i always found it to be easier to write code myself than to direct a junior developer.
the level of teaching involved would always mean the overall velocity of work slowed down.
some people say you can throw them the drudge work but i find that if you're doing coding right (e.g. you dont let your code base degenerate into a mess of boilerplate), there is barely any drudge work to do.
You're missing the real goal of directing a Junior, which is you're teaching them to be a team player, Junior devs will surpass your expectations, the rate at which they goof or are about to goof should decrease over time the more you mentor them. If you do it right, you not have a strong ally and coder under your belt, or would you rather someone else teach them their bad habits?
Im not missing the point that they need to be trained. I was quite explicit about that being necessary. Nor am I denying the benefits of them being trained.
I just cant pretend that I'll get extra productivity while I'm training them.
Certain professions lend themselves well to a apprentice/master framework because the apprentice doing the drudge work that requires less skill frees up time. This can increase overall productivity while the apprentice is trained. Development isnt really like that though.
Perhaps but at least when you are directing a junior developer, even if badly, you'll eventually get a non-junior developer on the other side. With an AI agent, you'll get ... what?
With current models, you're right, there will be nothing to show for the effort except the code itself.
I suspect that will change sooner or later. Models will be cultivated over time the way we cultivate full-time employees now, with an acquired awareness of what they're building, new skills picked up in the process, and insight into how the larger system works.
There's a long-running instance of the model on the provider that's allocated to my organisation? Or are you thinking more of a server-side memory system, similar to the (currently very fallible) ones like Honcho and Mem0?
What happens when my org stops paying that provider? Do I get to take the now senior agent with me to the next provider? Does the provider have to delete it (and all that learning is lost forever)? Does that now become a free agent that can be hired by the next organisation like an employee (one that probably doesn't know how to keep industry secrets to itself)?
Beats me, I haven't thought it through any further than that. But it seems obvious that the combination of a vast-but-static bag of weights and a tiny, ephemeral context has limited value, and will eventually be looked back upon as a transitional form.
Karpathy's notion of a self-maintained wiki may suggest a direction for work in that area. As he points out ( https://gist.github.com/karpathy/442a6bf555914893e9891c11519... ) this is basically the latest of several attempts at implementing Vannevar Bush's "Memex" concept. I think it will find application in some form at organization-wide scope.
If the trend towards centralization holds, then yes, who gets to own and maintain the state in a stateful AI model is indeed going to be a big question. I hope it doesn't hold.
the level of teaching involved would always mean the overall velocity of work slowed down.
some people say you can throw them the drudge work but i find that if you're doing coding right (e.g. you dont let your code base degenerate into a mess of boilerplate), there is barely any drudge work to do.