🎙️ Check out my new Marc Hoag AI Law Pod, a podcast with AI “guests” courtesy of Google's NotebookLM. Now that NotebookLM allows customization, the AI “guests” actually interact with me at times, and deep dive quite a bit beyond the bounds of each issue’s content. If nothing else, it’s a very cool alternative way to enjoy Future Perfect while driving.

📈 Why your company is struggling to scale up generative AI (The Economist)

• Big tech leads in AI growth: Amazon, Microsoft, and Google report strong growth in AI services, with Amazon Web Services’ AI revenue growing at triple-digit rates.

• Public adoption outpacing businesses: Generative AI adoption among individuals is widespread (39% of Americans), but only 5% of U.S. businesses have scaled it beyond pilot projects, with many caught in “pilotitis.”

• Limited revenue impact: While AI revenue is climbing, it remains a small fraction of overall earnings for major tech firms and consultancies like Accenture, which reports $3 billion in AI-related work.

• Corporate caution over risks: CEOs in regulated industries like healthcare and finance hesitate due to potential privacy, regulatory, and reputational risks, contrasting with big tech’s eagerness.

• High costs and uncertain ROI: Implementing large language models (LLMs) is costly, and many executives are concerned about long-term returns, with recent surveys showing a dip in enthusiasm for AI investment.

• Scaling challenges: Effective AI adoption requires well-organized data and modern IT infrastructure; legacy systems and fragmented data often slow companies’ ability to harness AI effectively.

• Growing demand for AI skills: Job postings for AI roles surged 122% this year, with AI-skilled employees commanding significant pay premiums, reflecting a shift as employees move faster than many corporate strategies.

⚡️ Meta’s AI Abundance (Stratechery)

• Meta’s AI potential is underestimated: Despite past skepticism, Meta now appears uniquely positioned to capitalize on generative AI in ways that could make it the most valuable company worldwide.

• Advertising benefits from AI integration: Meta’s generative AI is poised to supercharge digital advertising, optimizing ad content and targeting with greater accuracy while boosting ad revenue and efficiency, especially in the wake of Apple’s ATT changes.

• Long-tail advantage in digital ads: Meta’s advertising thrives on a long tail of smaller advertisers, strengthened by a scalable, automated ad-buying system and now, generative AI, which can customize creative content on a massive scale.

• New advertising channels like messaging: Meta is experimenting with text-to-message ads, especially promising in markets where e-commerce and customer interaction rely heavily on chat, now boosted by generative AI for customer support and engagement.

• Cost-efficient AI-driven content creation: Meta’s AI tools are being tested for content generation, allowing for limitless, customized media that can be shown to users, maintaining engagement and improving monetization through personalized ad integration.

• Revenue implications of generative content: Meta’s shift to infinite, AI-driven content expands ad inventory and could drive ad prices lower, increasing competition and making it nearly impossible for competitors to keep up in both user engagement and ad revenue.

• Innovations in XR with AI and generative UI: Meta’s ventures into augmented and virtual reality (AR/VR) rely heavily on generative AI, with tools like Orion that promise minimal, intuitive interfaces—unlocking new ways for users to interact seamlessly in immersive environments.

• Financial edge through in-house AI development: Meta’s commitment to developing its own Llama models ensures it remains at the forefront of AI, potentially even subsidizing its R&D by licensing out models to others with fewer use cases.

• Infinite content strategy strengthens Meta’s role: AI-driven content abundance reinforces Meta’s aggregator model, with generative AI making the platform ever more engaging and scalable—positioning Meta to dominate in an AI-driven landscape of infinite digital content.

🤯 OpenAI’s o1 model leaked on Friday and it is wild — here’s what happened (Tom’s Hardware)

• o1 Leak Reveals Model’s Potential: A leak last Friday allowed limited access to OpenAI’s upcoming o1 model, showing capabilities far beyond previous previews and marking a significant shift from traditional GPT models.

• Enhanced Reasoning and Analysis: Unlike GPT-4 and other predecessors, o1 employs a reasoning process that allows it to deliberate before generating responses, enhancing accuracy and depth in handling complex tasks.

• Multimedia Processing: The o1 model can analyze images and multimedia, a feature demonstrated in the leak, where it accurately assessed an image puzzle and described elements of a SpaceX rocket launch image in intricate detail.

• Advancement Over Preview Models: Initial previews of o1 used earlier GPT model architectures, but the full o1 model, trained from scratch, offers a substantial leap in performance and responsiveness, as confirmed by user testing during the leak.

• Official Release Anticipated Soon: OpenAI has indicated that o1 may be released in the coming weeks, likely after the U.S. Presidential election, as the model could mark a groundbreaking improvement in AI’s problem-solving and multimedia comprehension capabilities.

✨ PowerPrompt™

This should be a fun one for the family:

Design a themed scavenger hunt that combines local history [of your town], iconic landmarks, and fun trivia about the area. Create each clue in the form of a riddle or puzzle, ending at a special location where participants can uncover a hidden ‘treasure.’ Include additional hints for each stop, with each riddle progressively getting harder to keep things engaging. Conclude with a list of local businesses or attractions participants could explore along the way for bonus points!”

Bonus Tip: Use AI to generate creative riddles or trivia questions tied to each location, and format the entire scavenger hunt as a downloadable PDF to share with friends or family!

🧰 AI Tools & Resources

• Google Learn About: Another Google Labs experiment to learn about anything with AI.

• WalloAI: Chat with your spreadsheet.

• Fairly AI: An interactive chart to study global AI regulation

🎉 THAT’S ALL FOR TODAY!

See you next time! 👋

Looking for past newsletters? You can find them all here.