Palantir CEO Alex Karp has a blunt take on the AI industry's hottest buzzword. Speaking on the TBPN podcast at Palantir's AIPCon 10 conference, Karp compared 'tokenmaxxing'—the compulsive overuse of AI tokens—to a porn addiction. His point was simple: many companies burn through tokens that appear productive but deliver nothing real.
Internal Tool to Curb Habit
Karp revealed that Palantir built an internal tool to help enterprises break the habit. The tool's name is not suitable for broadcast, but he described it as a 'demasturbatory, get off masturbation' mechanism. 'People are just sitting there all day like a porn addiction,' he said. He repeatedly circled back to the comparison, even as hosts tried to steer him toward the AI side of the analogy.
Understanding Tokenmaxxing
Tokens are the building blocks of large language models, breaking words into numerical units—a single token is roughly three-fourths of a word. AI providers typically charge based on token consumption and model choice. Recently, parts of Silicon Valley have turned against tokenmaxxing, the culture that previously celebrated nearly unlimited AI use.
Backlash Against Tokenmaxxing
Companies like Meta and Amazon reportedly built internal scoreboards to track token consumption, treating raw usage as a proxy for productivity. Most have since cut back. Uber COO Andrew Macdonald said the rideshare firm struggled to connect rising AI bills to real returns. Karp noted that until about two weeks ago, executives felt they couldn't publicly question AI without looking foolish, even as many privately admitted it wasn't working.
Palantir CTO Shyam Sankar made a sharper argument on a recent earnings call, calling Palantir a 'no slop zone.' More tokens, Sankar said, mean more slop.
Limitations of LLMs
Karp did not dismiss AI outright. He called it real and useful for narrow tasks, like writing a report on China's GDP growth. The trouble starts with harder, domain-specific problems, such as finding a cheaper, legal, and ethical way to drill for oil and gas. These require precise, ongoing processes that LLMs can enhance but not replace, he argued.
Much of this doubles as a sales pitch. Karp advised prospects to spend a few days with frontier companies like OpenAI or Anthropic first—and call him when they're done, because clients usually return 'clamoring.'
