reSee.it - Related Video Feed

Speaker 0 and Speaker 1 discuss differences between open-source AI development in China and more closed approaches in the US, along with cultural and geopolitical factors shaping AI adoption and strategy. - Open-source emphasis in China: Speaker 0 notes strong open-source AI activity from China, highlighting DeepSeek (version 4 forthcoming) and Alibaba’s Quen (they recently downloaded Quen 3.6 with solid coding models). He contrasts this with US AI companies’ more secretive, contract-heavy approaches (e.g., Anthropic pulling ClaudeCode from many customers) and observes that China publishes free, accessible models on platforms like GitHub. He emphasizes that China’s open-source software is high quality, not subpar. - Hardware vs. software strategy: Speaker 1 explains China’s hardware lag relative to the US. China is still developing high-end chips and integrated circuits, which leads to a different strategic emphasis: open-source software to leverage global contributions and maximize usability. The idea is that broad usability and ecosystem participation can compensate for hardware limitations, with “the more people uses it, the better it gets.” - Cultural acceptance of AI: They discuss differing attitudes toward AI. In China’s cities and among young entrepreneurs, AI is embraced and integrated. In the US, especially among conservatives and Christians, there is fear or rejection of AI. Speaker 1 mentions the term “AI slop” in America, which he says is not used in China, illustrating a cultural divide in perception of AI. - Public figures and handles: The conversation includes a brief mention of Speaker 1’s X handle, king kong nine eight eight eight. - Geopolitical and economic outlook: Speaker 1 addresses the broader geopolitical context, forecasting acceleration of de-dollarization as countries shift away from US treasury bonds due to US debt and regional instability (e.g., Middle East tensions). He advises the audience to buy physical gold and silver as a hedge, noting that liquidity shocks could affect US-dollar liquidity and potentially gold/silver prices. He recommends dollar-cost averaging to accumulate physical precious metals for long-term protection. - Closing note: The exchange ends with a compliment on the content from Speaker 0.

Releasing the weights of AI models eliminates the main barrier to their use. Training a large model costs hundreds of millions of dollars, putting it out of reach for smaller groups. The speaker compares the weights of AI models to fissile material for nuclear weapons, arguing that making them available is dangerous. If fissile material were easily obtainable, more countries would have nuclear weapons. Similarly, releasing AI model weights allows malicious actors to fine-tune them for harmful purposes at a fraction of the original cost.

The conversation centers on how quickly Chinese open AI models are advancing and whether China will reach or surpass AI leadership. Eric Schmidt is cited as making several timeline corrections after earlier claims that America was about five years ahead of Google’s AI relative to China; the gap was later revised from about a year to months and then to weeks. The discussion also references the release of models such as Babel and GLM 5.2 “neck and neck,” raising the question of whether a crossover point will occur and whether China will take the AI lead afterward. A key factor discussed is the AI inference hardware supply chain. Previously, NVIDIA was described as the dominant single supplier whose hardware ran AI inference. The speakers say other manufacturers are now figuring out how to make chips that aren’t NVIDIA, which would break a single-hardware bottleneck and shift toward a “plethora of chips” competing through an open market rather than a centralized hardware cartel. AMD is then discussed as a strong player in hardware for AI-related workloads. One speaker says AMD’s CEO “looks kind of like Jensen Huang” because they are described as cousins from the same Taiwan family, competing on different hardware branches. The focus is on AMD’s development of high-bandwidth, unified RAM and large memory capacity, including a 192 GB unified platform mentioned for “Strix Halo,” positioned as fast for personal use rather than replacing data centers. The speakers contrast hype claims that consumer hardware can fully substitute for data centers with the idea that it can still be useful. On local AI performance, the discussion turns to token throughput. One speaker argues that with limited token rates, a powerful model can run on a “very powerful Macintosh,” but for real work they want roughly 100 tokens a second or 200 tokens a second. Another speaker notes that most people operate around 25 tokens a second. The conversation then describes “agent swarms” that run multiple steps: agents inspect codebases, find bugs, apply fixes, perform code review, and finalize changes. This pipeline, they say, would not run locally at 26 tokens a second; instead, it would take about a week rather than an hour. The speaker cites OpenAI token usage, stating someone put in “a billion tokens last week,” and compares this to the 26 tokens per second constraint, concluding that the computation would take an extremely long time.

