Gen AI - A collection of my thoughts
This time last year, I wrote about Gen AI use cases that excited me. A lot has happened since then, and I thought it would be a good time to revisit the topic. This note isn’t meant to be a comprehensive assessment of everything happening in the Gen AI space. Instead, it’s a collection of thoughts on areas that I personally believe hold promise (or don’t) and those that I’m excited about.
Turbocharging our curiosity
We’re living in the golden age of learning. Take any topic under the sun and an LLM can help you learn like never before. Learning can now be fun filled with practical, real world examples and implications. I can ask an unlimited of why’s and the LLM’s never get tired of responding. I can go down an infinite number of rabbit holes. They are also amazing at surfacing the right questions to ponder, helping us think critically and explore topics from fresh perspectives. And now I can do that via audio as well when taking a walk (via the podcasting feature in Notebook LM)! How I wish I had this available back in college!
I also use LLMs while reading books. Though the experience isn’t perfect (great use case for a startup to tackle!) and it slows down my reading, it has transformed how I read books. I can now pause and fully explore concepts before moving forward.
In my earlier note, I was excited about how LLMs can disrupt the ed-tech industry. A year down the line, I’m now convinced that this will happen sooner than later. While today’s models aren’t ready to create entirely new high quality curriculum, their ability to personalize existing ones is a game-changer. Imagine a curriculum that adapts to a student’s interests, making learning more engaging and accessible. I see this with my own and my son's learning – the AI adapts to our interests, making even complex topics engaging and accessible.
Everyone should have access to high quality education. High quality content can now be made accessible to any student around the world. LLM’s are already amazing at translating content; so language is no bar anymore. I am especially excited about how it can transform learning in a country like India. For us to become a developed nation, a high quality, accessible education for all is mandatory and I think AI can help us get there faster.
Hyper-personalized Experiences
The idea of personalized education can be extended to all of software. The idea of “software for one” will emerge where software can evolve uniquely for each user. As users of software, we all will potentially start with similar experiences while using an app but the experience can evolve based on how we use the app. The personalization will go beyond just rearranging buttons or changing colors. The software will actually learn how we think and work. This level of personalization means the same app can feel completely different for each user while delivering the same value.
Expanding the Pool of Software Engineers
As AI makes it easier and easier to write code and deploy new applications, in all likelihood, as an ecosystem we'll build more software - and existing software will get more complex and personalized. The addressable market for software will grow faster than our ability to automate its creation. We will not just build the same things more efficiently; we'll build things that was not possible until now. As a result, I expect the overall pool of software engineers to actually expand.
Building large companies with really small teams
Over the next few years, I believe we'll see a lot of AI-native companies building large businesses with a very small employee pool (with lot's of AI Agents!). We'll also see a lot of small companies with just one employee - the founder. They won't just be small teams using AI tools – they'll be small teams building and managing AI agents to deliver high quality outcomes. The question won't be how many employees you have, but how effectively you can use AI to create value.
Leveling the playing field for smaller companies
Previously only large companies could build complex software and leverage complex tech to differentiate in the marketplace. With AI making it easier to build software, it'll level the playing field for smaller companies as they potentially will have the ability to build the same software used by larger companies.
This shift will go beyond code. AI is democratizing everything from marketing to product design. A small e-commerce business can have the same level of personalization and customer insight tools that was once only available to retail giants. A two-person marketing team can create campaigns and content that can match those produced by large agencies.
This leveling of the playing field isn't just about replicating what big companies do – it's about enabling innovation from anywhere. Small companies can now experiment and iterate faster, and outpace larger organizations in some areas. The advantage will shift from who have the most resources to who can be the most creative and adapt to these new tools.
Ideas and Taste matter more than ever
In a world where AI can generate an infinite number of options, the ability to curate – to know what to keep, what to get rid of, and how to combine elements – becomes incredibly important. I see this in product development and product management, where success isn't just about generating features but about having the taste to know which ones truly matter. I see it in content creation, where the challenge isn't producing more content but creating experiences that resonate and stand out. The most exciting developments I'm seeing aren't about AI replacing human creativity, but about AI becoming a powerful tool that amplifies human ideas and taste. The key isn't in the technology itself, but in understanding how to wield it in to create magic.
Human bespoke experiences will be at a premium
"We will crave scarcity, human touches and stories. we’ll engage MORE when we see it, and less when we don’t - Scott Belsky"
As AI continues to improve and enables personalized experiences and efficient service, we humans will crave and appreciate experiences that are distinctly human even more - Restaurants creating unique dining experiences (and getting rid of QR code menus :)), event planners creating experiences that feel both intimate and grand etc. Combining digital efficiency with human connection will be a huge differentiators for businesses in the future. The future won't be about choosing between AI efficiency and human touch, but about understanding when each adds the most value and recognizing experiences that should remain deliberately, beautifully human.
AI as an Interface -
AI is changing how we interact with technology. Up until now, we've mostly navigated tech via apps, menus etc. We are now seeing a gradual shift to more natural conversations for a subset of use cases. Instead of learning different interfaces for different tasks, we're moving towards a world where the interface adapts to us, understanding our intent and context.
The initial attempts at this transition have had mixed results. Humane's AI Pin (which I personally loved) wasn't able to hit scale – the technology wasn't quite ready for prime time.
Meta's Ray-Ban smart glasses (and their experimental Orion Glasses) feel like a more practical step forward. They solve real problems while fitting naturally into our existing behaviors. What excites me most is how these new interfaces might help us from stop being glued to screens. The shift from screens to AI interfaces feels like a natural evolution with technology slowly fading away into the background
New Business Models
AI will enable new business models that weren't possible before. AI isn't just improving existing software, it's breaking through the traditional constraints of how software creates value. Software companies can now target markets that were previously out of reach, and serve customers who were previously unserviceable and break away from the per seat revenue model to more outcome based models. I highly recommend reading Aaron Levie's tweet/article on this topic - https://x.com/levie/status/1802512375939240065