>We are trying to fix probability with more probability. That is a losing game.
Technically not, we just don't have it high enough
You're doing exactly what you said you wouldn't though. Betting that network requests are more reliable than an LLM: fixing probability with more probability.
Not saying anything about the code - I didn't look at it - but just wanted to highlight the hypocritical statements which could be fixed.
If there's a legit, measurable performance or data integrity problem, start with that. If most of your production bugs come from a specific module or service, document it.
If it is only technical debt that is hard to understand or maintain, but otherwise works, you're going to have a tougher time of building a case unless you build a second, better version and show the differences. But you could collect other opinions and present that.
Ultimately you have to convince them to spend the time (aka money) on it and do it without making things worse and that is easiest to do with metrics instead of opinions
>I am not sure what has happened over the decades regarding actually being proud of the work you produce.
Because there's still people doing less work than you do for a bigger paycheck
Because you'd get fired or laid off for someone working for 1/2 to 1/4th of your pay
Because they make you jump through multiple rounds of interviews and technical tests while people above you have a far less barrier to being hired
Because someone stole credit for your work
Because you'd get re-hired and find a mountain of shit code from a company that off shored their dev team
Because companies stopped giving significant raises that didn't keep up with major inflation in the past few years, while your work might have gotten them many multiples more of profits
And why? Not for any good reason, no. Just because they can. Your landlord isn't content when you pay $2,500 per month for an apartment, no. They need $2,600. $5 isn't enough for a dozen eggs, it needs to be $6. And what if we slapped 10-200% tariffs on random things, depending on the day? Wouldn't that be neat?
The collective delusion it requires to not see what the problem is is astounding. It's actually quite depressing, because it makes me think we're never going to meaningfully solve this problem. Maybe companies have to start executing employees or something, I don't know. Maybe then people will be bold and decide to re-organize society.
>Nobody makes you use YAML and Docker and VS Code or whatever your beef is.
Not VS Code, but maybe YAML and Docker if your company is trying to align what tools it uses. C# places might still force you to use Visual Studio proper. Everyone says use the right tool for the job, but for bog standard CRUD web development, we do have a shitload of tools that work and there's multiple ways to get to a fast, working product.
I still chuckle that my laptop is 3 times as fast as the cloud thing that serves our CRUD app and we pay many more times for it, but also knowing full well I do not ever want to be RDP'ing into a production box again and pouring through IIS or Windows logs.
What I definitely do see is a degradation in making choices about when to adopt a more complicated technology because of the incentives of the hiring market.
People have loudly beaten the drum to keep your skills up to date and so people choose stuff that's popular because it benefits them personally, even when the product doesn't need it. This in turn leads companies to only select people who know that stack, and the more that companies do that, the more people make technical choices to get them the best job that they can handle.
We absolutely, very much 100% see that happening now with LLM AI if you ever needed a bigger piece of proof. Pretty much everything that is happening now has just been a louder example of every bad practice since the run up to the dotcom bust.
Because of that, I'd frankly never suggest running on-prem or building a local-only app unless there was a much bigger reason (legal, security, whatever) especially if the other products in the company have chosen the cloud.
Why? Because convincing the next job that that would have been the right choice is too hard.
Edit: and to someone else's point, I made the choice to be in the Microsoft/Azure/Windows hell hole but digging myself out and moving to something else is practically working a second full-time job and holding 2 ecosystems in my head at once
Technically not, we just don't have it high enough
You're doing exactly what you said you wouldn't though. Betting that network requests are more reliable than an LLM: fixing probability with more probability.
Not saying anything about the code - I didn't look at it - but just wanted to highlight the hypocritical statements which could be fixed.
reply