Meta restructures Reality Labs, unveils new AI models

Meta, the tech giant formerly known as Facebook, is once again making headlines with a significant restructuring of its Reality Labs division. The division, charged with bringing CEO Mark Zuckerberg’s metaverse vision to life, will now be split into two distinct units: “Wearables” and “Metaverse.” This strategic shift comes amidst a series of workforce reductions and evolving market dynamics. While the restructuring will involve some layoffs, Meta emphasizes that this is a limited measure. Meta remains a major player in research and development, investing $38.48 billion in 2023 alone — the third-highest sum of any public company that year.

A two-pronged approach

Meta’s restructuring of its Reality Labs division into two distinct units reflects a strategic bifurcation of its metaverse game plan. This approach allows Meta to simultaneously advance both the hardware and software considerations:

• Wearables: This unit will concentrate on the development of hardware novel devices, such as the Quest VR headsets and AR glasses. Meta has won accolades for the Meta’s Ray-Ban Smart Glasses from a number of professional reviewers.
• Metaverse: This segment will continue to focus on the software and platform side, working on creating the virtual environments, experiences and tools that will populate the metaverse.

On a related note, the company recently released five new AI models to the public: Chameleon, a mixed-modal model that can understand and generate both images and text; a multi-token prediction model that helps train AI models to predict words faster; JASCO, a text-to-music generation model that offers control over the generated music; AudioSeal, an audio watermarking technique for detecting AI-generated speech; and a geographic disparities evaluation code and annotations system to help improve diversity across generative models.

Job openings point to Meta’s future ambitions

A review of Meta’s current job openings reveals a significant number of these openings concentrated in AI research and development, particularly in areas like computer vision, natural language processing and generative AI. Computer Vision: Meta is seeking experts in computer vision to tackle challenges such as photorealistic capture, 3D computer vision, and XR people generation. Current job postings include “Computer Vision Engineer – Photoreal Capture” and “Research Scientist, ML and 3D Computer Vision.”

Particular focus areas for senior and specialist jobs include the following.

• Natural Language Processing (NLP): Meta is actively recruiting NLP specialists to advance language models and generative AI. Openings for “Fundamental AI Research Scientist, NLP – FAIR” and “Research Scientist, Language Generative AI (PhD)” showcase the company’s interest in developing AI that can understand and generate human-like language.
• Generative AI: Meta is investing heavily in generative AI, the technology behind tools like ChatGPT. Roles like “Software Engineer (Technical Leadership) – Machine Learning” and “Fundamental Research Scientist – Generative AI.”
• AI Infrastructure: Beyond research and development, Meta is also bolstering its AI infrastructure. Positions like “Software Engineer – Datacenter networking” and “Software Engineer, Systems ML – Frameworks / Compilers / Kernels” emphasize the importance of building the underlying foundation for its AI/metaverse tech.

Meta salaries for tech roles tend to be competitive but not the most lucrative in Silicon Valley where companies such as OpenAI sometimes pay $500,000 for senior AI talent.

For now, Meta’s in a meta-position

Meta’s restructuring of Reality Labs into “Wearables” and “Metaverse” units, coupled with its aggressive AI and VR/AR hires underscores Zuckerberg’s metaverse and AI vision. As the company continues to invest billions in R&D and release AI models at a steady clip, it’s clear that Meta is playing the long game. Yet its ambitious strategy comes at a steep price: Reality Labs reported $3.85 billion in losses for the first quarter alone, with cumulative losses exceeding $45 billion since late 2020. Yet there are also signs that the company is turning things around, with its stock up 43% over the past six months. While its smart glasses (promotional video below) have achieved a measure of success that Alphabet’s Google Glass never did, only time will tell whether Meta’s costly bet on the metaverse will cement the company’s position as a pioneer of the next digital frontier or whether it will continue to be more aspirational than transformative.