Sora Proves the AI Bubble Is Going to Burst So Hard

The speaker criticizes OpenAI's recent product, Sora 2, a video generation app, arguing it's a poorly conceived product with significant ethical and economic flaws. Sora 2 allows users to create deepfakes, leading to the humiliation of individuals, including its own CEO, and the defilement of historical figures' legacies. The app also facilitates the creation of copyrighted content and the spread of misinformation and racist material. Economically, Sora 2 is unsustainable, costing over $5 per video to generate and lacking a clear monetization strategy, contributing to a massive drain on OpenAI's resources. This, along with the broader AI industry's speculative investment and lack of tangible productivity gains, suggests the AI sector is a bubble poised to burst, with detrimental consequences for the general economy and public.

• Sora 2, an AI video generator, faces criticism for enabling deepfakes, copyright infringement, and the spread of harmful content.
• The app is economically unsustainable, costing OpenAI over $5 per generated video with no clear monetization plan.
• The broader AI industry is characterized by massive, speculative investment totaling close to $400 billion annually, with little evidence of productivity gains.
• Experts question the current AI development approach, suggesting that simply scaling models is not leading to transformative "super intelligence."
• The speaker argues the AI industry is a bubble, similar to past tech bubbles but larger, which, when it bursts, will negatively impact the broader economy and regular people.

OpenAI's Floundering and Sora 2 [00:00]

• OpenAI, despite being the most valuable private company, has faced recent setbacks, including technical difficulties with a secret hardware device and user disappointment with GPT-5.
• The release of Sora 2, described as a "Tik Tok knockoff with AI slop," is questioned for its purpose and impact, especially after the hype surrounding GPT-5.

And so after a long run of bad press, to prove the haters wrong, Sammy Alman released Sora 2, a Tik Tok knockoff chaka block with AI slop.

Sora 2's Ethical and Societal Concerns [01:07]

• Sora 2 allows users to create AI deepfakes of anyone, including enabling the use of the CEO's likeness for humiliation.
• Videos generated depict the CEO being physically, emotionally, and sexually humiliated, highlighting a lack of respect for individual likenesses.
• The app facilitates the creation of content that infringes on copyrighted works, with the company claiming surprise at the copyright drama.
• Sora 2 is used to create content defiling the memory of deceased individuals like Bob Ross and Martin Luther King Jr., without their families' consent.
• The widespread ability to create realistic videos with Sora 2 makes it difficult to discern real from fake, potentially leading to the spread of fake news and political propaganda.
• The app is used to generate racist content, stereotyping various ethnicities in disgusting ways.

So, Sora is not just making Nazi Spongebob. It's making a tool to defile the memory of the dead.

The Economic Unsustainability of Sora 2 [06:31]

• The economics of Sora 2 make "zero sense," with users generating videos that cost OpenAI significant amounts of money to produce.
• An estimate suggests it costs over $5 to generate each video, leading to substantial financial losses per user.
• OpenAI has no clear plan for monetizing Sora 2, with the CEO acknowledging unexpected high user generation rates and low audience engagement per video.
• The majority of generated videos are described as bad, not funny, uninteresting, or unwatchable.
• Even when a good video is produced, it's seen as a worse, more boring version of TikTok, which obtains its content for free.
• Sora 2 incurs significant costs for each video, with no monetization strategy, unlike even simpler apps that include ads.

My friend Ed Zitron estimates that it costs OpenAI over $5 to make every single video. That means that they are losing huge amounts of money on every single user.

The Broader AI Industry Bubble [08:53]

• The initial promise of AI to solve global problems has seemingly devolved into creating TikTok clones instead.
• The speaker suggests that AI leaders may lack genuine ideas or understanding of the technology's true purpose, indicating a potential bubble about to burst.
• The AI industry's rapid growth, fueled by massive investments following ChatGPT's success, is propping up the entire economy.
• Companies are spending close to $400 billion annually on AI infrastructure, with projections for continued increases.
• This AI investment eclipses spending on major infrastructure projects like the interstate highway system and is likened to an "Apollo program every 10 months."
• The AI investment has accounted for all GDP growth in the first half of the year and significantly drives the stock market.
• Tariffs on goods have been waived for AI gear to artificially boost economic appearance.
• The economy is described as one enormous bet on AI, with the hope of achieving "super intelligence."

Our entire economy has become one enormous bet on AI.

The Unrealistic Promise of Super Intelligence and AI's Economic Viability [12:32]

• The stated goal of AI development is to create "super intelligence" or AI smarter than humans.
• Concerns are raised that AI development might lead to increased social media addiction rather than empowerment.
• Predictions that AI will eliminate half of entry-level white-collar jobs in the next five years are questioned for their benefit.
• The immense cost of AI development is a significant hurdle, with AI companies needing to earn $2 trillion annually by 2030, a figure exceeding the combined revenue of several major tech companies.
• Current AI revenue projections fall far short of this target, indicating a potential shortfall of $800 billion by 2030.
• The AI sector is characterized as a bubble, where valuations are not supported by underlying value or future revenue.
• Studies suggest AI does not significantly increase worker productivity; in fact, it might decrease it.
• Many companies attempting to use LLMs have not turned a profit.
• The argument that AI models are rapidly improving is countered by the observation that recent releases like GPT-5 and Llama have not met expectations.
• Simply scaling LLMs does not guarantee transformative intelligence; they have specific uses and limitations.

The main impediment to building super intelligent AI might not be the technology. It might be the money.

Investor Irrationality and AI's Bubble Characteristics [18:34]

• Investors are described as having lost common sense, funding even the most nascent AI companies with billions of dollars based solely on the word "AI."
• Companies with no demonstrable product or clear answers about their AI initiatives are securing massive seed funding.
• The AI market is "frothy," with billionaires acquiring AI programmers like collectibles.
• Companies are resorting to "abracadabra accounting" to hide AI spending and inflate profits for investors.
• Unlike past long-term infrastructure investments like railroads, AI infrastructure has a short shelf life due to rapid technological advancements.
• The billions spent on AI processors could become obsolete within 5-10 years, leaving companies with billions in "junk."
• Leaders in the AI field, including Zuckerberg and Altman, acknowledge the possibility of an AI collapse or that investors are overly excited.
• Technological advances historically involve a torrent of cash, with many innovations accompanied by bubbles.
• The current AI bubble is considered the largest, exceeding the dot-com bubble and all previous tech bubbles combined.

The AI market is so frothy that the billionaires are now buying and selling AI programmers like their Pokemon cards.

The Inevitable Pop and Consequences of the AI Bubble [22:44]

• Tech investors are aware of the bubble and its potential to burst but believe they will not be the ones to lose money.
• The tech companies, banking systems, and key figures like Sam Altman are expected to be insulated from the bubble's collapse.
• Regular people, who own stocks through pensions and 401ks, will suffer the most significant losses when the bubble bursts.
• A stock market crash due to the AI bubble would impact retirement and college funds for millions of Americans.
• The economic repercussions will extend beyond stock owners, leading to decreased spending, job losses, and broader economic hardship.
• The hype cycle around AI is maintained by tech companies to keep investors and the public engaged, masking the lack of substantial progress.
• Products like Sora are seen as desperate attempts to maintain this hype and secure further investment, rather than delivering on promises of world-changing AI.

They know that they're shoving dollars into a bubble. They know that they have no plan. They know that money will be lost and they know that real people will be hurt.