OpenAI and Anthropic CEOs are both so annoying! Doomsday theories and relative deprivation make American people resent AI

Silicon Valley’s “All-In Podcast” points out that American society is developing a backlash against AI, with data centers becoming the target for anti-wealth sentiments and fears of unemployment.

When discussing the AI industry, the well-known Silicon Valley podcast “All-In Podcast” made a sharp observation: the attitude of American society toward AI is turning negative, and the most concrete outlet for this emotion is the data centers being rapidly built across the country by AI companies. The sources of this resentment may include AI doomsday theories, fears of job loss, or a deeper dissatisfaction: a new wave of technological innovation seems to only make a few people extremely wealthy, while the majority see no significant improvement in their lives.

Local governments in the U.S. have overturned data center construction cases

Chamath Palihapitiya stated on the show that the current problems facing the AI industry are not just model competition, capital expenditure, or computing power shortages, but that “the American people are increasingly resentful of AI overall.” He pointed out that this resentment may stem from AI doomsday theories, fears of unemployment, or a deeper dissatisfaction: a new wave of technological innovation seems to only make a few people super-rich, even “creating a batch of trillion-dollar billionaires,” while most people’s lives do not see obvious improvements.

Chamath believes that when this sentiment accumulates to a certain point, local communities are most likely to oppose data centers. For example, in the U.S., a local government initially approved a $6 billion data center project, but the supporting committee members were subsequently replaced in the election, and the new officials attempted to overturn the original decision. He sees this as evidence that data centers are no longer just infrastructure but have become political symbols of the AI industry and tech billionaires.

Another host, David Friedberg, offers a more direct perspective. He believes many Americans are actually “starting to really dislike the wealthy,” and data centers are a physical manifestation of this emotion. He describes data centers as one of the most obvious physical spaces where wealth is created in the U.S., and as machines that highlight the widening gap between tech elites, political connections, and billionaires in the eyes of ordinary people.

Friedberg states that, for most people, the benefits of AI are not concrete enough. Many hear daily that AI will change the world, reshape businesses, and boost productivity, but in their own lives, the real improvements they feel might be only using ChatGPT for medical advice, writing letters, or searching for information. In contrast, what they more directly feel is anxiety over job displacement, concerns about rising electricity prices, and the large data centers built by tech companies to train models.

Therefore, Friedberg compares data centers to “the tax targets of this era’s mansions.” If politicians previously attacked second homes, mansions, or private jets of the wealthy, in the AI era, data centers are the new target. They represent the progress of tech billionaires but also symbolize the progress that others do not feel.

David Sacks adds a policy and industry perspective, explaining why data centers are becoming unpopular in many U.S. states. First, many local communities worry about the massive electricity consumption of data centers, which could raise electricity bills for ordinary households. Sacks notes that some developers previously sought permits from local governments even without clear solutions for power supply, which sparked backlash from communities.

Second, he points out the combination of AI doomsday groups and anti-data center movements. Sacks believes that some groups claiming AI could bring catastrophic risks find it hard to persuade the public that “AI will lead to terminators,” but framing the issue around water and electricity consumption or community destruction makes it easier to mobilize local opposition. He criticizes that some anti-data center movements are essentially “packaged NIMBYism.”

David Sacks criticizes Anthropic’s AI doomsday theories

Sacks targets Anthropic. He believes that Anthropic has historically allied with AI doomsday groups and NIMBY organizations, perhaps because Anthropic did not plan to build large data centers itself but relied on hyperscalers for computing power. Opposing data center construction was thus a way to “throw sand” in the path of competitors like OpenAI and xAI.

However, as Anthropic grows and its computing needs surge, if it must enter the data center construction race itself, this strategy could backfire.

The show also mentions that one of the biggest current bottlenecks for AI companies is the shortage of computing power. Chamath points out that the market’s reaction to Allbirds’ transformation into an AI data center company, with its stock soaring, seems absurd but actually reflects that capital markets have realized “a severe shortage of computing power.” He states that the AI industry not only lacks GPUs but also land, electricity, data center shells, and local government permits.

This creates a paradox for AI companies: on one hand, companies like OpenAI, Anthropic, xAI, and Meta need more data centers to support model growth and revenue; on the other hand, societal resentment toward data centers is growing, and local governments and residents are increasingly likely to block these developments.

Chamath warns that if leading AI companies cannot secure enough computing resources, revenue growth may not slow because of product quality issues but due to a problem similar to Friendster’s: demand exists, but infrastructure cannot keep up, ultimately allowing competitors to overtake.

Sacks also believes that if data center construction faces too many restrictions in the U.S., computing power might shift elsewhere, such as to regions with cheaper energy, more favorable policies, or even allied countries. He points out that if the U.S. restricts domestic data centers while opposing allies from using U.S. technology to build AI infrastructure, it will only weaken America’s advantage in the AI race.

Silicon Valley investors: Altman and Amodei are unsuitable as industry spokespeople

But the most noteworthy part of the show is the assessment of the AI industry’s public relations crisis. Host Jason Calacanis bluntly states that one of the biggest problems now is that the industry’s spokespersons are too poor. Comparing the American society’s perception of AI with China’s highly positive attitude, he believes the U.S. AI industry’s messaging is almost entirely centered on fear, unemployment, and elite monopoly.

Jason further points out that the public image of the AI industry is also related to its representatives. He thinks that Anthropic CEO Dario Amodei’s long-term focus on disaster, security risks, and large-scale unemployment deepens public fears. Meanwhile, OpenAI CEO Sam Altman, often embroiled in controversy, is also unsuitable to be a convincing figure. Jason bluntly states that these two “cannot be the industry’s spokespersons.”

If the AI industry wants to improve its social image, it must be redefined by figures who can better explain public interests like healthcare, education, and housing.

He advocates that the AI industry should refocus its narrative on three areas that can genuinely improve people’s lives: healthcare, housing, and education. That is, AI companies should not only tell the market how many trillions of dollars they are worth or how much they can save in labor costs for enterprises, but also show ordinary people how AI can make healthcare cheaper, education more efficient, and housing issues easier to solve.

• This article is reprinted with permission from: “Chain News”
• Original title: “Sam Altman, Dario Amodei Are Too Annoying! AI Doomsday Theories and Relative Deprivation Fuel U.S. Public Resentment Toward AI”
• Original author: Neo