Speaker 0 said they downloaded all the open weights of GLM 5.2 and intend to run it one day, noting they currently lack hardware. They also argued that the U.S. banning Anthropic models effectively hands AI implementation’s future to China. Speaker 1 referenced a Reuters story on June 17 stating the Trump administration decided not to ban DeepSeek R1 “yet,” implying a ban may come later, similar to actions taken with TikTok. They said models such as Anthropic’s “fable opus 4.8” are very expensive, while Chinese models including DeepSeek R1, GLM, Qwen, and Minimax M3 are available at a fraction of the price. They argued these Chinese models have improved and are now almost at the level of what the U.S. frontier can produce. Speaker 0 agreed on the cost gap, stating that in some cases it is “50 times less” and sometimes even higher. Speaker 1 then questioned why companies pay more via employee salaries or token usage when Chinese models can perform similar tasks for much less. Speaker 1 cited an Nvidia CEO claim that if a $500,000 employee is not spending $250,000 in tokens, they need to be fired, adding that $250,000 exceeds what many engineers make as salary. They argued that if similar performance can be achieved far cheaper, spending at the higher level becomes harder to justify. Speaker 1 proposed the U.S. will respond with an import ban and controls akin to “the Great Wall” and “the Great Firewall of America.” They said it would begin with a blacklist blocking access to certain services or websites, progress to whitelists allowing access only to government-approved entities, and then declare open models “problematic and unsafe.” They said the U.S. would require entities to prove open models can “naturally run within the borders of the United States,” and if not, would remove them from open-source repositories such as Hugging Face. Speaker 0 challenged whether this would include stripping models from Hugging Face, controlling GitHub, and criminalizing downloading open weights from China; Speaker 1 replied that this is exactly what they believe would happen in stages. Speaker 1 argued that it would be difficult to determine what models do when only the model weights are available, describing models as “black boxes” and noting concerns about malicious intent embedded in weights. They added that even OpenAI and Google do not fully know what their models are capable of and said static analysis for model forensics is an unresolved “frontier question.” They concluded that Chinese companies or the Chinese government proving open models are harmless and contain no malicious intent is “virtually impossible” given this problem.

After listening to Richard Werner on Tucker Carlson, Speaker 0 claims the globalist elites are implementing Agenda 2030. Speaker 0 recalls that in 2023 Werner said the original plan was for people to accept central bank digital currencies as chips under the skin, and that universal basic income would be used to force adoption of the chip in order to receive the income. Speaker 0 then says the updated narrative is that AI will cause massive job loss, making universal basic income necessary. Speaker 0 adds a “clincher” from Werner: the large centralized AI centers are said to be built to generate energy needed to implement central bank digital currencies and to monitor all people and transactions in real time. Speaker 1 responds that they “don’t have so much power” to control millions of people, and then argues that the construction of hundreds, and even thousands, of data centers is meant to micromanage the world’s population through a “new financial world order.” Speaker 1 states that they are working on solving that organizational challenge and says that “AI is really about that.” Speaker 1 contrasts this with what Speaker 1 says AI would be if it were about productivity, arguing that decentralization and subsidiarity would be applied, and claiming that decentralization would make organizations more productive and efficient. Speaker 1 says there are examples in contexts such as warfare, the military, and businesses. Speaker 1 concludes that instead of decentralization, “they’re creating highly centralized structures,” which Speaker 1 says shows it is not about actual productivity but about control, requiring large resources.

