Tech entrepreneur David Sacks, co-host of the popular podcast “The Rainman” and newly appointed White House special adviser for AI and crypto, has recently posted on X that his predictions about AI’s future were wrong. Sacks declared that the fear of a single AI model dominating all is now unfounded, citing the emergence of high-performing competing models.
Sacks highlights the dynamic nature of the AI race, where five major American companies are competing to develop frontier models. This competition brings out the best in each company and helps America win the AI race. According to Sacks, this has avoided a monopolistic outcome that would vest all power and control in a single entity.
The co-founder also emphasizes the role of open-source models, which excel at providing 80-90% of the capability at 10-20% of the cost. He believes that American companies should compete in this area to offer more affordable alternatives to customers who value customization, control, and cost.
Sacks also notes that there will be a division of labor between generalized foundation models and specific verticalized applications. Instead of a single superintelligence capturing all the value, numerous agentic applications will solve ‘last mile’ problems, leading to innovation in the startup ecosystem.
Furthermore, Sacks stresses that AI models are still at zero in terms of setting their own objective function. Models need context, heavy prompting, output verification, and iterative process repetition to achieve meaningful business value. He cites a history of tech showing early dominant players falling to competitors not yet born, as seen with Google’s rise to dominate the search industry.
Sacks concludes that the current state of vigorous competition is healthy, propelling innovation forward and avoiding centralized control. This news is welcomed by experts who had feared China’s open models would dominate the global market.
OpenAI CEO Sam Altman will use his upcoming trip to Washington to argue that AI is already increasing American productivity and promising to keep AI “democratic” by making it accessible to as many people as possible.
Source: https://www.axios.com/2025/08/10/david-sacks-ai-goldilocks-scenario