Spreadsheets and Disposable Software
A trend I am seeing, and experiencing, is an increase in disposable software. Software that does a little thing, for a little while and then gets deleted. No long term plan for maintenance or support.
Like everything, there's positives and negatives to the change. Overall I see more software going the way of the spreadsheet: useful, accessible but ultimately disposable.
On the positive side of things I am a big supporter of any change that helps people make more things. That applies to helping new people make things they couldn't make before, and making it easier to make things you'd otherwise not have time to make. I've started to see a lot of projects that would otherwise never see the light of day. They're not all destined to be billion dollar SaaS apps but most of them aren't trying to.
People, like with spreadsheets, are building little things for themselves or for a small group of people.
I find amongst software people, those that were making things before, this shift has mixed feelings. The norm for a long time was to, whenever possible, avoid writing software over and over. We have millions of tiny little libraries and frameworks designed to ensure you're not reinventing things over and over.
There are arguments to be made that this had become a bit extreme and we were importing trees of dependencies to avoid writing basic functions but lets ignore that for now.
Where the negatives start to creep in are around the whole idea of disposal. The term itself is typically used in a negative light. Things created or destroyed without care or concern. This argument seems to be more popular amongst those who feel the power or water costs of AI don't provide an ROI.
Data centers certainly use their fair share of resources, both in training and in inference, but it's hard to think the right path here is shutting things down. There are endless other ways that folks could be spending their time but making things, for them or for others seems like a generally positive outlet.
There's no doubt going to be an initial burst of 'waste' as people try thing, make mistakes, build their 'fart apps' and then get bored. If we pay that price to give thousands or millions of people the tools to build things then maybe it's worth while. If nothing else I think more use will help more people understand these tools more, know their limits and dispel the hype.
After all, spreadsheets had their hype cycle too.