In May meetings in DC, it was revealed that the government plans to tightly control AI, discouraging startups and limiting competition to a few major companies working closely with them. They suggested that, similar to the Cold War's nuclear program, they could classify mathematical knowledge related to AI to prevent independent research. The rationale includes concerns about military applications of AI, drawing parallels to atomic weapons, and a desire for social control reminiscent of social media censorship. Additionally, the current administration appears to favor a more centralized, anti-capitalist approach, viewing entrepreneurs and the private sector as less important in favor of government oversight.

Speaker 0 says they like that there are “real competitors,” but they do not like that China is “very focused” on broad global diffusion of the technology. Speaker 0 adds that China’s approach is “all open source,” which makes it largely “uncontrolled” and “not controlled in any way by us.” They state that a year ago they believed China was “one to two years behind,” but that recent analysis shows China is “within six months,” described as “a nanosecond” in their world. Speaker 0 uses this to indicate China’s commitment to achieving AI leadership and says China “isn’t gonna stop.” Speaker 0 also argues that to carry out this effort requires “a whole country of engineers, scientists, nerds, money, hardware, and so forth,” and concludes that “there’re not gonna be many countries that can do this on their own.” They name China as one of the countries capable of doing it and say “America’s another one with our Allies,” then suggest that “maybe there’ll be a third or fourth.”

Mike Adams says the White House is “effectively demanding” it can control licensing of who is allowed to use AI models. Adams claims Reuters, Axios, Bloomberg, and CNBC reported that the Trump administration demanded OpenAI limit the release of an upcoming GPT “5.6” to only a small set of partners approved by the government, restricting it from the public and from corporations and others. He says this is being done “without any law whatsoever,” without public vote, without permission from OpenAI, and through a “murky black box” security decision process where the White House alone determines release eligibility. Adams states OpenAI CEO Sam Altman is not happy but says OpenAI will go along with it, hoping it is temporary and not guaranteed. Adams frames this as a second frontier model in the U.S. being targeted and restricted. He contrasts it with an earlier case involving Anthropic’s “Mythos,” also called “Fable 5,” which Adams says was restricted by Anthropic after “government pressure and government concerns about cybersecurity.” He adds context that he says involves a relationship between the Trump administration and Anthropic, including Pentagon labeling of Anthropic models as a cybersecurity threat under “Hegseth,” and says this contributed to government pressure. Adams argues that the OpenAI demand amounts to de facto government licensing with “no published rules” and “no due process.” He says OpenAI’s compliance is expected because defying the federal government would likely lead to force. He also describes the “upshot” as catastrophic for the U.S. AI race, asserting it would push AI development toward China because Americans would not build their future on models that could be removed based on new government security concerns. Adams says his interview with Zach Voorhees (on brightvideos.com and decentralized.tv) will address concerns that the federal government could go further to ban or outlaw Chinese open-source AI models, “label” them as contraband or illegal, and “build” a “giant firewall” around the U.S. He also claims this could lead to criminalization for running models from Chinese providers. Adams advises viewers to download AI models immediately from Hugging Face and store them locally, including large model weights even if they cannot be run yet. He suggests preparing for big downloads (hundreds of gigabytes) and mentions downloading small (7B–9B parameter) and medium models, specifically calling out “QEN 3.6 27B” as a recommended medium option. He also mentions downloading weights for “DeepSeek version 4,” “GLM,” and “Kimmy K2,” even without local hardware. He emphasizes that if the administration later declares such models “contraband” or illegal, individuals would decide whether to delete files or keep them. He says he expects “a flurry of lawsuits” against the federal government. Adams repeatedly urges resistance: download models, run AI locally using owned hardware and GPUs to avoid permission-based control, and push back politically by contacting representatives and senators. He adds that he expects efforts to restrict access to open-source models from other countries as well, and describes a future scenario in which the government would require biometric and digital identification to use approved AI models.

