HN Explorer

Ask HN: I think the AI terminology needs a reset

1 points by ddaniel10 ↗ HN
A friend asked me what a personal AI agent does. I started explaining, then I realized I was describing three different things.

I think we need better order and terms in this world.

So personal ai agent can be 3 different things..

Delegate agent: It's own inbox, number, and accounts. Like: "Hi, i'm Peter the agent of Ana, I'd like a table for four"

Twin agent: My accounts, but it says an agent is speaking. Like: "Hi, I'm Ana, through her agent, I'd like a table for four"

Ghost agent: My accounts, without mentioning AI. Like: "Hi, I'm Ana, I'd like a table for four"

Also the term "personal ai agent", I'm not sure if its the right term, like i think what makes this agent special that its "always on" right? not that its personal.. theoretically it can work for entire team..

I would like to know what you think and if you have better ideas

2 comments

[ 0.20 ms ] story [ 14.9 ms ] thread
josh-paul
Personally the term agent is too overloaded. An agent is the loop. That is it, this loop handles context management and continuity of the calls to the LLM. The LLM is a STATELESS REST API call. The harness is the tooling and ways to interact in with the local system for the model. When you start looking at the boundaries this presents, you get a more accurate reprensentation of what is happening. It also decouples the agent loop from the specific model. Model's become more interchangeable reasoning service calls.