimate scientific output.
- Heather (reading the abstract): “Our final analysis suggests this ability to evade interventions is enabling the number of fraudulent publications to grow at a rate far outpacing that of legitimate science.” (80:47)
- Mechanisms of Fraud:
- Paper mills (organizations literally selling fake or low-quality research, often across many countries)
- Brokers (entities connecting authors and predatory journals)
- Editors and conferences complicit in fraud or failing to uphold standards
- A reliance on the good faith assumption in a system where the “honor system” no longer suffices
b. Insights from Bret and Heather
- Systemic Incentives:
- Fraud is game-theoretically favored because it helps scientists advance. “The sky fell and nobody blinked," as Bret says.
- Bret: “How do you get a system in which you've got tens of thousands of people engaged in various different levels of fraud in every single discipline? You get it by creating a system of incentives where you get a positive feedback, where the people who are succeeding are the ones who are doing these things.” (105:17)
- Transparency Paradox:
- Journals like PLOS ONE, which disclose editor names, are more easily measured for fraud—ironically making them look worse by transparency, even though the real problem is likely worse in less transparent venues.
- Impact metrics and CVs can be gamed by publication mills and predatory, off-topic journals.
- Peer Review and Public Critique:
- The old model of peer review has failed; only post-publication review (like PubPeer) offers real error correction now.
- Bret: “Peer review is not review by peers. Review by peers is desirable...but we are always told about peer review and how it's the guide to quality science. Why didn't peer review catch all of this fraudulent work?” (100:49)
- Cumulative Effect Problem:
- Once fraud enters the literature, it contaminates all subsequent research, making it increasingly difficult to build reliable knowledge.
- Heather: “Unless your question is small enough that you can actually answer it empirically within your own lab or your own scientific ecosystem...it is becoming nearly impossible.” (110:24)
- Cherry-Picking as Rational Survival:
- With so much fraud, selectively trusting only high-confidence sources becomes necessary—even as “cherry-picking” is normally derided.
- Bret: "Cherry picking for only those papers you believe have a high likelihood of actually reflecting work that took place is the only way to get around rampant fraud." (113:55)
c. Implications and Outlook
- AI Magnifies the Problem:
- If AI models are trained on the literature, they will amplify, not correct, these errors:
Bret: “What do you suppose is going to happen when AI reads the literature?...It's going to be as dumb as your average PhD, which is not a good look. That's what's going to happen." (123:08)
- If AI models are trained on the literature, they will amplify, not correct, these errors:
- Restoring Scientific Integrity:
- True science now depends on “the keepers of the flame” who maintain high standards and integrity outside of corrupted institutions.
Notable Quotes and Memorable Moments
- “We are the kind of creature who, more than any other creature is in danger of taking some beautifully architected evolutionary thing and squandering it because we solve some problem that is our focus at this moment in time. And that is what's coming.” – Bret (55:52)
- “The honorable people aren't going to be able to compete. So if you want to find them, they're not going to be there.” – Bret (106:38)
- “If you have defected from the goal of discovering what is true in favor of gilding your CV, then you don't deserve to be in that field.” – Heather (108:05)
- “You need to find the keepers of the flame. Many of them will not be academics. Some of them will never have been academics. Some of them will have been exiled from academia. And they rekindle the fire of science.” – Bret (125:52)
Timestamps for Key Segments
- AI Fashion Show & Gender Roles: 14:04 – 43:49
- Specialization, Loss of Skills, and Family Erosion: 25:34 – 54:35
- Michael Tracey, Moral Panics, & Epstein: 65:08 – 78:42
- PNAS Paper and Scientific Fraud: 78:42 – end (approx. 128:00)
Tone & Style
The tone is thoughtful, sometimes sardonic, and deeply critical of complacency, approaching the subject matter from a blend of evolutionary science, cultural observation, and personal anecdote. Both speakers balance grave warnings (“the sky fell and nobody blinked”) with moments of self-deprecation and humor.
Summary Takeaways
- Both human relationships and the pursuit of truth are endangered by novel technological forces and perverse incentives.
- The loss of intergenerational knowledge and the outsourcing of human connection to technology—especially AI—harbor unseen dangers for individual and species-level wellbeing.
- Scientific fraud is now systemic, and incumbent institutions cannot be trusted to self-correct.
- True sensemaking, going forward, depends on individuals and small communities maintaining and teaching uncorrupted methods and standards.
For regular listeners, this is classic DarkHorse: high-level synthesis, challenging current orthodoxy, and offering a bracing (if sometimes bleak) analysis of civilization’s trajectory as seen through the evolutionary lens.
