[HN Gopher] The Cybernetic Teammate
       ___________________________________________________________________
        
       The Cybernetic Teammate
        
       Author : tobr
       Score  : 26 points
       Date   : 2025-03-22 12:52 UTC (10 hours ago)
        
 (HTM) web link (www.oneusefulthing.org)
 (TXT) w3m dump (www.oneusefulthing.org)
        
       | torginus wrote:
       | 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 :)
        
       | bwestergard wrote:
       | 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.
        
         | peterldowns wrote:
         | Yeah, but to be honest it seems like a reasonable eval. Having
         | to actually launch these products into the market would
         | introduce a lot of confounders!
        
       | peterldowns wrote:
       | 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.
        
         | mentalgear wrote:
         | Exactly. I think we need to start talking less about "AI" as
         | one overall term, and instead state the actual components that
         | bring high value (semantic search) vs those of often low or
         | unreliable value (code generation).
        
       | codr7 wrote:
       | 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.
        
       | Mistletoe wrote:
       | 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.
        
       | bob1029 wrote:
       | 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).
        
       ___________________________________________________________________
       (page generated 2025-03-22 23:01 UTC)