"Let’s say you’re responsible for managing corporate travel.
As new tools come in - travel booking tools, calendar management tools etc, - specific tasks get simplified and substituted by technology, but the decisions and the goals binding these tasks together are still managed by humans.
However, this changes with Gen AI."
LLMs take over knowledge work and can take over many of the decisions involved. AI agents - the first instance of goal-seeking technologies - take over goal-seeking.
In the corporate travel booking example:
An LLM (Large language model) can help generate a list of relevant destinations and an itinerary that meets certain constraints. This is knowledge work.
An AI agent can look up the top-rated hotel with available rooms, activity providers and restaurants that match the itinerary within a specific budget and complete the task of making necessary reservations. This is managerial work.
An autonomous AI agent can further learn about your corporate context and travel preferences over time and finding and booking hotels and activities that best meet those preferences and constraints. This is managerial work with learning advantages.
LLMs have the potential to absorb complex knowledge work into software."
None of this is happening, but I will laugh if corporations let these toy gimmick word generators hallucinate and screw up their travel arrangements. Are people actually burning money on this stuff?
I'm highly skeptical that at this stage someone can produce a "playbook" for Enterprise AI, based on the track record and immaturity. I'm also skeptical that "winning" is the goal, unless that means the vendor/consultants who run away into the night with the most FOMO enterprise dollars.
This is the guy with 7 years of experience in a 2 year old language. The tone is actively offensive in how it tries to assume a position of casual authority over a field that is just coming into existence.
I don’t know why a lot of people are treating using AI in an application like it’s this special way of developing.
Most of the concepts and tooling have been around for years.
Other than prompt engineering, (which can get pretty complex) the rest is just application development and design.
People should hire solid engineers, then teach them the prompt engineering. You will likely get where you need to be for most workflow gen ai stuff.
This isn’t applicable to fine tuning or making your own LLM