- AI Odyssey
- Posts
- Inside Meta’s big AI reorg
Inside Meta’s big AI reorg
PLUS: TikTok can generate AI songs, but it probably shouldn’t
Japan grapples with a severe labor shortage, impacting industries from construction to farming. As the nation faces a shortfall of 11 million workers by 2040, innovative solutions such as industrial robots in construction, self-driving vehicles in transportation, and robot labor in agriculture and retail become vital to address the demographic challenge and sustain essential services.
⏰ Top Headlines
🔗OpenAI has announced its first partnership with a higher education institution, Arizona State University.
🔗Australia creates a new advisory body to regulate AI risks.
🔗To cut food waste, supermarkets turn to AI to sell near-expired goods
🔗The rabbit r1 will use Perplexity AI’s tech to answer your queries.
🔗Google reveals $1bn UK data centre it says will create jobs and 'boost growth of AI'.
🔗AI voice-cloning startup ElevenLabs valued at $1.1 billion.
📲 Latest in AI…
META
Meta advances AI with Llama 3 training and organizational shift
Training Llama 3: Mark Zuckerberg reveals Meta has initiated training for Llama 3, the next-gen AI model, signaling a commitment to enhancing product capabilities through artificial general intelligence. Zuckerberg underscores Meta's pledge to open-source AI models and attracting top talent for ambitious problem-solving.
Organizational restructuring for AGI: To align with AGI objectives, Meta reshuffles its AI organizational chart. The Fundamental AI Research (FAIR) unit moves to the Family of Apps under Chris Cox, aiming for seamless synergy between AI research and product development.
AI infrastructure investment: Meta's dedication to AI is evident in hardware projections, targeting over 340,000 Nvidia H100 GPUs by year-end. The envisioned power equivalent of nearly 600,000 H100s reflects Meta's commitment to advancing AI capabilities.
Zuckerberg shared these updates in a concise Instagram video, outlining Meta's strategic moves toward artificial general intelligence.
SPONSORED BY WEBAI
In today's competitive landscape, leveraging AI can give your business a real edge. But developing AI solutions requires specialized skills and loads of time, putting it out of reach for many teams. What if there was an easier way?
Navigator by webAI makes AI accessible, allowing anyone to build AI-powered apps.
Accelerate your AI Projects: Get your AI models trained in minutes and tuned in hours.
Data & Model Ownership: Control and privacy are built in - train and run on your local devices
Designed for Everyone: Full code or drag-and-drop IDE
With Navigator, AI prototyping and deployment is finally open to all, regardless of technical expertise.
AI DEVELOPMENT
Davos 2024: From AI talk to action
AI dominated discussions at the World Economic Forum, with a notable shift towards actionable strategies. Boards press CEOs for tangible AI integration plans, moving beyond theoretical debates.
Storefront surge: Tech firms and consultants seized the opportunity, populating Davos with promises of AI-driven transformations. Accenture broke tradition, hosting a hands-on generative AI bootcamp personally led by CEO Julie Sweet. Accenture's sessions, attracting 150+ participants, delved into generative AI's risks, possibilities, and its impact on job roles. The move signifies a growing appetite for practical AI applications.
Executive demand: AI executives, including OpenAI's Sam Altman, were in high demand at official Davos sessions and side events, showcasing the industry's evolution beyond proof-of-concept stages.
Balancing act: Transforming with AI requires technical expertise and people skills. Accenture's research reveals only 10% of companies have all necessary components, emphasizing the importance of data infrastructure and HR systems.
Start small, learn big: For businesses not ready for a complete overhaul, Accenture suggests starting with existing tools like Salesforce or Microsoft 365, incorporating AI features to learn and prepare for future advancements.
Executives express optimism about AI's potential to address workforce shortages and boost global GDP, signaling a bullish perspective on AI's role in enhancing productivity.
AI SOCIALS
TikTok dabbles in AI songwriting, results not quite harmonious
TikTok is testing a feature called AI Song, allowing users to generate songs from text prompts using the large language model Bloom. Users input lyrics, and AI Song suggests sounds, letting them toggle the genre.
Early experiments and challenges: While not universally available, some TikTok users have tried AI Song with mixed results. Experiments show challenges with tuning, even with auto-tuning vocals. Notable attempts include quirky creations by users Jonah Manzano and Kristi Leilani.
History of AI songs on TikTok: AI-generated songs aren't entirely new on TikTok, with the infamous AI Drake and The Weeknd song "Heart on My Sleeve" gaining popularity. However, criticisms and out-of-tune results have surfaced.
Broader trend in generative AI: TikTok is not alone in exploring generative AI for user features. YouTube has also tested music creation tools, like Dream Track, allowing users to compose snippets in the style of popular artists.
Transparency measures: In response to the rise of AI-created content, TikTok has introduced features to identify such content on the platform. Additionally, updated rules require users to be transparent about their use of AI in their content.
💬 More on AI…
🔗Test yourself: Which faces were made by AI?
🔗AI-generated, virtual products are coming to a podcast near you.
🔗State of generative AI in the enterprise 2024.
🔗Tech firms points to AI as justification for layoffs.
🔗Samsung declares its latest smartphone marks a new dawn for mobile AI, but the reality doesn’t quite match the hype.
🔗AI-generated content is raising the value of trust.