What Did OpenAI Do This Week?

A DeepSeek storm + an early launch. Is OpenAI shaken? News count: +45

OPENAI COMES OUT FIGHTING RETHINKING.

What a week, and what a difference it made. OpenAI was in a tight spot, battling perception it was losing ground in the AI race to Chinese player DeepSeek - whom OpenAI suspects of IP theft [irony?]. Why? DeepSeek surged to the top of the App Store around the world this week amidst breathless claims of a US stock market crash. Kudos to DeepSeek for its ability to make a more efficient model without relying on as many advanced GPUs.

In response, OpenAI shipped more tools and features than your ‘average’ week [if it ever has those], and the o3-mini model preview launch was brought forward. Sam Altman summed up OpenAI’s fight back: DeepSeek's R1 is an “impressive model, particularly around what they're able to deliver for the price” and OpenAI “will pull up some releases.” Enter left, the early preview launch of o3-mini to “push the frontier of cost-effective reasoning.” O3-mini will be available to all ChatGPT users, the first-time free users of the chatbot will be able to try out OpenAI's reasoning models. A more technical early read on the o3 is here.

OpenAI also tried to strengthen ties with Washington. Pushing an ambitious data centre project and, oh yeah, quietly setting the stage for one of the biggest funding rounds in history [valuing the company at +$300 billion].

OpenAI researchers, engineers, and executives, including OpenAI CEO Sam Altman, also answered questions in a wide-ranging Reddit AMA. Altman admitted DeepSeek has narrowed OpenAI’s lead and acknowledged that OpenAI has been “on the wrong side of history” regarding open-source AI. While the company has dabbled in open-source releases before, it has largely favoured a closed, proprietary approach. Now, with rising competition and increasing scrutiny, OpenAI appeared to be rethinking that strategy - before the decision is made for them. Beyond prompting OpenAI to reconsider its release philosophy, Altman said that DeepSeek has pushed the company to potentially reveal more about how its so-called reasoning models show their “thought process.” [OpenAI employees were discussing this possibility but think the earlier models, not new ones.] Currently, OpenAI’s models conceal their reasoning, a strategy intended to prevent competitors from scraping training data for their own models. In contrast, DeepSeek’s reasoning model, R1, shows its full chain of thought. Is this a potentially seismic strategy shift?