Sam Altman releases three-point observations about AGI: AI usage costs are reduced by about 10 times per year

PANews February 10th news, OpenAI CEO Sam Altman said in a blog that AGI (General Artificial Intelligence) is gradually emerging, and its impact may surpass all previous technological changes. He put forward three key observations on the economics of AI:

1. The intelligence of AI models is logarithmic to computing resources —the investment in training computing, data and inference computing can increase infinitely, and achieve continuous and predictable returns.

2. The cost of using AI is reduced by about 10 times per year - for example, the word cost of GPT-4 in early 2023 decreased by about 150 times compared with GPT-4o's cost in mid-2024, far exceeding the growth rate of Moore's Law.

3. The linear growth intelligence brings super exponential socio-economic value - this drives investment to continue to increase and will not stop in the short term.

Altman expects AI agents (such as software engineering assistants) to become "virtual colleagues" and penetrate widely across industries. He believes that the economic impact of AI may be similar to transistors, the way in which the world changes is not easy to detect but extremely far-reaching. At the same time, the popularity of AI may lead to a sharp drop in the costs of certain commodities, while the prices of scarce resources such as land rise.

He stressed that policy and social consensus are crucial to the future of AGI. To ensure the widespread distribution of benefits of AGI, new economic models may be needed, such as providing “computing budgets” to everyone around the world. He believes that by 2035, everyone should be able to mobilize intellectual resources equivalent to global wisdom in 2025, which will greatly release human creativity.