OpenAI uses the latest and greatest chips to teach their engine how to replace humans. It is trying to teach it everything.
China’s DeepSeek took a much different approach. It broke up the teaching into common units much like humans have jobs. It then takes the input and determines the best unit to handle the input. Thus, it is far more efficient and effective than OpenAI. Would you ask a doctor about a legal case or a lawyer about a medical issue?
As for AI and robotics, we are just scratching that surface. AI and computers can think on the level of a mouse. The curve is likely to continue to increase at an accelerating rate. The bigger factor will be human acceptance.
The real AI advances will be very specific and likely assisting humans. A good example is CoPilot and Curser that help developers write and debug software. The more efficient a developer becomes the more software they can write. If we improve the productivity of software programmers by 30%, the GDP increases as software itself makes humans more productive.
We can look at cars with self driving modes but the humans still have to be involved thus no real boost to gdp. We can look at medical fields and I could see less doctors needed as AI can use massive amounts of imaging and data to help find common issues with people. This helps drive the cost of healthcare lower and improves the speed of service including early detection.
As for China, they have some very real constraints. the biggest IMO, they don’t truly understand the technology coming out of the west. They simply implement what they can steal and it is pretty close to the wests version. Why this matters? In AI, it is only going to do what we teach it. If China teaches it something wrong, it will always be wrong.
www.stocktitan.net
