GPT – Keeping a Firm Grasp on the Obvious

By Geoffrey Moore

Author – The Infinite Staircase: What the Universe Tells Us About Life, Ethics, and Mortality

By virtue of its amazing conversational capabilities, GPT (Generative Pre-trained Transformers) is taking the world by storm. This has sparked, among other things, a wave of commentary about the need to regulate AI. While legitimate in its own right, I fear it might be distracting us from the main event. 

GPT represents a deeply disruptive innovation in user experience, one that will reengineer a whole host of natural language interactions, many of which today are still unautomated.  Most importantly, it will do so in a remarkably non-disruptive way from the point of view of infrastructure-providing software vendors and the admins that support their applications. That is, it will generate maximal immediate benefit from minimal investment. This is a formula that drives mass adoption, or what we call the Tornado. 

One limiting factor for this tornado is the Large Language Model, and its potential to leak proprietary information and spread disinformation. This is where the regulatory concerns come into play. But for goodness’ sake, there is a massive target-rich environment of non-proprietary information causing widespread frustration and confusion that we can tackle immediately, on a low-risk, no-regrets basis. In particular, for subscription-oriented enterprises relying on user-centric consumption business models, eliminating any material amount of friction from the user experience will expand usage, reduce costs, and increase revenues. What’s not to like?

The real limiting factor, in my view, will be the available bandwidth of software application developers who have both the domain expertise and the GPT skillset to roll out these adaptations at scale. This is where the consultancies come in. In that context, it is no accident that this past week Accenture, having laid off some 50,000 employees to adjust to a sluggish market for general consulting, announced it will be adding 80,000 employees to accelerate their AI practice. Every enterprise on the planet will be GPT-izing their user-facing applications wherever they require the kind of fiddly bureaucratic procedure-following that keeps their call centers operating 24/7 and their customers perpetually on hold.

But the consultancies have to be careful too. They should keep in mind a lesson from the client-server tornado.  That too was gated by a labor shortage, spurring companies like Viant and Scient to leap into the fray. They were market cap darlings as long as the wave of adoption was at its crest but then fell into disfavor once it passed. User-interface technologies must commoditize quickly if they are to have persistent impact, so this phase of the AI revolution is not an anchor for sustainable competitive advantage.

Still, a boom is a boom, and much more welcome than a bust, so drivers, start your engines!

That’s what I think. What do you think?

Follow Geoff on LinkedIn | Geoffrey Moore Main Mailing List | Infinite Staircase Mailing List

________________________________________________________________

Geoffrey Moore | Infinite Staircase Site | Geoffrey Moore Twitter | Infinite Staircase Twitter | Facebook | YouTube