Sora's Potential Collapse: AI Video's Moment of Truth

OpenAI recently announced the discontinuation of its Sora application and associated video models, a mere six months after its initial launch.

This decision prompted a lively discussion on the latest episode of TechCrunch’s Equity podcast, where Kirsten Korosec, Sean O’Kane, and I delved into its implications for OpenAI and the broader industry. The move largely aligns with prevailing insights into OpenAI's strategic pivot towards enterprise and productivity tools, especially as it reportedly prepares for a potential initial public offering (IPO).

Kirsten notably characterized OpenAI’s choice to shutter Sora as “a sign of maturity that was nice to see in an AI lab.”

However, the cessation of Sora – coupled with ByteDance’s reported delay in the global rollout of its Seedance 2.0 video model – could also serve as a crucial reality check for developers of AI video tools and for proponents who prematurely claim these technologies are poised to replace Hollywood in the near future.

Below is an edited excerpt of our conversation, prepared for length and clarity.

Anthony initiated the discussion by underscoring that the shutdown extends beyond just the app itself. He noted that the app, in his view and perhaps others', was particularly unappealing, resembling "a social network without people, where it’s just nothing but slop." Beyond the app, it appears OpenAI is substantially scaling back its entire video development effort. Citing reports from The Wall Street Journal, which broke some of this news, Anthony explained that this strategic shift is driven by OpenAI’s pre-IPO focus on business, enterprise, and programming products, rendering consumer social apps and broader video initiatives a lower priority.

Sean concurred, admitting he never actively used the app, finding its premise off-putting for various reasons. He suggested the experience served as an internal reminder for OpenAI – not to diminish their achievements – of the "element of luck" that contributed to ChatGPT’s phenomenal success. While acknowledging ChatGPT’s undeniable value, evidenced by its reported usage numbers and sustained relevance over years, Sean pointed out a perceived hubris surrounding Sora’s launch. He likened it to an attitude of, “We built the most successful consumer product ever, and now we’re doing it again. And we’re going to bring in Disney and all this stuff.” Sean concluded that this outcome is a "harsh reminder" that there isn't always a direct shortcut to creating top-tier consumer products; sustained success requires genuine user meaning and engagement.

Kirsten, however, expressed commendation for OpenAI’s decision. She argued that while the "move fast and break things" mantra is often satirized, there is inherent value in companies that can rapidly iterate and then decisively discontinue underperforming products without succumbing to a sense of failure. She acknowledged the significant financial investment, referencing the "billion dollar deal" with Disney, and posed questions about the long-term value versus the actual expenditure, despite not having access to their balance sheets. Ultimately, Kirsten reiterated her view that despite the initial intrigue of what they could create, the decision to shutter Sora "showed a sign of maturity that was nice to see in an AI lab."

Anthony further elaborated on the implications for OpenAI, stating that the move appears entirely consistent with their evolving strategic direction and doesn't represent a significant setback for the future of generative AI. He found the situation particularly interesting within the video sector, noting its timing alongside reports of ByteDance’s generative AI video model, Seedance 2.0, facing delays. These delays are reportedly due to complex engineering and legal challenges, specifically concerning the integration of robust IP protections – an aspect seemingly not prioritized previously. Anthony framed this as a broader "reality check moment," recalling "hyperbolic statements," even from within Hollywood, that proclaimed, “We’re done, this is the future, it’s just typing in prompts and making feature films.” He concluded that, for a myriad of technical and legal reasons, such an outcome is "not that easy and we are very, very far from that happening."

Sean concluded the discussion by highlighting a significant underlying dynamic: this decision is one of several seemingly made since Fidji Simo assumed leadership of OpenAI’s day-to-day operations. He emphasized that Simo's arrival represents a profound shift within the company. Sean believes that as time progresses from her appointment to oversee the company's operations and, critically, determine the fate of its consumer products, it will become increasingly clear just how pivotal this period was for OpenAI.