Hugging Face co-founder Thomas Wolf just challenged Anthropic CEO’s vision for AI’s future — and the $130 billion industry is taking notice

Thomas monsterCofounder of AI Company Hug faceThe most optimistic meetings of the technological industry, claiming how the promises of the creators of today’s EU systems can handle the scientific revolutions, the most optimistic meetings of the technological industry have created more problems.

Saboteur Blog Post This morning was published on the private website, Wolf welcomes the widespread vision of Anthropic CEO Dario Amodei waiting for an advanced AI to deliver “Stated in the 21st century“Ten years of decades can happen in decades.

“I fear that the EU will not give us the 21st century” “Wolf, which is currently in the post of the current AI systems in the post”Yes country of men in the servers“Not”the country of geniuses“It imagines Amodei.

The AI ​​leaders emphasize that the technology of technology, which has the main effects of work strategies, research priorities and policy decisions, which have a growing division of technology.

From a flat student ‘Middle Medical Officer’ – Why does academic perfection not equal scientific genius

The worm justifies its criticism in personal experience. Despite being a flat student participating in the MIT, he clearly describes the “very average, pin, a secondary medicine” when he started his candidacy. This practice requires the compliance with academic success and scientific genius, past reward compatibility, the last compared to the latter compared to the formed mindset.

“The main mistakes that are generally made are usually the only good students to think of Newton or Einstein,” he explains the monster. “A real science test is Copernic against all knowledge about the days – in the terms of ML, ‘all training information’, we will be able to orbit the sun in another way.”

Amodei’s vision in October last year “Loving grace machines“Essay presents a radiant perspective in the 10x-100x human speed, which is a future exceeding the intellectual EU and Nobel Prize winners, can make progress in biology, neurology and other areas of a century.

Amodei, “Most Natural Infectious Diseases”, “Most Natural Infectious Diseases”, “Most Cancer Eliminating”, human life with human life with human life for genetic disease and a human life with human life. “I think the return to intelligence is high for these discoveries and everything in biology and medicine is mainly watching them,” he said.

Try the EU for fitting instead of creativity? Benchmark problem holding scientific discoveries

In the criticism of the Wolf, this main tension reveals a reality that often evades: Our evaluations are designed to compare disagreement with a way of thinking. Current AI Systems is superior to produce answers with the consensus of the existing knowledge, on the contrary, the control of the scientific revolutions, the paradigm is struggling with difficult concepts.

AI has invested very much to measure how the systemic systems can answer questions with the constructed answers, solve problems in known solutions and compliance with existing understanding. This creates systemic bias against systems that are more suitable than difficulty.

The monster is specifically criticizing the current AI evaluation standard “The final examination of humanity“And”Border math“It is more of the ability to create innovative assumptions or challenging existing paradigms, with difficulty questions with known answers.

“AI models, if we can find the correct answers to a number of questions we know, the test of these Etalons,” Wolf writes. “However, real scientific improvements will be advanced to answer known questions, but asking new questions and to ask general concepts and previous opinions.”

This marks a deeper issue that criticism is how conceptual intelligence is. Current focus on the number of parameters, training information volume and benchmark performance can be created by the AI ​​equivalent of excellent students, not revolutionary thinkers.

Billions are in danger: ‘Obedient students vs. How to shape the Revolutionary Dispute AI Investment Strategy

This intellectual division has significant results for the AI ​​industry and a wider business ecosystem.

The emergence of innovation from Amodei’s vision, computing power and more widely knowledge integration can be prioritized in an unpredictable innovation, an unpredictable size. This approach supports the strategies of favorable companies Anthropical, Open and other border EU laboratories raised collectively ten billions of dollars In recent years.

On the contrary, the prospects of the worm show that it is possible to protest, falsify and create novel hypotheses, and it is possible to prepare novel hypotheses and not to create a presentation methodology.

“We are currently currently building many submissive students,” he explains the monster. “It is ideal for today’s main purpose in creating great assistants and creating excessive appropriate assistants. However, they still do not give us scientific revolutions until you have a way to promote their knowledge and suggest the views against past training information.”

This argument for enterprise leaders who bets on the EU increases important strategic questions for an innovation. If Wolf is correct, the expectations of the revolutionary scientific progress must be greedy of the expectations that invest in the current AI systems. The real value may be in the placement of human-aI cooperation approaches, while the paradigm provides difficult intuitions in the development of existing processes in more growing processes or human-aI cooperation approaches.

$ 184 billion Question: Ready to convey AI to their scientific promises?

This exchange comes in a pivotal moment in the evolution of the AI ​​industry. For years, after an explosive increase in AI opportunities and investment, both the state and private stakeholders are increasingly aimed at practical returns from these technologies.

The latest information from Venture Capital Analytics Firm reached Pitchbook $ 130 billion in 2024Attractions of special interests with health and scientific discovery applications. However, questions about the material scientific progress from these investments were more insistent.

The Wolf-Amodei dispute represents a deeper philosophical division in the development of AI, which boils under the surface of the industrial talks. On one side, a scale of sustainable development, information volume and sustainable development in the training techniques in the data, which will give the revolutionary concepts of the revolutionary concepts. On the other hand, architectural skeptics may prevent the cognitive leaps that characterize the scientific revolutions of fundamental restrictions when the current systems are processed.

This dispute is especially important thing, both is the occurrence of two respected leaders in the development of the AI. Nor can it be fired uninforced or continuously for technological progress.

Outside the scalp: How can the EU need to think like the EU more scientific rebel

Tensions between these prospects are pointed to a potential evolution of how AI systems are built and evaluated. The critical of the worm does not offer to leave the current approaches, but rather new techniques and dimensions aimed at developing the control way.

In the article, Wolf offers the preparation of new criteria to test the scientific AI models “challenging their own training information knowledge” and “can admit to brave falsification approaches.” It is not for less AI investment, but a call for more thoughtful investment, for more thoughtful investment that takes into account the full range of cognitive capabilities needed for scientific progress.

This nuencing appearance accepts the many potential of the EU, when the current systems can be superior in special exploration types. The forward path covers the development of unilateral approaches to strengthen the strengths of the current systems when finding ways to solve the restrictions.

For institutions and research facilities navigating the AI ​​strategy, there are many results. Organizations must develop an evaluation framework for how the AI ​​systems respond to existing questions, but how new ones affect the new ones. Human-AI cooperation models, the paradigm of human professionals need to develop a cooperation models of AI’s pattern and calculation capabilities.

Find the average way: How can EU combine calculation power with revolutionary thinking

Perhaps the most valuable consequence of this exchange is the industry to direct a more balanced understanding of both the potential and restrictions on the EI. Amodei’s sight AI offers a compulsory reminder of the transformative impact that can be between more than one domain. Criticism of the worm It provides a necessary balance that really emphasizes specific types of cognitive capabilities needed for revolutionary progress.

This tension between optimism and skepticism between the industry is progressing, it is likely to manage the wave of innovation in the development of AI. By understanding both perspectives, organizations can develop a more nuanced strategy that increases the potential of existing systems, but also invests in approaches to resolve the restrictions.

So far, the opposite, on the contrary, the opposite, on the contrary, the contrary, on the contrary, we can only provide a more comprehensive approach to the development of our existence questions.