Sora rollout receives mixed response from AI filmmakers

Ten months later, through eye-catching, vivid video previews, OpenAI finally releases its AI video generator model Sorareleased to the public on Monday.

In the two days since, however, the debut hasn’t been perfect: Filmmakers early adopters of artificial intelligence have reported surprisingly inconsistent and unrealistic results from Sora, especially against leading competitors like Runway, Luma and others. Compared with artificial intelligence film creation tools, Hero, Kline and Tencent New Hunyuan.

Others have raised questions about OpenAI Content restrictions Violence and explicit content, even cartoons or non-serious visuals, are prohibited. According to a post from OpenAI, OpenAI has temporarily closed Sora account creation in response to unexpectedly high demand. OpenAI CEO Sam Altman spoke at X yesterday.

Sora’s rocky launch has already attracted some staunch AI critics, such as PR agents Ed Chitlen Suggesting this is a “bait and switch” to win positive press coverage for OpenAI even though the company is technically unable to actually deliver reliable inference models to the masses.

Reactions ranged from impressed to disappointed

Regardless, those who have been able to access the tool starting this week (or earlier, when OpenAI pre-seeds it to select alpha and beta testers) have reported a variety of experiences, ranging from underwhelming That’s disappointing, especially considering the price point for accessing it: $20 per month for 50 generations with a ChatGPT Plus subscription, $200 per month for unlimited generations with ChatGPT Pro.

“No, Sora hasn’t arrived yet!” wrote the creator Umesh on X. “Hailuo AI looks much better. I just tried four different generations of prompts to achieve the task that Hailuo AI does easily, and none of them succeeded.

Likewise, artists PurzBeats Posted on X Says Sora is “[p]Probably only the pro version is worth it” and they have experienced “[v]Among other complaints, “everything but the subject matter is very strange and unstable” for their generation.

“OpenAI has been lying to us!” wrote Indie filmmaker el cine on X. “It failed in every way, with most clips failing to work or even follow the prompts correctly,” they noted. They posted a clip of a generation in which people walked backwards with their legs facing opposite sides of their torsos and heads. Ultimately, they concluded: “Think twice before choosing a professional plan.”

Others were more impressed with the results, including futurist podcaster Ed Krassenstein, who called the model “amazing” In a post on X, based on his experience using X to create quick clips. he posted a four-minute long The Sora movie, produced by fellow creator KNGMKRlabs, features cavemen in a documentary-style show called “The First Humans,” which, in my opinion, looks very real and compelling.

Competitive markets reduce room for mistakes and tinkering

Still, Sora’s debut appears to be challenging, to say the least, as the AI-powered video generator struggles to compete for users and rolls out new features that allow the masses to create Hollywood-caliber movies.

For a real Hollywood studio OpenAI and rivals reportedly courtingcompetitors may currently have an advantage. For example, already Runway has signed an undisclosed deal with Lionsgate supply John Wick The studio has custom AI models trained on its catalog of more than 20,000 movies and TV shows.

Particularly for those looking to shell out the money for the “Pro” subscription tier, the question is whether Sora is worth it now, or whether other AI generators with similar or cheaper pricing structures are better options. Sora’s current production volume and relatively high entry price (unlike other AI video generators, it doesn’t offer a free tier) may make widespread adoption more challenging.

In response to these reactions, an OpenAI spokesperson emailed VentureBeat the following statement, excerpted from Sora’s official blog post: “The current model still has room for improvement. It can have difficulty simulating the physics of complex scenes, and it may not understand specific instances of cause and effect (for example: a character may not display a marker after biting a cookie). The model can also confuse prompts spatial details, such as distinguishing left and right, or it is difficult to accurately describe events unfolding over time, such as specific camera trajectories.“

An OpenAI spokesperson also pointed out: “We’ve seen huge demand for Sora.”