Discussion about this post

User's avatar
DinoNerd's avatar

Complexity is not the same as change.

The new dis-improved user interface that was just downloaded to my phone probably isn't more complex than the last one. It's just different. Users have to deal with increasing complexity, because of remembering dozens of different user interfaces to accomplish the same damn thing. But if they climbed into some vendor's walled garden, they could sharply reduce that number, particularly if they purchased new tools - that did nothing new or better that they needed - long before the older ones had become useless. It still wouldn't be as stable as a hand axe, or even a typewriter, but it'd be a bit easier to cope.

We do get complexity on the inside. I worked as a performance engineer for cell phones. Layers and layers of software that no one fully understands, all of it changing routinely in ways never clearly documented. My job was to understand the complexity better than most, with the aid of tools that let me look inside often proprietary boxes, and reduce the incidence of unwanted consequences - if I could.

This complexity supposedly provides benefits - but the main one seems to be faster time to market. We'll just reuse some pre-existing software that doesn't quite fit our use case, adding a layer of shims to make it fit. Except that software is reusing some other software, adding its own shim. Pretty soon it's shim-city, and both performance and reliability are terrible. But we got that new UI on user devices faster than ever, and the person in charge got a promotion, sparing him from the need to address his bugs. (It's usually a "him", in my experience.)

LLMs add additional complexity, with even less documentation. No one knows what they are doing, and that's by design. I'm kind of glad I retired before they became prevalent. But it's basically the same old ever-growing complexity.

It's important to note that the price of all this complexity is that things don't work very well. The nth-order interactions lead to lots of breakage. But that's also a cost of the constant churn - newly modified software will have fewer of its bugs revealed and repaired. No wonder it's easy to conflate them.

Overall, I'm less sanguine than you about the benefits of all complexity-inducing creativity, at its current scale, while recognizing that some of it has been useful to me.

Expand full comment
Alternative Lives R Available's avatar

I see it a little differently:

Humans are mostly fixers of problems rather that innovators from scratch. So a problem comes along, and we solve it by adding a layer of complexity to fix the original problem. But then this results in more, but smaller problems, so we fix each of those too, but of course that creates even more problems that need fixing, and yet another layer of complexity. So we aren't so much innovators as developers, not so much radical designers as iterative fixers.

The problem is that such an approach leads to layer upon layer of complexity, often getting further and further away from the original simple concept of the genuine innovator. The Austin Mini by Issigonis becomes the BMW so-called 'Mini' a 4x4 truck. The VW Campervan becomes the 20 ton motorhome with jacuzzi bath with a toy hauler on the back.

And an education to do useful work becomes a $250,000+ debt to join an intellectual arms race for esoteric theoretical qualifications and sheer numbers of academic research papers published, that are all-too-often useless to everyone still living in the real world!

And the mobile phone that makes phone calls and sends texts, becomes a supercomputer Smartphone, that knows exactly where you are every minute of every day, watches your every move and everyone around you, can measure your heartbeat and how much sleep you get, and sends all that information back to........ who exactly? Palantir? CIA?

And the funniest thing? YOU PAY PEOPLE FOR THAT!!!

George Orwell would never have dreamed that a government could get people to pay every month to get spied on!

😅😂🤣😂🤣

Expand full comment
10 more comments...

No posts