| The economic, political, and media frenzy in the United States surrounding artificial intelligence (AI) evokes Alan Greenspan’s now-famous phrase, “irrational exuberance,” coined during the dot-com bubble. Investments planned or considered for AI development—including the acquisition of electronic components, the creation of data centers, and ensuring power supply—are reaching colossal amounts, while the potential revenues remain uncertain. Likewise, the impact of AI on employment and economic growth is still largely unknown. On Monday, January 27, the Chinese AI startup Deepseek disrupted Silicon Valley by achieving comparable results with only a fraction of the resources of American firms like OpenAI, Google, Meta, Microsoft, and Anthropic. DeepSeek spent only $5.6 million to train their V3 large language model (LLM), matching GPT-4. For a critical assessment see (1) Although the stock values of major players rebounded the following day, a deep sense of doubt lingered. The New York Times (2) summerised the critical issues as folllow : – Do leading A.I. companies like Google, Meta and the privately held OpenAI and Anthropic deserve their astronomical valuations ? – Do companies need to spend hundreds of billions on vast data centers powered by hugely expensive chips from Nvidia and others? Consider that OpenAI and its partners have promised to spend at least $100 billion on their Stargate project, or that Microsoft said it will spend $80 billion, or Meta $65 billion. – Does America need the huge uptick in electricity generation that has fueled a run-up in utility stocks? At the Davos Forum, Stanford Digital Economy Lab, Director Erik Brynjolfsson contended that with AI “there’s an amazing hype and billions of dollars are being wasted” as investments have yet to generate the desired returns. “We saw it with electricity, the steam engine, early computers, we’re seeing it now,” Erik said. “The real challenge, the bottleneck, is figuring out how to identify business value.” (3) Beyond geopolitical relations between the United States and China, the existential question raised by Deepseek concerns the openness to competition enabled by open-source software. In 2023, Meta provided free access to its AI technology, and it appears Deepseek leveraged this technology along with other open-source software available online. Yann LeCun, an early AI pioneer and Meta’s Chief AI Scientist, stated on LinkedIn that those interpreting DeepSeek’s success as China surpassing the U.S. in AI development are misreading the situation. “The correct reading is: ‘Open source models are surpassing proprietary ones,'” he said. Dr. LeCun added, “Because their work is published and open source, everyone can profit from it. That is the power of open research.” This raises the question: Why has no European startup anticipated or followed an approach comparable to Deepseek? Another instance of “irrational exuberance” concerns the impact of AI on employment and economic growth. During Greenspan’s era, the stock market soared, but between 1995 and 2000, the U.S. experienced strong economic growth, productivity gains, and full employment. Greenspan attributed this golden age to advancements in information and communications technologies (ICT). What Can Be Predicted About Artificial Intelligence? We can expect AI to contribute significantly to research across various fields. By definition, we do not know where research will lead in the future. However, it is clear that technological progress resulting from R&D is the main driver of long-term growth, as highlighted by economists like Philippe Aghion. Regarding the complementarity or substitution between labor and AI, we can speculate, but definitive conclusions remain to be demonstrated in practice. What is known is that the more perfect and widespread the substitution, the greater the inequality between workers and capital owners is likely to become. (1) Konstantin F. Pilz and Lennart Heim, “What DeepSeek Really Changes About AI Competition”, Just Security Org, February 3, 2025, https://www.justsecurity.org/107245/deepseek-ai-competition/ (2)”DeepSeek Forces a Global Technology Reckoning,” New York Times, Jan. 27, 2025 (3) Stanford Digital Economy Lab, https://mailchi.mp/stanford/digdig-01312025 |
- Comment
- Reblog
-
Subscribe
Subscribed
Already have a WordPress.com account? Log in now.
Monsieur Stekke 