The Arrival of DeepSeek Marks the Beginning of Tangible Artificial Intelligence
In a thought-provoking piece by renowned economist Daron Acemoglu, published in the Financial Times, he ponders the potential future decline of the American economy, particularly in the realm of robotics. The piece begins by acknowledging the current tumultuous nature of our times and wondering if we're witnessing the start of a radically different era.
The discussion then leaps forward to 2025, imagining a historian looking back on this period as the beginning of American decline. The reasons for this decline, according to the piece, could be numerous. One prominent theory is the impact of DeepSeek, an LLM developed in a Chinese hedge fund incubator that performs tasks with ChatGPT-like sophistication at a fraction of the cost.
Sumant Wahi, one of Man Group's tech PMs, believes that this could have major implications for robotics. For decades, humanoid robots have been considered theoretically possible but impractical due to high costs, energy requirements, and overheating issues. However, DeepSeek has blown open these assumptions, suggesting that AI models can be trained and run on cheaper, less power-hungry hardware.
This could drastically lower the barriers to robotics, potentially making humanoid robots an economic inevitability. In fact, the cost of robots has already fallen dramatically, and Tesla's Optimus robot is expected to cost just $50,000 per unit, excluding software. If mass production drives costs down into the sub-$20,000 range, humanoid robots could become a viable option for consumers, especially in labor-intensive industries like nursing and food service.
The debate over whether robotics will take off in industrial settings or at the consumer level is no longer a clear-cut choice. As AI efficiency continues to improve, both markets could see significant growth. This could have major impacts on markets, from productivity to inflation expectations.
However, this rapid advancement in robotics also raises concerns about the potential displacement of jobs, particularly in sectors like logistics, agriculture, and manufacturing. The speed at which automation could occur could create significant policy challenges.
In this rapidly changing landscape, it's crucial for current administrations to consider strategies to circumvent these challenges. This could include investing in education and training, developing regulatory frameworks, fostering public-private partnerships, offering incentives for investment, and addressing economic challenges such as high interest rates and global economic instability.
In conclusion, the potential impact of DeepSeek on robotics is significant and far-reaching. The cost curve for AI and robotics is bending faster than expected, and we're entering a phase where AI and robotics are converging at a rapid pace. The future of AI isn't just digital, but physical. And it's coming much faster than anyone expected.
- Jensen Huang, the CEO of Nvidia, has been a strong advocate for robotics and AI, suggesting that the converging technologies could revolutionize various industries by 2025.
- In the realm of futurology, the integration of AI and robotics, fueled by advancements like DeepSeek, could lead to the widespread use of cobots in labor-intensive industries, such as nursing and food service.
- As a historian looking back on this period, they might ponder the inevitable impact of AI and robotics on job markets, particularly in sectors like logistics, agriculture, and manufacturing.
- At the 2022 Nvidia GTC conference, Huang showcased 'Omniverse Render,' a platform that uses real-time simulations and AI to optimize and automate render processes, further underscoring the potential of AI in the robotics sector.
- Predicting the future of AI and robotics, Acemoglu's piece highlights the inevitability of these technologies in various industries, challenging policy-makers and thought leaders to consider how they can adapt and prepare for the future.
