I don’t know where else to share this so I am starting here.
I’ve been using Sora for a few days now, accumulating several hours of experimentation. Unfortunately, my experience has been underwhelming compared to my expectations. While I appreciate the potential of the tool and have had some slight enjoyment, particularly in generating short clips of cute animals or simple, fun scenes, its inability to produce anything of substantive quality is deeply disappointing.
Positives:
Sora’s text-to-video capability does offer a bit of entertainment when used casually. Experimenting with prompts to see what it generates can be amusing, especially for simple, static scenes. There’s a certain novelty in seeing how the AI interprets creative prompts, but this novelty wears off quickly.
Areas for Improvement:
- Challenges with Complex Actions:
One of the primary issues is Sora’s struggle with dynamic or complex actions. For instance, I tried to generate a video of a Norse man jumping aboard a longboat while it was docked. Despite several attempts and prompt revisions, the results were unsatisfactory. The generated video either had the character hopping aimlessly around the boat or behaving entirely inconsistently with the described action. Sora seems to lack the ability to follow nuanced prompts effectively or self-correct its output. - Distorted Movements and Hallucinations:
While occasional hallucinations are expected in AI-generated content, the degree to which Sora produces distorted or mutated actions—especially involving people—renders it almost unusable for creative storytelling. Characters frequently have disjointed limbs or perform incomprehensible movements, which undermines the output’s quality and usability. - User Experience and Expectations:
As someone who is not a professional artist, author, or videographer, I was excited by Sora’s promise as a creative tool for the public. However, the performance falls significantly short of what I would expect from a product marketed as “ready for public use.” It feels more like an early beta product than a polished tool. This gap between expectation and reality leaves me feeling misled by the advertising surrounding Sora. - Value Proposition and Pricing:
The pricing structure is also difficult to justify. While I currently have the Pro subscription, I am considering canceling it due to my dissatisfaction. The idea of paying $200 for incremental upgrades like a slight increase in resolution or video length feels unreasonable, especially when the base performance is so lacking. Sora’s current capabilities do not align with its pricing model, which only amplifies my disappointment.
Broader Implications:
From my experience, it’s evident that Sora is not yet a viable replacement for traditional artists, directors, or other creative professionals. If this tool is meant to be the “next step in creativity for the public,” it needs significant development to meet that promise. Moreover, there is a lingering concern that the truly powerful AI tools may remain exclusive to select entities, leaving the public with underperforming alternatives.
Conclusion:
I had high hopes for Sora as a fun and imaginative outlet, but it has consistently failed to meet even modest expectations. If significant improvements are not made soon, I cannot justify maintaining my subscription. I urge the development team to address these issues before marketing Sora as a publicly ready product.