The AI Daily Brief: Artificial Intelligence News and Analysis

OpenAI's New Deal

37 min

This episode discusses OpenAI's new policy document, 'Industrial Policy for the Intelligence Age,' which proposes various societal and economic shifts for the AI era but faces criticism for its lack of concrete commitments and failure to articulate AI's benefits. Meanwhile, Anthropic's revenue has surged to an annualized $30 billion, potentially surpassing OpenAI's reported run rate, driven by enterprise customers and a massive new compute deal with Google and Broadcom. The podcast also touches on Google's productization of Gemma 4 and Meta's upcoming model release and internal 'token maxing' culture.

Summarized by Podsumo

🎧 Listen 🎙️ Ask about this episode

✨ Key Takeaways

1

Anthropic's annualized revenue has reached $30 billion, a 3x increase since last year, potentially surpassing OpenAI's reported run rate, fueled by enterprise customers and a significant compute partnership with Google and Broadcom for 3.5 gigawatts of capacity.
2

OpenAI's 'Industrial Policy for the Intelligence Age' proposes ideas like worker perspectives, a public wealth fund, and tax modernization but is criticized for being too technocratic, failing as PR, and lacking specific commitments from OpenAI itself to fund or implement these policies.
3

Both OpenAI and Anthropic face sky-high model training costs (OpenAI $30B this year, Anthropic $28B by 2028), leading them to present alternate profitability metrics excluding these costs, with OpenAI expecting cashflow positive by 2030 and Anthropic by 2028.
4

Google launched an on-device AI dictation app using Gemma 4, showcasing its commercial viability and potential for local models, while Meta prepares its new model and grapples with an internal 'token maxing' culture where engineers are encouraged to consume vast amounts of AI tokens as a proxy for productivity.

💬 Notable Quotes

"“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. Bizarre to have another definition of earning simply because we don't like the costs.” — Rah Malawalia"
"“Hard to believe that just 18 months ago, Anthropic was broadly considered the odd man out of the AI race, with an ambiguous business plan and no clear funding model. Not so anymore.” — John Arnold"
"“How does measuring productivity by total token consumption make any sense at all? Comparing it to Chairman Mao requiring peasants to smelt steel in their backyards during the Great Leap Forward, which of course led to tons of useless low grade steel, Joe continues, real backyard steel furnaces vibe in my opinion.” — Joe Weisenthal"