In a Reddit AMA, Sam Altman admitted that DeepSeek has lessened OpenAI's lead and said OpenAI has been “on the wrong side of history” in terms of open sourcing
To cap off a day of product releases, OpenAI researchers, engineers, and executives, including OpenAI CEO Sam Altman …
Context & Ripple Effects
Altman had already called DeepSeek’s R1 impressive for what it could deliver at its price and said OpenAI would accelerate releases after praising DeepSeek’s price-performance. The AMA extends that competitive acknowledgement into a public reassessment of OpenAI’s stance on open sourcing.
The comments also follow an AMA response that OpenAI was discussing releasing model weights and research but did not rank it as its highest priority while leaving weight releases unresolved. They put openness alongside capability and release speed as a live strategic question for the company.
First-order effects
- OpenAI’s leadership has publicly narrowed its claimed lead over DeepSeek, increasing pressure to demonstrate differentiation through forthcoming products and releases.
- Altman’s statement gives OpenAI users, developers, and observers a clearer signal that the company is reconsidering its historical posture toward open sourcing, without committing to release weights.
Second-order effects
- DeepSeek and other developers of more open AI systems gain validation that openness and price-performance can shape the competitive agenda of leading closed-model providers.
- A possible shift in OpenAI’s approach would force customers and developers to reassess the trade-off between proprietary hosted models and models they can inspect or deploy more directly.
Third-order effects
- If leading labs increasingly treat openness as a competitive response rather than a peripheral governance issue, model access could become a more important axis of AI market structure alongside raw capability.
- The episode also keeps the tension between OpenAI’s commercial imperatives and its broader mission in view, echoing earlier reporting on misalignment between its nonprofit and profit sides.
The trend: Competition from capable, lower-cost and more open AI models is pushing frontier labs to revisit whether closed distribution alone remains a durable advantage.