Urgent.News

What's breaking now, across thousands of outlets.

AI

US competition watchdog expands investigation of Anthropic and OpenAI

Federal Trade Commission will demand information and seek testimony from the executives of leading AI companies

US competition watchdog expands investigation of Anthropic and OpenAI

At OpenAI's DevDay keynote, the company made a notable announcement regarding its marketplace, mirroring AWS's approach. OpenAI is offering a list of 32 partners for customers to use in place of their OpenAI spend commitments. This move indicates OpenAI's strategy to win on the merits of its models rather than through partnerships.

The company's partnership with Baseten, an open model provider, suggests a deeper philosophical shift in its approach to competition. In contrast, Anthropic prefers to use its own models exclusively, displaying a more rigid stance. The marketplace could become a decisive factor in customer decision-making, potentially leading to OpenAI's dominance in steering the customer's next purchase.

However, the fine print reveals that OpenAI's eligibility decisions and billing processes are cautious, limiting customer access. As vendors gain access to committed budgets, they risk relying on OpenAI's decisions, potentially leading to a marketplace dominated by OpenAI. The competition between OpenAI and Anthropic echoes the cloud war of the past, with the former recognizing the power to control customer decisions, while the latter maintains a more rigid approach.

Written by urgent.news from The Register's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at ft.com →

More in AI

What agent frameworks cost on the wire: measurements from agentic-arena

Most framework comparisons argue from feature lists. I wanted numbers, so agentic-arena holds the model, tools, datasets and iteration budget fixed and measures what each framework does differently.

  • Vanilla baseline uses 753.5 prompt tokens per item in a 15-item tool-use task
  • smolagents framework uses 3,295.5 tokens, a 3.90x increase over vanilla baseline
  • smolagents is the only framework that can handle three consecutive 429 errors

More from Wednesday 30 September →