Comment by maccard
3 days ago
Two points. First of all, yes I would. There's a reason we use existing tools and don't have every junior programmer write a TGA viewer when they want to view an image that is sent to them.
Secondly, I used time as a proxy for cost. A single junior can "only" spend their own 15 minutes, but an agent can farm out to N subagents to do the work in 15 minutes, and spend $3000 in tokens building something. As you said, scale matters, so if they do this once a day for a month, it costs the same as paying 12 juniors to avoid asking a single question.
There used to be a time when you had to write in assembly and count bytes to fit your code into the vanishingly small ROM you had available.
Nowadays, you can just write in python with no care for the hundreds of thousands of cycles and megabytes of memory you are wasting.
All indications today point towards the same happening with the cost of intelligence.
The principle that you should not reinvent the wheel (unless there's a good reason to) has not changed.
Code is a liability. Even if it is a small cost (in terms of time and money) at the time when code is generated, it can be a huge burden later, especially when you think about security vulnerabilities.
If this happened during my work, I would be very mad -- I would absolutely reuse an existing tool instead of expecting Claude Code to come up with its own half baked, bug ridden implementation of a common tool. Any (capable) human developer would have stopped and discussed with the team how to proceed.
In fact, I cannot tell how many times similar situations have happened where LLMs "made a decision" without consulting with me.
A very many of us do care for the cost.
There’s also a very significant difference between “choose python” and “completely ignore all existing material and reinvent visualisation”.
Also, by all indications the costs of LLMs are _rising_ not falling as the tech progresses.