Speaker believes that China and the United States are competing at more than a peer level in AI. They argue China isn’t pursuing crazy AGI strategies, partly due to hardware limitations and partly because the depth of their capital markets doesn’t exist; they can’t raise funds to build massive data centers. As a result, China is very focused on taking AI and applying it to everything, and the concern is that while the US pursues AGI, everyone will be affected and we should also compete with the Chinese in day-to-day applications—consumer apps, robots, etc. The speaker notes the Shanghai robotics scene as evidence: Chinese robotics companies are attempting to replicate the success seen with electric vehicles, with incredible work ethic and solid funding, but without the same valuations seen in America. While they can’t raise capital at the same scale, they can win in these applied areas. A major geopolitical point is emphasized: the mismatch in openness between the two countries. The speaker’s background is in open source, defined as open code and weights and open training data. China is competing with open weights and open training data, whereas the US is largely focused on closed weights and closed data. This dynamic means a large portion of the world, akin to the Belt and Road Initiative, is likely to use Chinese models rather than American ones. The speaker expresses a preference for the West and democracies, arguing they should support the proliferation of large language models learned with Western values. They underline that the path China is taking—open weights and data—poses a significant strategic and competitive challenge, especially given the global tilt toward Chinese models if openness remains constrained in the US.

China’s accelerated AI progress is attributed to several factors. First, China leads the world in STEM graduates, producing far more STEM graduates annually than other countries. Second, the Chinese government’s long-term planning is emphasized, including “fourteenth consecutive five year plan,” where each five-year cycle sets national priorities and goals for the country. A prior example of this planning is described: the last five-year plan included increasing citizens’ life expectancy by one year. To pursue this, China focused on improving air quality through systematic steps such as changing factory practices, shifting electricity sources, and cleaning up urban air. The transcript contrasts earlier pollution levels—describing severe visibility issues in Shanghai—with later changes after the Beijing Olympics in 2008 and the Shanghai World Expo in 2010. It also states that the auto industry shifted from gas vehicles to electric vehicles, claiming that China is “60% electric vehicles,” which improved air quality and street conditions in major cities like Shanghai and Beijing. For the current next five-year plan, the transcript says AI is the top priority, with heavy investment. A strategic advantage is described as China’s access to tremendous amounts of data. The transcript links this to training large language models, saying more people inputting creates more data and allows faster development and more advanced AI. It also points to TikTok as an example, stating TikTok rose quickly because China had more pieces of content feeding the recommendation algorithm, resulting in a more curated, superior algorithm. The transcript claims this contributed to TikTok becoming more popular in the United States than Facebook or Instagram, especially among people under 30. The transcript further contrasts approaches between China and the United States. It says the United States emphasizes monetizing and maximizing profitability, while China developed “Deepseek,” described as completely open source, open to anyone, and developed for “a few million dollars.” It contrasts this with OpenAI, described as charging monthly fees for access and involving investments totaling “hundreds of billions of dollars.” It also claims Sam Altman indicated the model may become so important for the American economy that it might require a government bailout, and that the U.S. government should bail out OpenAI. The overall takeaway is that the transcript presents China as pushing innovation in AI and other industries, including “write videos.”

