Our IT department and leadership are more schizophrenic than usual around AI. Top leadership wants it bad, pushing big initiatives. Risk management layer, predictably, is more cautious - requiring analysis and approvals and so forth - this is driven into IT with things like redirects from open internet AI services to internally hosted alternatives (sensible), but the internal services aren’t completely up-to-speed with tools like Cursor. So, I have approval for Cursor, and IT is helping me make it work around their filters, but then again - once in a while the Cursor software magically uninstalls itself overnight. Luckily, the work in progress is still there and when I re-install Cursor it picks up where it left off.

We clearly have conflicting interests at work behind our company computers.

If you ask an AI agent (like the one on Google search) how to disable the prompt dialogue it will usually tell you something that works. This is actually nothing new, back in Windows ME days I spent 8 hours with search engines figuring out how to de-feature Windows ME until it worked like Windows 98SE, that was actually a pretty good computer/OS for the day - after I turned off all the ME garbage.

Famous criticism from the testing department “we don’t write the buggy code…”

I’ve had some success taking the buggy output of Claude (which Claude gets stuck in a loop trying to fix), fixing it with GoogleAI, then feeding the result back to Claude which can then follow the working patterns and make working extensions…

Still, anything past “microservices” seems to be inadviseable to trust to any current AI.

I had an Atari 800, and the manual for it was pretty complete.

At office in 1991 I started programming for IBM PCs in DOS, that was a big step into the void.

had comprehensive manuals

You must not have coded for DOS.

In the late 80s there were no man pages, we had a 50 page paper manual - and it was mostly useless.

I do most of my cursing during preliminary test…

Nah, the good ol’ days were when websites provided value to their users. The only competitively sustainable business model is the reverse: where users provide value to the website owners.

You know what you didn’t mention? Security. Privacy. Authenticity.

a clean and functional site

But the specs no longer call for a clean, functional site. Today’s “professional” web development specs call for user tracking, customized content per user per platform per locality the page is being served to. Then there’s the backend dashboard services showing the overlords user behavior patterns and how they change based on tweaks made to their ui, I wouldn’t be surprised if the algorithms auto-tune presentation to optimize behavior.

It is always different. I was born because times before I was born were different, my parents’ friends were being drafted and sent to Vietnam to die, or worse: come back mentally or physically screwed up. Apparently, being a new parent was just enough consideration to keep a college enrolled man from being taken - so they did that and here I am.

If it’s any consolation, there was a nuclear sized crater of a job market for all engineers (software and otherwise) after Reagan’s 8 were up. The .com aftershocks were pretty huge, and the 2008 housing crisis hit everything really hard too. Then in 2012 I got laid off due to the end of our Afghanistan debacle, then there was that pandemic thingy…

So, yeah, the current foolishness is going to make a hell of a mess, but there has always been a huge mess either cleaning up, or coming soon for the past 35 years, and longer I’m sure, those are just the ones I’ve been hit by.

Does no-one remember the image-magick anymore?

I think that’s all about who’s doing the polling and how they are doing it: https://abcnews.go.com/538/trump-approval-polling-average-works/story?id=117999010

My question: was the whole Canada / Mexico thing just a smoke screen for the newly increased China tariffs?