🕵🏼‍♀️ From AI Insights in 2024 to AI Action in 2025

Hello Sunshiners & Happy New Year!

If you're anything like me, you're likely still processing the sheer velocity and scope of GenAI's evolution throughout 2024. It wasn't just a year of incremental changes; it was a period marked by rapid innovation, unexpected breakthroughs, and, at times, a fair bit of head-scratching complexity. In my own quest to make sense of it all, I’ve spent considerable time immersed in a wide array of podcasts and articles, wrestling with the implications of each development—to the point where my Meta Ray-bans might actually need replacing!

To bring some clarity to this landscape, I've decided to approach this wrap-up in two parts, drawing from the expertise of one of my favorite GenAI analysts (and someone who I hope to bring as a guest speaker to an AI event here in the desert!). First, I'll share a distilled analysis of the top 15 most impactful AI products of the year, a perspective meticulously crafted by NLW. Second, I'll offer a month-by-month recap of the most significant AI news, tracking the ebbs and flows of progress as they unfolded.

It's this shared understanding of 2024’s landscape—both the standout products and the unfolding narrative—that I believe will naturally inform, and I hope, inspire our strategic approach as we define ambitious and impactful goals for our Palm Springs Coachella Valley AI Innovation Studios in 2025. I’m eager to share what I've uncovered and to begin this collective exploration with you as we chart our course together. Ready to venture forth?

PS - I’m gonna need a third screen for 2025!

The 15 Most Important AI Products of 2024

2024 was a landmark year for AI, with groundbreaking products reshaping industries and setting the stage for the future. Here’s a deep dive into the most influential AI innovations, from honorable mentions to the top game-changers.

The List at a Glance

• Honorable/Dishonorable Mentions: AI wearables (Humane Pin, Rabbit R1) and Apple Intelligence.

1. Eleven Labs and Heygen – AI audio and video avatars.

2. Devin – AI junior developer.

3. Enterprise AI agents (Sierra, Decagon, Ada, Glean).

4. Venice AI and Grok – uncensored AI advocates.

5. Meeting note-takers (Fathom, Read, Fireflies, Otter, Granola).

6. Suno and Udio – music generation tools.

7. Llama 3.1 405 B – Meta’s open-source GPT-4 class model.

8. ChatGPT’s Advanced Voice Mode with Vision.

9. Salesforce’s Agent Force – automating labor.

10. Video generation tools (VEO2, Sora, Luma Labs, Dream Machine, Runway, Pika).

11. OpenAI’s O1 model – improved reasoning.

12. Perplexity – the Google Search challenger.

13. Claude (Anthropic) – the product-driven LLM.

14. Cursor – the go-to AI coding tool.

15. Google’s NotebookLM – revolutionizing knowledge exploration with audio overviews.

The Countdown - From 15th to 1st

Honorable/Dishonorable Mentions

• AI Wearables: The Humane Pin and Rabbit R1 flopped, proving 2024 wasn’t the year for AI wearables. Despite the hype, these devices struggled with form factor, capabilities, and user adoption.
  Takeaway: AI wearables aren’t dead, but they need a major rethink to become practical and appealing.

• Apple Intelligence: After years of anticipation, Apple’s AI strategy underwhelmed. Delays and lackluster features left users questioning its value.
  Takeaway: Even tech giants can stumble in AI. Execution matters as much as vision.

15. ElevenLabs and Heygen

These tools are transforming audio and video creation, but they’re still waiting for mainstream adoption. ElevenLabs is pushing hard to become a key player in AI agent creation, while Heygen’s video avatars are gaining traction in niche markets.
Takeaway: AI audio and video tools are powerful but need broader use cases to go mainstream. We’re working on offering training in our Creator Track.

14. Devin

Devin promises to be a junior developer in your pocket, but at $500/month, it’s still a luxury for most. Its ability to build software from the ground up is impressive, and it’s currently positioned as a complement to human developers rather than a replacement.
Takeaway: Devin is a glimpse into the future of no-code tools, but affordability and scalability remain challenges.

13. Enterprise AI Agents (Sierra, Decagon, Ada, Glean)

These AI agents are already handling customer service for big enterprises, proving AI’s practical value. They’re low-hanging fruit for enterprises looking to dip their toes into AI.
Takeaway: AI agents are no longer theoretical—they’re here, and they’re solving real business problems. How can we provide this quality of CS solution to our local businesses?

12. Venice AI and Grok

Both platforms champion uncensored AI, pushing for privacy and open knowledge mediation. Venice AI focuses on privacy and uncensored models, while Grok (from XAI) advocates for less restricted AI.
Takeaway: As AI mediates more of our knowledge, uncensored and private-by-default models will become increasingly important.

11. Meeting Note-Takers (Fathom, Read, Fireflies, Otter, Granola)

Meeting summarizers are now a must-have in enterprises, with Granola leading the minimalist UX trend. These tools save time and improve productivity, but their full potential is still being explored.
Takeaway: Meeting AI is the first step toward broader AI adoption in the workplace. I like these tools, but still wish for a wearable to record, summarize and organize my IRL meetings and conversations.

