AI: Platforms and Applications

...the perennial conflict, comes to AI

At the dawn of every potential platform in technology, that needs the vigorous participation of third-party developers to build applications and/or services on top for consumers and businesses, there remains an uneasy balance of ambitions.

From the platform’s perspective, the clarion calls goes out, ‘Come one, come all, come and build something amazing on our platform’ for our millions/billions/jillions of users. But at the same time, the mostly unspoken warning also goes out: ‘don’t build something so great that we’ll build it ourselves’. Sometimes that’s said in hindsight that ‘it was obvious that Platform X was going to go and build Application and/or Service Y’. It’s obvious, isn’t it?

Like clockwork, in this AI Tech Wave too, this is happening again as OpenAI unveiled its upgraded GPT-4 Turbo on its first ‘DevDay’, with the opportunity for Developers to build millions and presumably ‘jillions’ of GPTs for users, be they consumers or businesses. And of course a plan to roll out an ‘App Store’ a la Apple, where said developers could sell their wares. Lots of possibilities for where OpenAI is heading.

But at the after-parties and gatherings post the Keynote, Sam Altman's advice to developers had more nuance. He said about building on OpenAI's new GPT and GPT store platform:

“I think, on the whole, most developers will be happy for the additional distribution that ChatGPT provides for what they create. But as we say, don’t build a thin wrapper on top of OpenAI. We are planning to build the obvious features that you would expect for a robust platform over time, and there is an enormous value to build.”

Key phrase here is ‘obvious features’. It’s never clear what’s obvious in the beginning. It’s always obvious in hindsight. For every Lotus 123 and Wordperfect (spreadsheet and word processing) application on the emerging PC platform by Microsoft and Intel, it was ‘obvious’ that Microsoft would at some point offer Excel, Word, and eventually Outlook and ofcourse, Office.

It was obvious in hindsight that as Facebook rolled out its Developer platform in the late 2000s at its F8 Developer conference, that developers were welcome to build applications like Zynga’s enormously popular ‘Farmville’ and so many others, but it was obvious that Facebook may roll their own down the road.

It was obvious also when so many App Developers rolled out ‘Flashlight’ apps for the iPhone App Store in the 2010s, that Apple would eventually make a ‘Flashlight’ app of its own in iOS. Ok, that was partly tongue-in-cheek, but you get the point. Apple did roll their own when it came Messages and Notes for example on its platform.

It was said after OpenAI’s DevDay, that thousands of AI startups were in trouble given the range of tools, features and services OpenAI rolled out vs what so many AI startups had been working feverishly to build. But the buzz and race is already on to fund the next generation of applications on OpenAI’s next iteration of their platform. Be they GPTs, Apps, Services, and of course ‘Smart Agents’ with a wide and deep range of capabilities.

Indeed, Google is rumored to be investing additional sums in Character.ai, a private AI unicorn that kicked off offering millions of AI driven ‘smart agents’ to almost a 100 million users at one point.

So regardless of whether App and Service developers on platforms are potentially subsumed by competition from the underlying platform, the opportunity to build said application of service is always going to be tempting for developers and investors in the early days.

But it’s a cat and mouse game, with a number of elephants also trundling around at speed driven by the same forces of fear and greed. It’s important for all to stay nimble. Stay tuned.

(NOTE: The discussions here are for information purposes only, and not meant as investment 

advice at any time. Thanks for joining us here)