Key facts
- Anthropic President Daniela Amodei discussed the 'tokenmaxxing' trend in AI spending.
- Amodei anticipates significant improvements in AI models over the next two to eight years.
- She believes businesses will find novel applications for AI tools.
- Anthropic does not use an AI usage leaderboard for its employees.
- Claude Code's high token usage contributes to the 'tokenmaxxing' discussion.
- High AI development costs are driving firms to seek public market capital.
- Anthropic has confidentially filed paperwork for an IPO.
Daniela Amodei, president and co-founder of Anthropic, addressed the 'tokenmaxxing' trend, where companies spend heavily on AI with uncertain business returns. Speaking at the Bloomberg Tech conference, Amodei expressed optimism about future AI capabilities, stating that current advancements are not the final stage. She anticipates significant improvements in AI models over the next two to eight years and believes businesses will find novel applications for these tools. Amodei also hopes AI will become more integrated into daily work routines, enhancing productivity. She clarified that Anthropic does not maintain an AI usage leaderboard for its employees, though general usage of its Claude products is tracked. Claude Code, Anthropic's AI coding tool, is a significant contributor to the 'tokenmaxxing' discourse due to its extensive token consumption. The discussion follows reports of internal AI usage leaderboard takedowns at companies like Amazon and Meta. Amodei also stated that the high cost of developing AI models is driving firms like Anthropic to look to the public market for capital, days after the company confidentially filed paperwork for an IPO, with a potential valuation of $960 billion.