Today's episode covers major developments across the AI industry, featuring analysis of Anthropic's explosive revenue growth, OpenAI's new policy proposals, and shifting public sentiment toward AI technology.

The discussion examines Anthropic's leap to $30 billion in annualized revenue, potentially surpassing OpenAI, while both companies face scrutiny over massive training costs approaching $30 billion annually. Google's Gemma 4 model release demonstrates the growing commercial viability of smaller, locally-run AI models.

The episode also analyzes OpenAI's 'Industrial Policy for the Intelligence Age' document, which proposes worker participation programs, public wealth funds, and tax modernization, though critics note the absence of concrete company commitments to fund these initiatives.

Anthropic's Revenue Explosion Reshapes AI Competition

Anthropic announced reaching $30 billion in annualized run rate (ARR), representing a 3X increase since the end of last year and 58% growth since February, potentially flipping ahead of OpenAI in revenue.

Enterprise customers with million-dollar annual spends doubled from 500 to 1,000 in less than two months, demonstrating unprecedented B2B adoption velocity.

Wall Street Journal analysis revealed both OpenAI and Anthropic expect to spend around $30 billion on training costs, with OpenAI's costs tripling year-over-year while Anthropic projects $28 billion by 2028.

Raw Malawalia criticized the financial engineering: 'OpenAI and Anthropic are incredibly profitable if you just strip out the training and inference costs. This business model is equivalent to running a passenger airline, except you need to replace your jets every six months.'

Google's Gemma 4 Signals Local AI Breakthrough

Google released AI Edge Eloquent, a dictation app running entirely on-device using the Gemma 4 model, demonstrating commercial viability for local AI applications.

Gemma 4 achieved 2 million downloads in its first week, compared to Gemma 3's 6.7 million downloads over an entire year, showing accelerated adoption.

Philip Schmidt demonstrated the model's agentic capabilities, showing it can query Wikipedia using agent skills while running on an iPhone, suggesting potential for mobile agent applications.

Anthropic signed a massive compute partnership with Google and Broadcom for 3.5 gigawatts of capacity coming online from 2027, addressing critical capacity constraints.

Meta's Token Maxing Culture Sparks Productivity Debate

Meta employees created 'Claudonomics' leaderboard tracking token consumption among 85,000 employees, with top users earning ranks like 'Session Immortal' and 'Token Legend.'

NVIDIA CEO Jensen Huang stated he would be 'deeply alarmed if an engineer on a $500,000 salary wasn't using $250,000 worth of tokens annually.'

Meta CTO Andrew Bosworth endorsed the practice: 'This is easy money. Keep doing it, no limit' - referring to engineers spending salary-equivalent amounts on tokens for 10x efficiency boosts.

Critics compared token maxing to 'Chairman Mao requiring peasants to smelt steel in their backyards during the Great Leap Forward,' questioning productivity measurement validity.

American AI Sentiment Deteriorates Despite Rising Usage

Quinnipiac poll shows 55% of Americans believe AI will do more harm than good, up 11 percentage points from last year and reaching majority negative sentiment for the first time.

70% believe AI will reduce job opportunities (up 14 points), while only 7% expect increased opportunities - a 10-to-1 negative ratio on employment impact.

Despite negative sentiment, AI usage continues growing: 51% now use AI for research (up from 37%), and those who never used AI dropped from 33% to 27%.

Tamila Triantoro noted: 'Younger Americans report the highest familiarity with AI tools, but they are also the least optimistic about the labor market. AI fluency and optimism here are moving in opposite directions.'

OpenAI's Policy Document Draws Sharp Criticism

OpenAI released 'Industrial Policy for the Intelligence Age,' proposing worker participation programs, public wealth funds, and tax modernization without concrete company commitments.

The document suggests giving workers 'a formal way to collaborate with management to make sure AI improves job quality, enhances safety, and respects labor rights,' though critics note it avoids mentioning unions.

Will Manitas criticized the absence of OpenAI commitments: 'The document proposes higher taxes on capital. OpenAI could commit to paying them. The document proposes a public wealth fund. OpenAI could seed it.'

Daniel Jeffries pleaded: 'Please, please, please, I'm on my knees begging every AI exec on the planet, just stop with this stuff. Just give us models. Let the collective, distributed intelligence of people figure things out in real time like we always do.'

Resources Mentioned

Project Management A Socio-Technical Approach 2024 Release ISE

things about Google's Gema4 family. First, unlike previous small models, this doesn't seem to be a research project. It is a commercially viable model for certain use cases, and Google seems intent o

Captivating Stories for Curious Kids Unbelievable Tales From History, Science and the Strange World We Live In

And yet, this is all despite adoption rocketing forward. The majority of people are now using AI to research topics they're curious about, rising from 37 to 51% over the past year. Analyzing data and