I've been maintaining web applications since the early 2000s, back then if your server was ok, your site was ok, period.
Now if any of {AWS, github CI, cloudflare, google cloud, claude, microsoft, a dozen auth providers, etc} is down, your application is surely going down.
3. complex products are maintained by few people because of economy, hence proper safeguards are not being put in place (either on-time or at all)
4. LLMs enabled us to find more issues in our products, hence more patches are coming, but since business is not going to wait for proper fix, some companies are adding "temporary" fixes (forever)
and many more reasons/causes, but has it increased, maybe, maybe not, maybe it was always buggy we just didn't feel it, because maybe airports weren't heavily relying on a single communication channel and so on
We don't actually apply the principles of Engineering[0] when we build things that turn out to be critical to the stability of civilization. We have millions of people calling themselves engineers who just aren't. We've normalized ignoring this.
I mean, who in their right mind routes safety critical notifications through email?[1] Email is a best effort system, not a reliable one. Even this non-engineer understands that difference.
Also, our operating systems are unfit for purpose. See also, any of the dozens, maybe hundreds of rants I've spewed about this previously.[2]
Aviation is lauded as an example of engineering principles being applied to achieve extreme safety and reliability. Yet, even they failed to foresee the recent outage in the US, they failed to protect against the SPOF.
It’s true that IT workers are generally less qualified but I think we should also recognise that the realm of software is harder. Yes, harder. Physics is predictable, it works the same way everywhere in the world. Computers are completely unpredictable, a software flaw in one area can have far reaching effects. It’s not like a random switch failing in the cockpit of one aircraft can take down the whole fleet. Software is more akin to psychology than to engineering.
A combination of AI slop, LLMs multiplying the rate of bad actors’ ability to discover security defects, and in some cases failure to invest in or put even borderline-competent people in charge of key systems.
I might add that growing disaffection among those of us working in software and IT — due to layoffs, AI culture, other impacts of late-stage capitalism on employment, and just the world going to hell — isn’t exactly helping either.
Dependency between products has increased a lot.
I've been maintaining web applications since the early 2000s, back then if your server was ok, your site was ok, period.
Now if any of {AWS, github CI, cloudflare, google cloud, claude, microsoft, a dozen auth providers, etc} is down, your application is surely going down.
difficult to measure this.
1. dependency on products increased
2. products became complex
3. complex products are maintained by few people because of economy, hence proper safeguards are not being put in place (either on-time or at all)
4. LLMs enabled us to find more issues in our products, hence more patches are coming, but since business is not going to wait for proper fix, some companies are adding "temporary" fixes (forever)
and many more reasons/causes, but has it increased, maybe, maybe not, maybe it was always buggy we just didn't feel it, because maybe airports weren't heavily relying on a single communication channel and so on
We don't actually apply the principles of Engineering[0] when we build things that turn out to be critical to the stability of civilization. We have millions of people calling themselves engineers who just aren't. We've normalized ignoring this.
I mean, who in their right mind routes safety critical notifications through email?[1] Email is a best effort system, not a reliable one. Even this non-engineer understands that difference.
Also, our operating systems are unfit for purpose. See also, any of the dozens, maybe hundreds of rants I've spewed about this previously.[2]
[0] https://www.in.gov/doe/files/cf-eng-principles-engineering-3...
[1] https://news.ycombinator.com/item?id=50509411
[2] https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...
Aviation is lauded as an example of engineering principles being applied to achieve extreme safety and reliability. Yet, even they failed to foresee the recent outage in the US, they failed to protect against the SPOF.
It’s true that IT workers are generally less qualified but I think we should also recognise that the realm of software is harder. Yes, harder. Physics is predictable, it works the same way everywhere in the world. Computers are completely unpredictable, a software flaw in one area can have far reaching effects. It’s not like a random switch failing in the cockpit of one aircraft can take down the whole fleet. Software is more akin to psychology than to engineering.
A combination of AI slop, LLMs multiplying the rate of bad actors’ ability to discover security defects, and in some cases failure to invest in or put even borderline-competent people in charge of key systems.
I might add that growing disaffection among those of us working in software and IT — due to layoffs, AI culture, other impacts of late-stage capitalism on employment, and just the world going to hell — isn’t exactly helping either.