A recurring criticism of Sora 2 is that most people still can’t use it. The app remains largely invite-only, which has created a confusing and often unfair rollout. Some users receive access without applying, while others wait for weeks without any response. This uneven distribution has led to online “invite trading” and resentment among those locked out.

Veo 3, by contrast, is available to a wider audience in the creative industry, particularly through beta partnerships and test projects. That doesn’t make it fully public, but it has avoided the perception of being a gated product. For many professionals, this has meant more opportunities to experiment and share practical feedback.

For many Sora 2 users, the biggest barrier isn’t the model’s technical ability but what they’re allowed to do with it. Content filters block references to copyrighted material, political themes, and even vague allusions to violence or nudity. Several early adopters said the app felt more permissive at launch, only to become heavily restricted days later. The result is a creative experience that feels constrained and unpredictable.

Veo 3 faces a different criticism. It gives users far more freedom to generate footage, but artists argue that this freedom comes at the cost of precision. Fine-grained adjustments — such as lighting passes, shot composition, or multi-layer outputs — are largely unavailable. In VFX production, where small visual details often matter most, that lack of control makes Veo 3 unsuitable for anything beyond certain use cases like background shots or rough concept work.

Sora 2 has been widely criticised for inconsistent results. Users report repetitive visuals, shallow storytelling, and frequent generation failures. The gap between promotional demos and real-world performance is a common theme, and many say the final product feels no better than existing tools. Processing delays, interface glitches, and a repetitive content feed have further eroded confidence.

Veo 3’s reception is more positive on the quality front. Many artists admit they struggled to tell some generated clips apart from real camera footage, and they see value in using it for specific tasks. Yet the realism doesn’t always translate into reliability. The absence of post-production controls and the inability to produce outputs suitable for layered compositing mean it’s still not ready to replace traditional workflows for high-end projects.

The user sentiment around Sora 2 often goes beyond technical complaints. Many feel OpenAI overpromised and underdelivered — launching with flashy demos, then tightening features and moderation once public interest spiked. The term “enshittification” appears often, reflecting a belief that the app is already declining in quality and user focus.

Veo 3’s criticism is less about corporate strategy and more about industry impact. Artists worry about how easily clients might accept “good enough” AI video, reducing demand for traditional work. Others fear legal uncertainty — particularly around ownership and copyright — could make professional adoption risky. Ethical concerns, including potential misuse for deepfakes or fabricated evidence, also weigh heavily in discussions.

Sora 2 promised a creative playground but delivered a restricted, inconsistent experience that is frustrating.

Veo 3, gets some points for its realism and niche utility but lacks the control and legal clarity needed in professional settings.

Neither tool should be rejected outright, but both merit scrutiny. Users should have access without barriers, creative freedom without overreach, quality that matches the hype, and clarity about how these tools will fit into their workflows.

Until those demands are met, both Veo 3 and Sora 2 will remain powerful but imperfect and overhyped tools.