The AI Economy is about to change

The PrimeTime

868,758 views • 4 days ago

Video Summary

The AI economy, particularly tokenomics, is showing early signs of strain as companies grapple with the high costs of AI development and deployment. Anthropic's "painted door" test, where they subtly removed access to their Claude code from a $20 plan, leading some users to unknowingly pay $100, highlights a move towards monetization. Similarly, GitHub Copilot shifted from a usage-based model to token-based, effectively increasing costs for those using more expensive AI models. This is driven by the immense financial pressure these companies face; OpenAI, for instance, is reportedly burning through billions monthly despite substantial investments. In contrast, Google, with its vast resources, can absorb these costs more readily, allowing them to invest heavily in AI without immediate profitability concerns. This economic reality suggests that the era of free or heavily subsidized AI access is likely ending, forcing companies to find more sustainable business models to cover the "Uber-expensive" AI usage.

One striking example of this financial pressure is Uber, which reportedly spent its entire year's AI budget within just four months, indicating the rapid and uncontained consumption of AI resources when usage is encouraged without clear cost controls.

Short Highlights

Anthropic conducted a "painted door test" by removing access to Claude code from a $20 plan, leading some users to pay $100.
GitHub Copilot transitioned from a usage-based model to token usage, impacting costs for users of more expensive AI models.
OpenAI is reportedly spending $5 to $7 billion monthly, consuming a $120 billion investment in 18-24 months.
Google is investing over $100 billion annually in AI and remains profitable, allowing for sustained development.
Uber spent its entire year's AI budget in four months, illustrating the high cost of AI usage.

Key Details

The "Painted Door" and Shifting AI Economics [00:07]

Anthropic employed a "painted door" or "fake door" test by altering pricing display to gauge user willingness to pay more.
Specifically, they removed Claude code usage from the $20 plan, causing a subset of users to be charged $100 per month without explicit price changes.
This experiment highlights a growing need for AI companies to recoup development and inference costs, which are substantial.

"Instead of doing a price change, they just simply removed clawed code usage from the $20 plan."

OpenAI's Financial Strain and the Need for Monetization [02:18]

OpenAI's $120 billion investment is projected to last only 18-24 months, requiring $5-7 billion in monthly expenditure.
The company is losing billions of dollars and must find ways to increase revenue to remain viable.
This financial pressure necessitates exploring new monetization strategies, like the one tested by Anthropic.

"They had to do this test because they have to know how much money they can make because if they don't make some sort of change, they're going to continue to lose billions of dollars."

Microsoft's GitHub Copilot Adjusts Its Pricing Model [03:02]

Microsoft's GitHub Copilot has shifted from a model based on a certain amount of actions to token usage.
This means that using more expensive AI models incurs higher token costs, reducing the number of calls one can make within a given plan.
This change reflects the varying costs associated with different AI models and the need for economic viability.

"If you use a more expensive model, well guess what? You got to use more of your tokens. Therefore, you won't get as many calls out of it."

Google's Financial Advantage in the AI Race [05:57]

Google is the real winner, pouring over $100 billion annually into AI development.
Unlike startups, Google can afford these massive investments and still generate profits, allowing them to innovate without immediate investor pressure.
This financial stability allows Google to compete aggressively on the AI frontier and potentially secure market dominance.

"They are pouring like a hundred plus billion dollars a year into AI and they can just do that. And guess what? After they pour hundred billion, $200 billion into AI, they still make money. That's wild."

The Escalating Costs of AI Usage: The Uber Example [06:48]

AI is proving to be significantly expensive, with companies struggling to manage consumption.
Uber reportedly spent its entire year's AI budget within four months, highlighting the unpredicted and rapid expenditure.
This raises questions about how employees are utilizing AI and the sustainability of widespread, unmanaged usage.

"Uber just got done claiming that within four months they spent their entire year's budget on AI."

The Future of AI: Viability and Reduced Access [07:24]

While AI is unlikely to be abandoned, the era of free or overly subsidized access is ending.
Companies will find ways to make AI economically viable, likely involving more controlled usage and increased costs.
The trend suggests a decrease in the amount of free usage users can expect, as companies shift towards more sustainable monetization strategies.

"They will find a way. Sure, it may be in a couple years, but you know what? They're going to find a way to make this thing viable. But for now, we're starting to see the cracks. Things just can't be as free as they once were."

The Excitement of AI and Personalized Experiences [08:35]

Despite the economic challenges, the potential of AI and technology remains incredibly exciting.
The ability to build almost anything imaginable is a unique privilege, making this an unparalleled time to live and innovate.
AI can serve as a powerful tool for learning, offering personalized documentation and explanations for complex code.