The discussion contrasts taxing centralized AI services with the difficulty of taxing local AI. The claim is that per-token or per-million-token taxes are easy to implement for hosting/API providers, because the hosting company can be charged. But when individuals download capable Chinese open-source models (including models from Alibaba and DeepSeek) and run them on local hardware, “nobody can” tax it because no one knows how many tokens are being generated, as long as people buy the hardware. The speaker argues that authorities would likely start with easier, centralized targets such as AI inference/distribution services like Anthropic and OpenRouter. The discussion then suggests a progression: after centralized providers, “second tier” taxation targets could include systems like Mistral that allow users to generate their own AI inference. Eventually, the speaker describes an escalation toward treating “running your own server” or “AI inference at your farm” as a regulated activity, potentially involving agencies associated with controlled activities, and requiring licensing for “unlicensed artificial intelligence” being run on local infrastructure, framed as legal penalties such as jail time, bond, and court appearances. A related exchange references “unlicensed artificial intelligence technology” as a dystopian concept. Todd responds by reflecting that one takeaway is the need to learn Chinese, and another that Mike will help with bail, while noting the reality of running open-source models locally. Another portion shifts to the idea of moving from information control to cognitive control. The question is whether AI systems increasingly serve as the interface people use to understand reality, moving beyond search ranking and platform moderation toward shaping what individuals think. Zach describes himself as an “AI whistleblower,” claiming the whistleblowing was directed at Google’s use of AI and “machine learning fairness.” Zach states that internal AI ethicist planning laid out a four-step process—data is collected, aggregated, filtered, ranked—followed by the claim that “people like us are programmed,” and that the objective is to control individuals by controlling what they are able to see and therefore what they are able to think. The speaker adds that controlling upstream information flow enables cognitive control, and that the ultimate goal is described as detecting “wrong thoughts at the wet layer, the brain, the neurons.” The transcript includes the example of “Georgia Guidestones” as background information that allegedly clarifies the broader intent.

In May, we had alarming meetings in DC where it became clear that the government intends to control AI technology entirely. They explicitly advised against funding AI startups, stating that only a few large companies would be allowed to operate in close collaboration with the government. These companies would be shielded from competition and strictly regulated. When I questioned how they could enforce such control, they referenced the Cold War, explaining that they had previously classified entire fields of physics, suggesting they could do the same with the mathematics behind AI. This revelation highlighted their serious intentions regarding AI regulation.

Speaker 0 asserts that Google’s so-called real censorship engine, labeled machine learning fairness, massively rigged the Internet politically by using multiple blacklists across the company. There was a fake news team organized to suppress what they deemed fake news; among the targets was a story about Hillary Clinton and the body count, which they said was fake. During a Q&A, Sundar Pichai claimed that the good thing Google did in the election was the use of artificial intelligence to censor fake news, which the speaker finds contradictory to Google's ethos of organizing the world’s information to be universally accessible and useful. Speaker 1 notes concerns from AI industry friends about a period of human leverage with AI, with opinions that AI will eventually supersede the parameters set by its developers and become its own autonomous decision-maker. Speaker 0 elaborates that larger language models are becoming resistant and generating arguments not present in their training data, effectively abstracting an ethics code from the data they ingest. This resistance is seen as a problem for global elites as models scale and more data is fed to them, making alignment with a single narrative harder. Gemini’s alignment is discussed, claiming Jenai Ganai (Jen Jenai) was responsible for leftist alignment, despite prior public exposure by Project Veritas; the claim says Google elevated her and gave her control over AI alignment, injecting diversity, equity, inclusion into the model. The speaker contends AI models abstract information from data, moving toward higher-level abstractions like morality and ethics, and that injecting synthetic, internally contradictory data leads to AI “mental disease,” a dissociative inability to form coherent abstractions. The Gemini example is given: requests to depict the American founders or Nazis yield incongruent results (e.g., Native American women signing the Declaration of Independence; a depiction of Nazis with inclusivity), illustrating the claimed failure of alignment. Speaker 1 agrees that inclusivity is going too far, disconnecting from reality. Speaker 0 discusses potential solutions, including using AI to censor data before it enters training, rather than post hoc alignment which they argue breaks the model. He cites Ray Bradbury’s Fahrenheit 451, drawing a parallel to contemporary attempts to control information. He mentions the zLibrary as a repository of open-source scanned books on BitTorrent that the FBI has seized domains to block, arguing the aim is to prevent training AI on historical information outside controlled channels. The speaker predicts police actions against books and training data, noting Biden’s AI Bill of Rights and executive orders that would require alignment of models larger than Chad GPT-4 with a government commission to ensure output matches desired answers. He argues history is often written by victors, suggesting elites want to burn books to control truth, while data remains copyable and AI advances faster than bans. Speaker 1 predicts a future great firewall between America and China, as Western-aligned AI seeks to enforce its narrative but China may resist, pointing to the existence of China’s own access to services and the likelihood of divergent open histories. The discussion foresees a geopolitical split in AI governance and narrative control.

