When the AI Bubble Pops: Layoffs, AI Winter, and What Comes Next

Asian Dad Energy

114,504 views • 15 days ago

Video Summary

The video discusses the current state of AI investment and its potential for a significant market correction, drawing parallels to the dot-com bubble. The speaker, an ex-big tech engineer, details a hackathon experience where an AI customer support solution failed to impress a business leader due to its inability to address the most critical 10% of cases, highlighting the challenge of demonstrating real business value with AI. This anecdote serves as a microcosm for the broader industry, where billions are invested in AI with unclear returns, and many AI efforts are reportedly losing money, with companies like OpenAI allegedly losing significantly more per API call than they earn. The speaker predicts a potential AI "winter" characterized by reduced funding, project cuts, and mass layoffs across the tech sector and supporting industries, with the possibility of broader economic fallout including bank failures and a general depression if the bubble bursts.

An interesting fact from the video is that OpenAI may be losing six to seven times the amount of money per ChatGPT API call than they earn in revenue.

Short Highlights

A big tech company's AI hackathon yielded an AI customer support agent designed to handle 90% of cases, but a business leader deemed it a "glorified FAQ page" due to its inability to solve the remaining 10% of critical issues.
Billions of dollars are being invested in AI with unclear returns, and most AI efforts are reportedly losing significant amounts of money, with OpenAI allegedly losing 6-7 times revenue per API call.
The current AI hype is compared to the internet bubble of the 2000s, but on a much larger scale, leading to big tech companies employing "tricks" like inter-company product purchases and offshore layoffs to maintain appearances of growth and drive stock valuations.
The speaker predicts a potential AI "winter" leading to reduced R&D funding, project cuts, and mass layoffs across tech and supporting industries (semiconductors, data centers), with potential contagion into the general economy, financial sector, and even bank collapses.
Coping strategies for individuals include doubling down on core engineering skills, learning to use AI as a tool, growing a T-shaped skill set, building a substantial emergency fund (2+ years), reducing living expenses, and creating alternative income streams.

Key Details

An AI Hackathon and the Reality of Business Value [00:00]

The speaker, an ex-big tech engineer with 25 years of experience, recounts a company-wide AI hackathon aimed at monetizing billions invested in AI models and infrastructure.
His team developed a proof-of-concept for an AI customer support agent, using a large language model enhanced by RAG (retrieval augmented generation), intended to resolve 90% of customer support cases.
A business leader, Leo, critiqued the idea, stating that the remaining 10% of cases consume over 90% of the human support team's time, rendering the AI solution a "glorified FAQ page" that offered no real business value.

"So our solution is more like a glorified FAQ page rather than something that delivers actual business value."

The Broader AI Investment Landscape and Financial Losses [03:52]

The speaker posits that the challenges faced in the hackathon are mirrored across the entire industry, with hundreds of billions, even trillions, invested in AI yielding unclear returns.
Most AI efforts are described as losing substantial amounts of money, with OpenAI reportedly losing six to seven times the revenue it earns per API call to ChatGPT.
This situation is compared to the dot-com bubble, but on a significantly larger scale, where many ideas received funding but few generated profit.

"In my opinion, these are all either work in progress items or flatout science fiction that is unlikely to be realized in the extreme near future."

Tactics to Maintain AI Growth Appearances [05:31]

Big tech companies are allegedly employing various "tricks" to justify their AI spending and inflate stock valuations.
These tactics include buying AI-related products from each other to create an illusion of rising demand.
Mass layoffs of US-based employees are framed as AI efficiency gains, while simultaneously replacing those roles with cheaper offshore labor.
Accounting manipulations, such as reclassifying revenue from existing products as "AI revenue" after adding an AI feature, are also used.

"Tricks like accounting manipulations to show AI income from nowhere."

Predicting the AI Bubble Collapse and Its Ramifications [07:11]

The speaker outlines several potential triggers for an AI bubble collapse, including the failure of weaker AI-only companies like OpenAI, a major investor panic, or a social/political backlash against AI's environmental impact (e.g., data center energy consumption).
The overvaluation of "MAG7" companies and tech's dominance in the S&P 500 (40-50%) makes the market vulnerable to a correction when investors realize the promised near-term returns are not materializing.
A potential "AI winter" could ensue, characterized by reduced R&D funding, project cuts, and a shift in focus from broad AI feature development to efficiency gains.

"When that happens, the bubble will burst."

Downstream Economic Impacts of an AI Crash [09:14]

A crash would directly impact big tech companies, leading to significant reductions in market capitalization and substantial layoffs across the tech sector, affecting core AI jobs and adjacent roles.
Supporting industries, including semiconductors, electronics, networking, and construction (AEC companies), would face project halts and mass layoffs due to reduced demand.
The contagion could spread to the general economy, impacting real estate markets, the financial sector (potentially leading to bank collapses), and causing a broader liquidity crisis.
This could cascade into a general economic depression, characterized by mass unemployment, decreased consumption, and potential social destabilization akin to the Great Depression.

"This could even lead to bank collapses."

Individual Coping Strategies Amidst AI Uncertainty [12:12]

For those employed in engineering, the advice is to double down on core engineering and problem-solving skills, learn to leverage AI as a productivity tool, and develop a "T-shaped" skill set by acquiring ancillary skills in business development, product management, and QA.
Outside of work, it's recommended to significantly bolster emergency funds to cover potentially years of unemployment, reduce living expenses, and create diverse income streams through side hustles or personal businesses.