The fact that you think this is a seriously practical suggestion shows that, very fortunately for you, you have no idea what goes in a household like this. I say this with sincerity, it's a good thing that you don't realize the problem with this suggestion.
Agree - if my 10yo son was in charge of getting his own food for school from our well stocked fridge, I know he'd struggle to get any kind of balance.
I know from friends in education who see kids come to school who have had to try to find food to bring in. I live in a relatively well off area but kids still come to school with stuff like just dry bread or a single cereal bar as there just isn't anything to hand for them to bring in for school - looking after their kids welfare isn't a given for a sad and sizeable minority.
You know there are countless parents that abuse their kids, physically, emotionally, sexually. They rape their kids and strangle them until they pass out. They hit them in places where bruises aren't visible. They break bones and threaten the kids with death if they dare tell anyone or show their injury or pain.
Why do we suddenly need to especially think about the other side of this because tax money is involved? Now we need to care?
Yes. I wonder if the path forward will be to create systems of agents that work as a team, with an "architect" or "technical lead" AI directing the work of more specialized execution AIs. This could alleviate the issue of context pollution as the technical lead doesn't have to hold all of the context when working on a small problem, and vice versa.
This is kind of what the modes in roo code do now. I'm having great success with them and having them as a default just rolled out a couple days ago.
There are a default set of modes (orchestrator, code, architect, debug, and ask) and you can create your own custom ones (or have roo do it for you, which is kind of a fun meta play).
Orchestrator basically consults the others and uses them when appropriate, feeding in a sensible amount of task definition and context into the sub task. You can use different LLMs for different modes as well (I like Gemini 2.5 Pro for most of the thinking style ones and gpt o4-mini for the coding).
I've done some reasonably complicated things and haven't really had an orchestrator task creep past ~400k tokens before I was finished and able to start a new task.
There are some people out there who do really cool stuff with memory banks (basically logging and progress tracking), but I haven't played a ton with that yet.
It's your comment that seems to add the least, because the majority agree with the OP. The point is that ergonomics is not as good, of course there are contrary opinions. The existence of a contrary and minority opinion doesn't detract from the point.
It's like I've always said about self driving cars. Self driving cars always seem around 5 years away, until suddenly they're 6 months away. There's no in between. We're seeing that with Waymo.
It doesn't really make explain it because then you'd expect lots of nonsensical lines trying to make a sentence that fits with the theme and rhymes at the same time.
With non-unique units paladins can only be countered by upgraded halberdier, heavy camels or a huge mass of arbalesters. Monks counter them in small numbers but in late game when they're massed they're almost unstoppable unless you have halb/camel in equal numbers.
Well even arbalesters get slaughtered by Paladins, unless we are able to push them behind a wall or a corner with only a very tiny attack area for the Paladins.
I think even in open ground if you have a big enough group you can still trade well. Ethiopian arbs fire 25% faster. If you micro the arbs to focus on individual paladinsthey can one shot paladins and their attack surface area increases with the sqrt() of their numbers. For sure it's a lot more work than spamming halb.
reply