The Cybernetic Teammate

the trouble with this one is that only observes one day workshop settings

Interesting, I wonder how much of the value of team+AI is the sense of constant momentum and progress. In my experience the utility of vibecoding is overrated BUT the utility of having a constant, enthusiastic, broadly-knowledgable assistant is underrated. I've been moving a lot faster with the help of AI but it's almost entirely acting as a "super rubber duck" that helps me explore ideas or overcome roadblocks simply by being there and mildly helpful. This compounds when working on a team; because I'm making progress, other people are excited and try to make progress, which makes me excited, etc.

What I'm very much missing so far is research into longer term consequences for learning. I get that we could potentially gain something from using AI, I would like to know what we're losing, because it's obvious to me that it's a trade of some sort.

This article hints on it in the form of reduced frustration. Frustration and tearing your hair out seem to be tightly linked to learning from my experience. To the point where I wonder if it's possible to even get the same results without the experience.

Their primary measures of group and individual performance are ratings of "quality", "novelty", etc. by technical and business experts on a ten point scale.

So people working with LLM bots can improve ratings of their producy R&D work product by professional peers. That's an interesting finding.

Of course professionals can be wrong in their assessment of how market actors will value their peers' work.

I am finding there is value in having a tight loop between human & machine (LLM). Trying to design LLM "agents" that take a single human command and then run off into the dark forest for a billion tokens is not the right path.

We should be designing "agents" that aggressively involve the user at every possible opportunity. In all contexts of use, you should have something like an AskUserQuestion() function that can be called in parallel each turn. Then, the user serves the AI in answering these questions before they get what they originally asked for. Effectively, this makes the user do a big information loop that could have otherwise been skipped in their initial prompt, but guiding the user when they are not 100% focused can help a lot.

I thought this was what we were aiming for this entire time, but it seems like the goldilocks zone is a bit more complicated than is preferred with fly-by-night AI vendors. To build this requires deep understanding of the "happy paths" through the business. How could you expose functions to the LLM if you don't know what they are or are experiencing disagreement regarding the process around them as a team?

The ultimate conclusion I wind up with is that very few teams & products are organized enough to plug directly into an LLM agentic cybernetic whatever you want to call it. If you cant get the team & customers all on the same page about what the product is, I don't know how the AI is going to help. Once you do get everything standardized enough, you may find you don't care about the purported benefits of AI anymore (you've already realized them).

Next time I go to the market, I will ask 0.3 standard deviations worth of potatoes as that's apparently a concrete quantity now :)

What do 776 professionals at Procter and Gamble do? I'm just wondering how working with AI fits in with what they must do all day.

Besides discontinue my favorite shampoo of all time, Herbal Essences Blue Ginger & Micellar Water.