10. Suno and Udio

These music generation tools went mainstream, with AI-generated tracks even making it into viral diss tracks. Musicians are embracing these tools as part of their creative process.
Takeaway: AI is becoming a collaborator in creative industries, not just a replacement. Part of SunshineFM’s roadmap is to play AI-generated music created by our local talent.

9. Llama 3.1 405 B

Meta’s open-source model closed the gap with GPT-4, making open-source AI a serious contender. Its performance is nearly on par with the best closed-source models.
Takeaway: Open-source AI is catching up fast, challenging the dominance of big players like OpenAI and playing the role of counter-balancer.

8. ChatGPT’s Advanced Voice Mode with Vision

This multimodal feature is changing how we interact with AI, though it’s still in its early days. It allows users to engage with AI through voice and visual inputs, creating a more immersive experience.
Takeaway: Multimodal AI is the future, and it’s already here. I have a team of virtual co-hosts, co-authors (see above byline), and co-founders as experiments in multi-modal AI.

7. Salesforce’s Agent Force

Salesforce is experimenting with AI agents that automate labor, signaling a shift in enterprise workflows. Its pay-per-interaction model is a bold new approach to AI pricing.
Takeaway: AI agents are poised to transform how businesses operate, starting with Salesforce. Yes, there will be consultants in the middle.

6. Video Generation Tools (VEO2, Sora, Luma Labs, Dream Machine, Runway, Pika)

Video AI had a slow start but ended the year strong, with tools like Sora and Pika unlocking new creative possibilities. These tools are making video production faster and more accessible.
Takeaway: Video AI is catching up to text and image generation, opening up new creative and commercial opportunities. My colleagues in Hollywood are not as thrilled about this, so it’s on us to help them out. Changing environment has Hollywood fading, but I can see a future where writers, craftspersons, artists et al gather together for a Desert Film Collective of sorts, and yes using GenAI.

5. OpenAI’s O1 Model

O1 focuses on improved reasoning, hinting at the next evolution of LLMs. It’s not perfect, but it represents a new direction for AI development.
Takeaway: The next generation of AI will focus on reasoning and problem-solving, not just scaling. Is this a new biosynthetic pathway for AI?

4. Perplexity

Perplexity is beloved by users and is the first real challenger to Google Search in years. Its focus on UX and research capabilities has made it a favorite for knowledge workers.
Takeaway: Great products can disrupt even the most entrenched giants. I use Perplexity every day, and when I ask folks I meet IRL, they’ve never heard of it. Yes, I live in an AI bubble.

3. Claude (Anthropic)

Claude’s product-first approach made it a favorite for coding and enterprise use, doubling its market share. Its Artifacts feature revolutionized how users interact with LLMs.
Takeaway: Product thinking is as important as model performance in AI.

2. Cursor

Cursor became the go-to AI coding tool, with developers flocking to its seamless workflow. It’s a prime example of AI’s impact on software development.
Takeaway: AI is transforming coding, making it faster and more accessible. Cursor, Replit, Plumb and others will be at the heart of our reskilling and upskilling our local workforce.

1. Google’s NotebookLM

NotebookLM’s audio overviews revolutionized how we explore knowledge, making it the most important AI product of 2024. Its ability to turn dense documents into engaging audio summaries is a game-changer.
Takeaway: NotebookLM shows that AI can create entirely new ways of interacting with information. It doesn't just give you summaries; it lets you converse with your notes, extract unexpected connections, and transform your thinking. So while it may seem like a novelty, it shows to me how GenAI will not just change our relationship with information but will change the very way that we work, study and grow.

Key Takeaways from 2024

1. User Experience is King: Products like Claude and Perplexity show that great UX is as important as powerful AI.

2. AI Agents are the Future: Tools like Salesforce’s Agent Force and enterprise AI agents are automating tasks at scale.

3. Open-Source is Catching Up: Meta’s Llama 3.1 405 B proves open-source models can rival closed-source giants.

4. Multimodal AI is Here: ChatGPT’s Advanced Voice Mode with Vision and NotebookLM’s audio overviews are just the beginning.

5. AI is Transforming Creativity: Tools like Suno, Udio, and video generation platforms are reshaping music and video production.

2024 was just the beginning. As these products evolve, they’ll shape the next decade of AI innovation. They will certainly shape our plans for 2025 and our AI Innovation Studios. Yes, it’s a thing!

Your turn. I'd love to hear about your experiments. Which of these products are most excited about? Do you use Sora or VEO2 for video? Suno or Udio for music? How about NotebookLM? Whatever you’re experimenting with, I want to hear about it.

Shoot me an email at [email protected] or join the AI Founders Club (and soon-to-be AI Innovation Studios). Bonus points if you've got thoughts on my 760-SUNSHINE concept – I'm seriously considering making that happen. Happy New Year too!

Researched and written with the help of SatGPT and BuzzAI - both Generative AI Agents