Lumian Gen AI Newsletter Issue #47

Welcome to the 47th edition of the Lumian Weekly Gen AI Newsletter!

AI isn't changing the world just yet. Hundreds of millions have tried ChatGPT, but most haven’t come back. Every big company has done a pilot, but few are deploying. Why? Because LLMs are impressive at first but not completely useful for everyday life. The key is finding product-market fit—building technology that people genuinely need and want.

When the iPhone launched in 2007, it wasn’t an immediate hit. It took several years and the introduction of the App Store in 2008 for it to start gaining real traction. Apple had to learn what users wanted and build a product that fit those needs. Today, over 2 million apps are available, and the iPhone is a cornerstone of modern life. But it didn't happen overnight.

Consider cloud computing. When Amazon Web Services (AWS) launched in 2006, it was a niche product. Companies were hesitant to move their infrastructure to the cloud. It took years of convincing and demonstrating the benefits before it became the standard. Now, cloud services are a multi-billion-dollar industry, but it required a slow and steady adoption process to reach this point.

E-commerce is another example. The dotcom bubble of the late 1990s saw a surge in online shopping ventures, many of which failed. However, the vision wasn’t wrong—it just needed time. Fast forward to today, and e-commerce giants like Amazon dominate the retail landscape. It took nearly two decades for online shopping to capture 20% of US retail sales.

ChatGPT rocketed to 100 million users in 2 months. But most people played with it once, then wandered off. It’s like they visited a new restaurant, tried the food, and decided they’re fine with their usual takeout.

Some see millions trying ChatGPT as a success. Others see most not returning as a problem. It turns out, convincing people to change habits takes more than a shiny new toy. Plus, the best versions often hide behind paywalls.

Enterprises are experimenting with LLMs, but few are diving in headfirst. Bain’s data shows lots of interest, but real deployments are scarce.

The rush to integrate LLMs came from competitive and market pressure, and the belief that this is the next big platform shift. These things look like finished products, and companies have tons of cash from the last decade, leading to enormous investments in what’s still experimental tech.

This rush means we skipped the slow grind of finding product-market fit. The web, e-commerce, and the iPhone all had to grow and learn to become useful. LLMs went straight to ‘it’s for everything!’ before meeting actual users.

Most technologies follow an S-Curve. In the beginning, progress is slow because the technology is new and unproven. This is the bottom of the "S". Once it finds its product-market fit, it rapidly grows and improves—this is the steep middle part of the "S". Eventually, growth slows again as the technology matures—this is the top of the "S".

Some believe LLMs might bypass the slow start of the S-Curve and go straight to rapid growth, forming a "J-Curve". This would mean immediate, explosive adoption and impact. However, this hasn't happened yet.

Some see this surge in investment as a bubble waiting to burst. But others see it as a bet that LLMs are a technology needing proper customer discovery to find their true purpose.

The Dotcom bubble’s dreams eventually came true, and AI maximalists might be right too. LLMs might revolutionize software and automate new tasks, leading to unprecedented growth.

Here’s the reality: This AI hype is just the beginning of something bigger. The startups betting on LLMs are early to the game, planting the seeds for future innovation. The real transformation isn’t happening this year, or even next, but it will come. AI has immense potential. The magic show has set the stage; now, it’s time to get to work and build the technology that will truly change lives. Let’s embrace the journey and keep moving forward.

Happy reading! 📚🤖🎵

In this week’s issue:

• News Flash: Microsoft x OpenAI, Stability AI Release, KyutAI's Voice AI

• AI Frontier: AI Twitter tools you can use today

• Fundraising: The biggest deals in AI

• Nerd Out: Technical and Business Content for Everyone

🚀 AI in Practice

Cutting-Edge AI Twitter Tools You Can Use Today

• TwitterBookmarks - Transform Twitter Bookmark Clutter into Clarity

• Readsss - Listen to your Twitter feed like an audio newsletter