Is it worth the hype?

You must be a blessed soul if you had not heard about generative AI (GenAI) and all the buzz around it. Meaning, you have been living under a rock or in a cave!

The entire tech world is dancing on chatbots, mid-journeys, and nothing less than artificial intelligence. If you don’t get it, at least fake it. That’s what many of us do.

The last three years (since 2021) saw US$50B investment in GenAI. The year 2023 saw 426 deals closing US$22B.

Amidst all this FOMO, it's worth looking at it through a product lens.

Does GenAI truly justify the expectations and enthusiasm surrounding it? Is it really the game-changing innovation it's made out to be? Or is this just another overhyped tech that will ultimately underwhelm? 

Now don't get me wrong, the capabilities of these systems are truly impressive, would have seemed like science fiction only a few years ago.

So, is it desirable?

Absolutely. With the power to create engaging, informative, and coherent content in the form of text, image, or video, GenAI cannot only help create new experiences in a product but also save hours, if not weeks and months, worth of manual efforts. From Amazon’s summarised customer reviews to WhatsApp’s AMA AI, the power is being felt by billions around the world.

Is it feasible?

It depends. Creating AI large language models from ground up is no child’s play. It requires $$$. Consider the GenAI technology stack. To oversimplify, assume that it has three layers. (Refer to the book Reimagined by Shyvee Shi et. al.)

The bottommost layer is the foundation of GenAI. It contains the Holy Grail: the computation, the cloud infrastructure, and the large language models. No wonder this is the playing field of big tech companies like Nvidia, OpenAI, Microsoft, Google, Amazon, and the likes.

The second layer above the foundation is the layer of tools. From fine tuning of models to data preparation and management and application development frameworks, this layer helps in realising the GenAI transformation vision.

The topmost layer is the application layer where the products integrate GenAI technology to create new product experiences or enhance productivity, or improve process efficiency or a combination of above.

It is easier and economical to play at the topmost layer building applications. No wonder there are 70,000+ companies working to build apps.

But the key question is: who actually makes money?

OpenAI raised US$11.3B since 2015. Reports suggest that OpenAI burns US$700,000 per day while its annual revenue has climbed over US$3B. Along with LLM providers, the cloud infrastructure companies also benefit. In essence, only a handful of big companies are making money now.

And what happened to these 70,000+ companies?

Nothing yet. It is not even clear whether they ever will.

According to a recent report by The Ken, this scenario is no different even for the frontrunners like Accenture and Salesforce despite Salesforce’s claims that introduction of more than 50 AI tools among its workforce has saved 50,000 hours of work in 3 months, that is someone's 24 years of work time compressed to 3 months. The interviews with venture capitalists also show little promise for more investment.

The reasons being

• there are no significant applications today even after a good two years.

• (according to David Cahn of Sequoia) "there is a big gap between the revenue expectations implied by the AI infrastructure build-out, and actual revenue growth in the AI ecosystem, which is also a proxy for end-user value".

There is a lot of scepticism in air despite all the initial excitement. Is this another blockchain in the making? And, in 2-3 years time, we're left wondering "whatever happened to that whole generative AI thing?"

The future may be bright, but the path there is anything but certain.