Breaking news

Alex Karp Criticizes AI Labs For Focusing On Token Consumption

Enterprise Frustrations With AI Frontier Labs

Palantir CEO Alex Karp said some enterprise customers are becoming frustrated with frontier AI companies, arguing that providers are often focused on increasing AI usage rather than solving specific business problems. Speaking with CNBC’s Sara Eisen, Karp criticized what he described as “tokenmaxxing”, an approach that prioritizes greater consumption of AI services instead of delivering measurable outcomes for customers.

Rising Costs And Implementation Challenges

Karp’s comments come as companies continue to increase spending on artificial intelligence tools while facing growing questions about costs and returns on investment. He argued that large language models remain important, but said the greatest value will come from how businesses implement the technology. “It is not that large language models aren’t crucial for the world,” Karp said. “It’s just the implementation is where the value is, certainly in the next seven years.” According to Karp, execution and integration will play a larger role in determining commercial success than access to the models themselves.

Market Dynamics And IPO Momentum

The remarks were made as the AI sector continues to attract significant investor attention. Several leading companies, including OpenAI and Anthropic, are moving toward public market debuts while competition among AI providers continues to intensify.

OpenAI recently confidentially filed for an initial public offering, while Anthropic has also been linked to IPO plans following strong growth in its valuation. The developments reflect continued investor interest in artificial intelligence despite growing scrutiny over infrastructure costs and profitability.

Political And Strategic Implications

Karp also addressed the broader implications of AI development and its influence on public policy. While he has publicly disagreed with some industry leaders on aspects of AI regulation and deployment, he acknowledged the role prominent executives and researchers play in shaping the future direction of the sector.

The discussion highlighted ongoing debates around governance, adoption and the long-term economic impact of artificial intelligence as governments and businesses increasingly incorporate the technology into decision-making processes.

Looking Ahead

Karp said businesses should remain focused on practical outcomes rather than ideological debates surrounding AI. As adoption accelerates, technology companies are likely to face increasing pressure from customers and investors to demonstrate that AI investments translate into measurable business results.

A New Twitter-Inspired Social Network Is Taking Shape

A new social network called Twitter.now is entering the market, with a founding team that includes former Twitter trademark counsel Stephen Coates. The service is being developed by startup Operation Bluebird.

As Ars Technica reported, X sued the company last year and asked a Delaware judge to block the launch. Operation Bluebird argued in a petition that X had abandoned trademarks including “Twitter” and “Tweet.”

Coates has said the project is not an attempt to recreate the original Twitter. In a LinkedIn post, he described the platform as a new public space focused on trust, transparency and user choice.

AI System To Rate Posts

Twitter.now is currently being tested, with early access priced at $20. Its main feature is VERA, an AI system designed to evaluate posts, verify claims and provide sources and context.

Posts receive a trust score, with users eventually able to set a minimum score to filter their feeds. The company says this approach will give people more control over what they see instead of leaving those decisions entirely to an algorithm.

Moderation Remains A Challenge

Scaling moderation will be one of the platform’s biggest tests. Social networks have repeatedly struggled with content moderation as their communities grow, and newer platforms such as Bluesky have faced similar criticism.

Operation Bluebird says VERA will form the basis of its moderation and verification system. A second version is already planned, with expanded tools that would let users set a specific trust threshold for the posts appearing in their feeds.

For now, Twitter.now remains in an early testing phase, combining the familiarity of the Twitter name with an AI-driven approach to evaluating online information.

eCredo
Aretilaw firm
The Future Forbes Realty Global Properties
Uol

Become a Speaker

Become a Speaker

Become a Partner

Subscribe for our weekly newsletter