- Sam Altman suggests a “compute budget” to ensure widespread AI access.
- He warns of potential power imbalances between capital and labor due to AI.
- Altman reiterates the proximity of AGI, while acknowledging its imperfections.
- He emphasizes the importance of scaling AI systems and the decreasing cost of AI usage.
- OpenAI anticipates making unpopular safety-related decisions regarding AGI.
Originally reported in TechCrunch: OpenAI CEO Sam Altman has acknowledged that the benefits of artificial intelligence may not be evenly distributed and proposed a “compute budget” as a potential solution. In a recent blog post, Altman outlined his vision for ensuring widespread access to AI technology, emphasizing the need for new ideas to address potential societal challenges.
Altman expressed concern about the potential for AI to exacerbate the power imbalance between capital and labor, suggesting that early intervention may be necessary. He introduced the concept of a “compute budget,” a mechanism to enable broader access to AI capabilities, although he admitted such solutions may be complex to implement. The increasing impact of AI on the labor market, with reports of job cuts and downsizing, underscores the urgency of addressing these concerns. Experts have warned of potential mass unemployment if AI’s rise isn’t accompanied by appropriate government policies and reskilling initiatives.
Altman reiterated his belief that artificial general intelligence (AGI), which he defines as AI systems capable of tackling complex problems at a human level across multiple fields, is on the horizon. He cautioned, however, that AGI will not be perfect and will likely require significant human oversight. He anticipates that AGI systems will excel in certain areas but struggle in others, and that their true potential will be realized through large-scale deployment.
Despite the significant investments required to develop and deploy AGI, Altman pointed out that the cost of using AI capabilities is decreasing rapidly, roughly by a factor of 10 every 12 months. This trend is supported by the emergence of capable and affordable AI models from companies like DeepSeek. While training and development costs are also potentially declining, both Altman and Anthropic CEO Dario Amodei have emphasized the need for substantial investments to achieve AGI and beyond.
Addressing the future of AGI systems, Altman indicated that OpenAI will likely make “unpopular” decisions regarding safety. He reaffirmed the company’s commitment to responsible AI development, even as it transitions to a more profit-driven corporate structure. OpenAI reportedly aims to achieve $100 billion in revenue by 2029.
Altman stressed that OpenAI’s goal is to promote individual empowerment while preventing the misuse of AI by authoritarian regimes for mass surveillance and control. He acknowledged the need for greater transparency and individual control over AI technologies, including potentially open-sourcing more of their work. He also addressed the company’s relationship with Microsoft, stating that they fully expect to remain partners for the long term, despite previous reports of a contractual definition of AGI that could have triggered renegotiations.
Altman’s blog post precedes the AI Action Summit in Paris, where other tech leaders are expected to share their perspectives on the future of AI.
