Meta has announced plans to invest between $115 billion and $135 billion in developing AI superintelligence, but the tech giant faces challenges with its latest AI initiative. The company’s new image and video model, codenamed Avocado, has reportedly underperformed in internal tests compared to competing models from firms such as OpenAI, Google, and Anthropic.
According to a report in the New York Times, anonymous sources within Meta disclose that Avocado failed to surpass Gemini 3.0 during evaluations. The model was initially slated for release this month, but disappointing test results have led to a delay in its launch until at least July.
In a notable turn, Meta is reportedly considering licensing Google’s Gemini in the interim to enhance its AI offerings. Spokesman Dave Arnold emphasized that the upcoming models will demonstrate both quality and an upward trajectory in performance, stating, “As we’ve said publicly, our next model will be good but, more importantly, show the rapid trajectory we’re on, and then we’ll steadily push the frontier over the course of the year as we continue to release new models.”
Meta’s commitment to AI development is evident, with a notable investment of $14.3 billion in the startup Scale AI. CEO Mark Zuckerberg has articulated the company’s ambitious goal to achieve a state of AI “superintelligence,” where AI capabilities exceed human intelligence.
In addition to Avocado, Meta is working on another fruit-themed AI model named Mango, both designed for image and video generation. This focus on AI aligns with the company’s interests in enhancing its wearable technology, including its rapidly growing line of smart glasses.
However, Meta’s efforts in the wearable market are also encountering hurdles. The company is now facing a class action lawsuit, with plaintiffs alleging that it misled consumers regarding how footage captured by its smart glasses may be processed. This follows reports from Swedish newspapers Svenska Dagbladet and Göteborgs-Posten, which revealed that contractors hired to label video content were exposed to sensitive and private scenes recorded by users. These contractors, based in Kenya, reported encountering intimate moments, including bathroom visits and sexual encounters.
The scrutiny surrounding Meta’s AI development and wearable technology underscores the complex challenges the company faces as it seeks to establish itself as a leader in AI and related sectors. As it navigates these obstacles, the broader implications of Meta’s advancements in artificial intelligence and consumer technology will likely resonate across the industry.
See also
OpenAI Integrates Sora Video Generator into ChatGPT, Boosting Multimodal Capabilities
AMD Silo AI Partners with UniMoRe to Develop Multimodal AI for Robotics by 2026
NBCUniversal Launches AI-Driven “Your Bravoverse” on Peacock Mobile, Curating 5,000+ Hours of Content
LinkedIn Reveals LLM-Based Feed Overhaul, Boosts Content Relevance by 30x with GPUs
Amazon Reviews AI Development Practices After Website Outages Affecting Millions
