Sam Altman Hints at More Open-Source AI as OpenAI Moves Toward AGI

OpenAI is exploring the open-sourcing of AI as it moves toward Artificial General Intelligence (AGI), CEO Sam Altman said in a recent blog post.

“AI will seep into all areas of the economy and society; we will expect everything to be smart. ” Altman wrote. “Many of us expect to need to give people more control over the technology than we have historically, including open-sourcing more.”

Altman outlined three key economic trends in AI—intelligence scaling predictably with resource investment, a rapid decline in AI costs, and super-exponential socioeconomic value from increasing intelligence.

He noted that AI token costs have fallen 150 times between early 2023 and mid-2024, a pace exceeding Moore’s Law.  “The cost to use a given level of AI falls about 10x every 12 months, and lower prices lead to much more use,” he said. 

He warned that while AGI’s benefits could be widely distributed, its impact will be uneven across industries. Some sectors may remain largely unchanged, while scientific progress is expected to accelerate significantly.

“The balance of power between capital and labour could easily get messed up,” Altman said, suggesting that intervention might be necessary. “OpenAI is considering strange-sounding ideas such as providing a compute budget to enable global AI access.”

AI agents, which Altman described as virtual co-workers, are being rolled out. He projected that software engineering agents could eventually handle tasks similar to junior developers at top firms, though requiring supervision. He also speculated on the impact of deploying such agents at scale across multiple industries.

Altman acknowledged risks in AGI deployment and noted that public policy will play a crucial role in shaping its integration. He cautioned against authoritarian control of AI and stressed that individual empowerment should be prioritised.

OpenAI continues to launch products “early and often” to allow society and technology to co-evolve, Altman wrote. He emphasised the long-term goal of making AGI a tool that enhances human capability, stating that by 2035, “anyone should be able to marshal the intellectual capacity equivalent to everyone in 2025.”