Interview, Fireside Chat, Panel, Conference Presentation
E147: TED goes woke, Canada's Nazi blunder, AI adds vision, plus: who owns OpenAI?
Coleman Hughes on TED's Ideological Capture
- Hughes accepted an invitation from TED founder Chris Anderson to speak about his upcoming book, The End of Race Politics, which argues for colorblindness in policy and personal life, prioritizing class over race as a variable for assistance.
- Although 95% of the 2,000-person audience responded positively, Hughes reports that a small group of staff and attendees, including the "Black at TED" employee group, internally opposed the talk.
- Following the talk, Hughes was informed that the video would not be released as intended; instead, staff demanded a debate video be created to accompany the original talk to "appease woke staffers."
- Hughes refused to have the talk and debate merged into a single video, fearing it would imply his argument required an opposing view to be valid, eventually agreeing to a release where the talk and debate are posted separately two weeks apart.
- Hughes discovered that the view count of his TED talk on the TED website was 73,000, significantly lower than the 450,000–800,000 views of five other talks released in the same period, leading him to believe the organization intentionally suppressed the content.
- The suppression was only made public after blogger Tim Urban tweeted about the low view count, prompting Hughes to release the video externally.
- Hughes characterizes TED's reaction as granting a "heckler's veto" or "crybaby's veto," where subjective emotional discomfort is prioritized over the free exchange of well-reasoned ideas.
- Hughes argues that TED has shifted from its original mission of "ideas worth spreading" to a one-sided platform that excludes perspectives like the nuances of why voters supported Donald Trump or critiques of identity politics.
- Chris Anderson admitted in a public post that some people were "genuinely hurt" by Hughes' talk, a sentiment Hughes critiques as conflating disagreement with physical harm or safety threats.
- Hughes notes that while the stereotype suggests people of color oppose colorblindness, many immigrants and minorities resonated with his message because they find racial identity less central to personal interactions than progressive institutions suggest.
- Hughes observes that institutional capture in organizations like TED is a "bottom-up" process where staff ideology overrides leadership intent, making it difficult for leaders like Anderson to enforce free speech without firing a significant portion of their workforce.
- When asked for advice to CEOs past the point of ideological capture, Hughes stated he has "no good advice," describing the situation as intractable due to the sheer scale of staff values that contradict leadership.
Institutional Capture and Policy Implications
- Hughes argues that welfare reforms since the late 1960s created structural incentives that contributed to social issues, suggesting current solutions often treat symptoms rather than root causes.
- Hughes opposes race-based hiring in venture capital and startups, asserting that hiring decisions should ignore race and gender, as signaling that identity is primary creates vulnerabilities to superficial appeals.
- Hughes warns that treating race as a "deep feature" of identity from the start of employment can lead to future meltdowns where companies feel obligated to address race-based grievances.
- Hughes suggests that the Democratic Party suffers from a disconnect between the "Twitter commentariat" and the actual views of the broader electorate, many of whom find colorblind arguments uncontroversial.
- Barry Weiss, a guest on the show, advocates for founding new institutions with the correct ethos from the start rather than attempting to reform captured ones like TED.
- Hughes references Brian Armstrong's strategy at Coinbase, where he established a clear mission and asked employees who did not align to leave, as a potential model for leadership to reset an institution's values.
Canadian Parliament and Ukrainian Nazi Controversy
- The Canadian Parliament gave a standing ovation to 98-year-old Yaroslav Hunka, a member of the Waffen-SS Galicia Division, after introducing him as a Ukrainian war hero at the invitation of House Speaker Anthony Rota.
- Speaker Rota resigned following the incident, and Prime Minister Justin Trudeau blamed Rota solely, claiming he did not know Hunka's background, though the episode reveals a lack of due diligence regarding WWII history.
- Chamath Patihapthia argues the incident reflects a broader pattern of "woke" virtue signaling and incompetence in Canadian leadership, noting Trudeau's nepotism and lack of political experience compared to his father.
- Chamath asserts that the present political narrative (support for Ukraine) is being used to whitewash history, ignoring the fact that Ukrainian nationalism has deep roots in neo-Nazi ideology.
- Chamath details how the 2014 Maidan coup was backed by Western officials and involved figures from the Svoboda party, whose original logo was a Nazi-inspired "Wolf's Angle."
- Chamath points out that the Azov Battalion, a neo-Nazi group, was integrated into the Ukrainian military and received support from Western governments despite its ideological composition.
- Chamath argues that the U.S. State Department and CIA knowingly aligned with these groups during the 2014 coup for political expediency, creating a problematic moral precedent for current foreign policy.
- Chamath contends that while President Zelensky is Jewish and likely not a Nazi himself, the Ukrainian government's willingness to ignore the Nazi elements within its nationalist movement is a deliberate political choice.
- Jason Friedberg adds that Canada's history includes bringing in former SS veterans in the 1940s and 50s for their anti-communist stance, creating a legacy where the presence of Nazi collaborators was normalized to serve anti-Soviet goals.
OpenAI, AI Hardware, and Future Computing
- OpenAI is in advanced talks to raise over $1 billion from SoftBank's Masayoshi Son and Apple designer Jonathan Ive, potentially to develop a consumer "iPhone of AI."
- OpenAI is simultaneously discussing a secondary share sale that could value the company between $80 billion and $90 billion, roughly three times its previous valuation.
- Despite reports that Sam Altman owns zero equity in OpenAI, Chamath speculates Altman controls the company through a non-profit foundation and utilizes a "capped return" investment model to retain residual value while avoiding tax and regulatory scrutiny.
- Chamath argues that the capped return model (e.g., capping investor returns at 100x or $300 million) functions like a sold call option, allowing the founders to retain the massive upside potential if the company reaches a trillion-dollar valuation.
- OpenAI recently announced multimodal capabilities, allowing users to input images, voice, and code to receive text or visual outputs, such as interpreting a "no parking" sign to determine parking availability.
- Google is integrating AI into core services like Flights and Docs, enabling complex natural language queries across multiple data points (e.g., comparing flight options with layover and seat preferences).
- Friedberg predicts that Large Language Models (LLMs) will evolve from chatbots into a new "kernel process" or operating system that runs apps in-stream, replacing the traditional app store model with dynamic, voice-driven services.
- Friedberg cites a paper demonstrating that LLMs can improve autonomous driving by acting as a communication layer between sensor data and control systems, allowing for human-like interpretation of complex environments.
- The future computing interface is predicted to be multimodal, combining voice, visual input from cameras/AR glasses, and audio output, rendering the concept of distinct "apps" obsolete in favor of in-stream services.
- Users are increasingly adopting external memory tools like "Reflect," which logs meetings and creates backlinks between concepts, serving as a bridge toward fully automated AI personal assistants.