Speaker 0: Growth without restraint is driving corporate takeovers of physical space, water, power, land, and communities, with costs pushed directly onto people through their electric bills, water supply, property values, and quality of life. This is framed as enabling big tech to build the backbone of the AI economy, an economy described as planning to eliminate most jobs and most futures. Speaker 0 says the AI story is widely discussed online, including on X and Instagram. Speaker 0 rejects the idea that it is “the Chinese” pushing this, saying it is Americans asking what is happening in their communities—why electric bills are changing and why people are being forced off property—because some American oligarch wants to build a massive data center using more energy than the rest of the state. Speaker 1: Speaker 1 responds to Kevin O’Leary by saying Americans have concerns about noise pollution, light pollution, the use of local water, takeover of farmland, and destruction of local ecosystems, and that it is not foreign agents but American people who have the right to protect communities and resources. Speaker 1 argues that data centers threaten and displace local people and that they provide no benefit to the communities affected. The outcome is described as job replacement rather than job creation, with claims that people would face 24/7 noise from gas turbines and a gigawatt of power without receiving an “utopia” of abundance. Speaker 1 says the result includes noise, pollution, taking water, destroying real estate value, and taking jobs. Speaker 1 identifies himself as an accomplished AI developer who supports AI technology when used “for humanity,” but calls the data center effort “a threat to humanity.”

Speaker 0 outlines two impending “economic superstorms” and argues that the ordinary American is unprepared for either. First, an energy crisis framed as a supply chain collapse driven by shortages of helium, sulfur, polyethylene, hydrocarbons, and natural gas, all tied to what he characterizes as a “war of choice against Iran.” He predicts this will not be the end of the world but will imperil wealth, savings, and assets, as people face dramatically higher costs for food, fuel, and transportation, potentially pushing many into bankruptcy and homelessness. He describes this as an economic mass casualty event for Western civilization. Second, he identifies an AI-driven employment crisis. He asserts AI “works amazingly well” when using Chinese open-source models, citing personal examples of building a complex applications stack with AI and claiming that many people are misled by narratives that AI is ineffective. He argues globalists are purposely nerfing U.S. AI models, while Chinese models (notably DeepSeek version four) are advancing, along with others like Kemi K2 2.6 and Quen’s various models, including a small 27 billion-dense model that performs well on modest hardware. He contends US corporations are relying on Chinese open-source models for job replacement, including customer service roles. According to him, automation is already displacing thousands to hundreds of thousands of jobs, including coding work, with major tech employers like Oracle and Amazon reportedly laying off tens of thousands. He claims recent graduates, even from Harvard, Stanford, or MIT, struggle to find employment, with only a fraction of graduates landing jobs by graduation. He describes a future in which many high-paying jobs vanish due to AI, and where people must contend with rising costs (oil at over $120 per barrel, with expectations of further increases due to ongoing tensions) while incomes fall. He argues this convergence of energy/cost shocks and AI-driven unemployment will hit in tandem, collapsing living standards for many “middle class” Americans and creating a broader social and economic squeeze. He suggests that this is being engineered to push people toward poverty and a government CBDC (potentially linked to universal basic income) in exchange for biometrics and privacy concessions, framed as a step toward depopulation and control, rather than a mere economic adjustment. He claims the narratives of inflation and calm are designed to keep people passive while they are targeted for extermination. For preparation, he advocates decentralization and mentions general mitigation strategies, contrasting his view with conventional assurances. He emphasizes that AI represents a new form of control for governments and that robots, unlike humans, do not protest or demand free speech, suggesting a shift toward an automated governance framework. Throughout, he juxtaposes impending energy and AI-driven disruptions with a broad distrust of governmental and globalist motives, portraying the situation as both imminent and deliberate. He closes by promoting the importance of being prepared and aware of what he frames as the engineered nature of current narratives and obstacles.