A Strategic Power Play
While Scale AI claims it “remains an independent leader” committed to serving all its clients, I have serious doubts about whether companies like OpenAI will continue working with a firm so deeply tied to Meta. The conflict of interest appears unavoidable when a direct competitor partially owns your data labeling partner.
This acquisition represents a fundamental shift in Meta’s AI strategy. Rather than simply competing with existing models, they’re positioning themselves to control a critical piece of infrastructure that everyone else depends on. It’s like owning the factory that makes parts for all your competitors’ products.
Scale will continue to partner with leading AI labs, multinational enterprises, and governments to deliver expert data and technology solutions.
But will those partnerships remain as fruitful when Scale’s leadership is focused on building Meta’s superintelligence capabilities? I’m skeptical.
Apple’s Quieter Approach
Meanwhile, Apple took a more measured approach at WWDC, focusing on practical AI applications rather than grand visions of superintelligence. Their most useful feature was live translation, which works across messages, FaceTime, and phone calls—all running entirely on a device without sending conversations to the cloud.
Apple also revealed they’ve been developing their own image generation model comparable to DALL-E or Midjourney. While not groundbreaking, it shows Apple isn’t sitting on the sidelines in the generative AI race.
What’s notable about Apple’s approach is its focus on privacy and on-device processing. They’ve opened up their on-device Apple Intelligence to developers, allowing apps to leverage internal models without needing to connect to cloud services. This privacy-first approach stands in stark contrast to Meta’s data-hungry strategy.
The Video Generation Revolution
While Meta and Apple make strategic moves, the real innovation is happening in AI video generation. Several developments this week show how quickly this technology is advancing:
- Meta rolled out video editing with AI, allowing users to transform outfits, locations, and lighting in videos
- Midjourney began its “video rating party,” suggesting a public release of their video generation model is imminent
- Google launched VO3 Fast, which is twice as fast and is now available to Gemini Pro subscribers
- Leonardo AI added access to VO3, making it available for just $10/month (compared to Google’s $250/month plan)
The speed at which these tools are developing is remarkable. Just months ago, generating coherent AI videos was nearly impossible for average users. Now, multiple platforms offer increasingly sophisticated options at accessible price points.
The Legal Battles Begin
As AI capabilities expand, so do the legal challenges. Disney and Universal are suing Midjourney for generating images that allegedly infringe on their intellectual property, including characters such as Wall-E, Star Wars figures, and Shrek.
This lawsuit could set important precedents for AI-generated content. The outcome will help define whether AI models can legally generate content resembling copyrighted characters or if such outputs constitute infringement.
The case highlights a gray area between human creative expression (like cosplay) and machine-generated content. If I dress as a stormtrooper and make a video, am I infringing on someone’s copyright? What about if an AI generates that duplicate content? These questions will shape the future of generative AI.
What This Means for the Future
The Meta-Scale AI deal signals a new phase in the AI arms race—one where controlling the infrastructure becomes as important as building the models themselves. As companies position themselves for the development of superintelligence, we can expect to see more strategic acquisitions targeting the AI supply chain.
For users, this competition drives rapid innovation, as evidenced by the explosion in video generation capabilities. But it also raises questions about the concentration of power and whether a few tech giants will control the future of AI.
I believe we’re witnessing a pivotal moment in AI development. Meta’s acquisition isn’t just another business deal—it’s a fundamental restructuring of the AI landscape that could determine which companies lead the next generation of artificial intelligence.
Frequently Asked Questions
Q: What makes Meta’s acquisition of Scale AI so significant?
Scale AI provides labeled training data to nearly all major AI companies. Meta’s acquisition of a 49% stake means it now influences a critical resource that its competitors rely on. Additionally, Scale’s founder joining Meta to lead a “superintelligence lab” signals a major strategic shift in the AI industry.
Q: How does Apple’s AI approach differ from Meta’s?
Apple focuses on privacy and on-device processing, developing AI features that don’t require sending data to the cloud. Their approach emphasizes practical applications, such as live translation and image recognition, rather than the pursuit of superintelligence. Meta, by contrast, is making bold moves to control AI infrastructure while explicitly pursuing the development of superintelligence.
Q: What are the most exciting developments in AI video generation?
Several platforms made significant advances recently: Meta introduced AI video editing features, Midjourney began testing its video generation model, Google launched a faster version of VO3, and Leonardo AI made Google’s VO3 available at a much lower price point. These developments are making high-quality AI video generation accessible to average users for the first time.
Q: What’s at stake in the Disney/Universal lawsuit against Midjourney?
This lawsuit will help establish whether AI-generated content that resembles copyrighted characters constitutes infringement. The outcome could set important precedents for what AI models can legally generate and may determine how strictly copyright laws apply to AI outputs compared to human creative expressions.
Q: What is superintelligence, and why is Meta pursuing it?
Superintelligence refers to AI systems that can outperform humans across virtually all cognitive tasks. It goes beyond artificial general intelligence (AGI), which aims to match human capabilities. Meta is pursuing superintelligence to position itself at the forefront of AI development, potentially gaining competitive advantages across its products and services while establishing technological leadership in what could be the most transformative technology of our time.
