I go to this wonderful tech community thing called the Rocky Mountain AI Interest Group in Boulder, CO when I am not travelling for work. First of all, it is run fantastically by Dan Murray who has been at the forefront of creating other communities in the past. Credit RMAIIG and Dan for the above image.

I barely know anything about AI, I have realized as I have sunk… I mean dived deeper into it. Approaching AI with a beginners mind is simple. But I’ve created my first custom GPT- a project charter wizard. That is a not so fancy way of saying I added context and examples to GPT 4 around the way I think charters should be done and to encourage people to actually do them in a way that grounds projects in the manner intended.

Last night, April 10, was an expo sponsored by the group that included everything from LWIB(Look What I Built) to commercial presentations from start ups to Oracle. Thanks to the people that visited my LWIB demo for the patient support of my truly simple effort, during the expo portion of the evening.

I was struck by not just the variety of ideas but by the leveraging of solutions across platforms and tools and applications. These left me with a few impressions.

My first impression is that integration of LLMs across tools and platforms will be complete and this aspect won’t threaten jobs as much as it will allow people to finally do their jobs. One of the solution’s demos was implemented in Slack because it was easy, and then people don’t need yet another window i.e. the ChatGPT “context box.” The second was that you could actuate any number of tools through agents and through AWS or other numerable web services to answer, for example, a simple question. It was something like, “which of my users in AWS don’t have MFA turned on? Who do I need to go yell at?” All in Slack, the cloud architect agent activated a series of steps that you could see performed and at the end of which the AI Agent responded “Good news, you don’t need to go yell at anyone; all you current AWS users have MFA.”

At least one security professional I know would cry with joy rather than be threatened by this use of AI.

Another large commercial entity demo’d their early efforts at an AI customer service tool that will track the “mood” of the customer and the human service agent during a call. This has all kind of implications but they admittedly were trying to improve the experience of the customer and the well being of their staff; though applications of this type will be used nefariously I imagine as well.

While some of the larger concerns will build their own LLMs, I think over time, LLMs have fighting chance to be more like a new language than something one particular company “owns” despite the fact that I am currently willing to shell out 20 bucks a month to play around on ChatGPT 4.

I think the real beauty may be in the potential of use cases to spread across all platforms and all applications so that by trying to own the innovation of genAI, you will exclude yourself from the community rather than control it. I don’t know, I am definitely a beginner and whether you want it or not, something is coming. As usual, the challenge will be in how we prepare our people to manage the change.