AI Won't Be the Next New Platform

Did we learn nothing from Blockchain?

It’s a typical Saturday in March, and I’m listening to the insanely popular (and insanely divisive) All-In podcast. They share a clip from Doug Leone that I completely disagree with and inspired this essay (at 7:52). Read below to find out why!

Here’s what Doug said.

I actually think that AI is the next platform shift in the same way that mobile was the one before, internet was the one before, so I think AI is real. But I said earlier we're going to overestimate it in a short term we're going to invest in everything in the same way that in 1999 we invested in everything.

But then Google came out of that. Or Facebook came out of that. So I think you have to have a good head on your shoulder, where you don't practice fomo, where you don't chase every company.

AI is real, AI is the next platform, but how do we not invest in everything that walks. How do we make certain investments based on Market Maps, based on thought processes, that are more rational and not do every investment just because every other Venture firm is doing every investment.

I agree with Doug on the point that AI is real and that it could be huge. Like really huge. What I disagree with is how AI lifts the market up. I don’t think it looks like the next Facebook or Google, where there’s an opportunity to own a huge chunk of the social media market or the search market. I don’t think it looks like mobile, where there’s only room for two, maybe three, big platforms with high profit-margin app stores.

Once Android and iOS became the dominant platforms for mobile, there was a choke point that could be leveraged to scale and earn massive profits as the only game in town to buy your apps from for each respective type of device. I don’t see how a similar model could exist for AI.

A building block, not a platform

The market is full of gaps and cracks. Some things suck about every job at the task level. Things that are missing in every app that creates friction when people use it. Generative AI will be the mortar to fill some of those gaps. There will be a few big general models like ChatGPT, accessed via web-based UIs or APIs, sure, but most generative AI will show up as a feature of an app you already own. Or you’ll buy it as a customizable service for a specific use case or workflow you have.

AI will be the mortar we use to fill the gaps in workflows or apps that suck - the things people don’t want to do or aren’t efficient at doing. It will be the building blocks of platforms, not the platforms themselves.

There will be a lot of narrow, purpose-built LLM models that are very accurate and have little-to-no chances of hallucination due to their more narrow focus. In this way, I think AI, specifically generative AI, will be more like Linux, python, or Javascript - a new building block that makes up the next generation of LAMP-esque stacks that developers use to build applications. Microsoft 365 will continue to be the platform with AI features. AWS will continue to be the platform but with LLM services.

Are LLM models the new Big Data?

Perhaps databases are the best metaphor for this new generative form of AI. They store data, they answer questions via queries (like prompts), but instead of having to learn SQL and understand the schema to be successful, you ask questions like you’re talking to another human. The big data/NoSQL movement resulted in some big winners we all know the names of - you could point to Snowflake or MongoDB. However, you could just as easily roll your own using open-source software or use the big data/NoSQL database options that AWS, Azure, or GCP provide.

The signs are all there. Open-source generative AI is set to pass up commercial offerings quickly. They’re not worrying about legislation or signing an agreement to pause. They’re on fire. Amazon has already announced its generative AI offering. Microsoft, Canva, Adobe, and others are already shoehorning it into every product they have that has anything to do with content creation. The next big Google or Facebook to benefit from this AI trend is… Facebook and Google, and all the other tech giants already positioned to put it in the public’s hands nearly overnight.

Conclusion

That’s not to say that we won’t see a drove of AI startups start popping out of stealth later this year and throughout next year, but they’ll have to be positioned to pivot quickly or prepare for an early exit to avoid becoming obsolete very soon afterward.

The mistakes made with blockchain are instructive here. Investors put money into it as if it were the next huge platform, but it just turned out to be another type of database. One that didn’t really solve any critical problems but instead introduced new ones.

AI will be revolutionary. It will change how software is designed and will change how we work. It won’t fail like blockchain did since it has a real utility that hundreds of millions of people are enjoying on a daily basis already, but it also won’t be the next big platform.