Analysts

Where we are going (next year)

If I didn't enjoy my job so much maybe I would become an analyst. It's so much fun to have a soap box to stand on from which you can state your opinions loudly with apparent authority (since you are an analyst, after all). But what to make of wildly different opinions from analysts? For example on CNBC we have the following (as reported in Futurism):

"The bottom line is, right now, everyone’s talking generative AI, Google, Amazon, Qualcomm, Meta," CCS Insight chief analyst Ben Wood told the broadcaster. "But the hype around generative AI in 2023 has just been so immense, that we think it’s overhyped, and there’s lots of obstacles that need to get through to bring it to market."

But then Gartner (which despite not being on CNBC is a bit more recognized as a leading analyst firm on Tech issues than CCS Insight) comes out with this (as reported on by Venturebeat):

“Generative AI is not just a technology or business trend — it is a profound shift in how humans and machines interact,” Gartner distinguished VP analyst Mary Mesaglio said in an opening keynote. “We are moving from what machines can do for us to what machines can be for us.”

Gartner goes on to advise clients that they should make AI adoption an enterprise priority over the next 12-24 months... So are there obstacles and is it overhyped? Or is it a profound shift and is there an urgent need to prioritize adoption?

I think that there is a confusion that we all fall into (including those of us that are analysts) when talking about a general-purpose technology such as foundation models. The desire is to think of it as one thing, but it is an enabler of many things. So talking about the technology being over-hyped ignores the nuance of all of the different ways the technology is being used. Are there topics/domains for which the application of foundation models is still some time away? Yes! And are there obstacles for adoption in some areas of the enterprise? Yes! But are there other areas where adoption is already rapidly happening, and we are moving from what machines can do for us to what machine can be for us? Also yes! The trick here is in expecting this to be a much more nuanced and evolving set of use cases.

Here is an analogy that I hope will help prepare you for what will happen over the next 12-24 months based on my own personal experience of the evolution of the "world wide web" in the mid 1990s:

1. Early on the idea of websites was exciting but limited. Most websites were static purveyors of information. In the early 1990s I was on a mailing list that was sent out every week and announced that a few new websites had been created. Regardless of the topic of the website, those of us receiving the email would excitedly visit just to see and understand how people were using this new Internet capability. At that point no one really knew what the "business" opportunity was.
2. Within a year of my starting to receive these emails they had gone from once a week, to every day, and then stopped. It was an impossible mission to report on new websites via email and newly created (and web based themselves) officious oracles like YAHOO were doing the work of organizing all of the websites into an easily navigable taxonomy of content. Now I could just see the new websites that appealed to me. Actual business models begin to emerge and people started talking about how someday people might actually purchase things online.
3. In 1996 I worked on a project whose stated goal was to have human editors write a review of "every website" and provide them with a quality rating. Lore had it that one of the senior executives of a tech publishing company had sat at a baseball game with his son and sketched out the plan, thinking that there were really only about 50,000 websites that needed to be reviewed. We built a digital sweatshop and ran almost 24 hours a day with an amazing (for the time) editorial production system. Already by the end of 1996 there were roughly 250,000 websites in the world. By 1998 there were 2.5 million. Today almost 1.8 billion (although, what's a website anymore?)

Detractors will say yes, and look at the terrible dot com bubble which burst in 2000 when everyone realized that the valuations of these companies had far outpaced their performance. And I am not denying that bubbles happen (tulips, etc.) But I don't believe that 2024 is the bubble year. More like the year when we give up on weekly emails as the way to keep up on how fast this is changing.