Most software today is built to ship fast.
Not to survive.
I’ve worked on systems expected to outlive their creators.
We optimized for:
- constraints
- clarity
- failure being visible
Not speed.
This is the Greenfield vs Amnesia problem.
1/2
Lovable has a mass data breach affecting every project created before november 2025.
I made a lovable account today and was able to access another users source code, database credentials, AI chat histories, and customer data are all readable by any free account.
nvidia, microsoft, uber, and spotify employees all have accounts. the bug was reported 48 days ago. its not fixed. They marked it as duplicate and left it open.
https://t.co/yYvJnAkST5
Systems are not lost to failure, but to forgetting. When decisions cannot be undone, change becomes collapse, and rebuilding becomes inevitable.
What’s one constraint you still refuse to drop?
I read every reply.
Most software today is built to ship fast.
Not to survive.
I’ve worked on systems expected to outlive their creators.
We optimized for:
- constraints
- clarity
- failure being visible
Not speed.
This is the Greenfield vs Amnesia problem.
1/2
@vctrstrm I agree.
My gripe is when an experiences dev break constraints, he has a mental model that allows him the insight to do it. He knows what, why and when.
In a Garry's List kind of situation, with 30k LOC/day, do you reckon anyone knows what constraints got broken?
We built software that lasted decades because we obsessed over constraints, failure modes, and entropy.
AI just made 'good enough' scale to millions of lines overnight.
What’s one constraint you refuse to compromise on, no matter what?
Drop it below. I’m reading every reply. #SystemsProgramming
@ThePrimeagen This isn’t intelligence.
It’s anthropomorphism at scale.
If you describe systems like organisms, you stop reasoning about constraints and start projecting intent.
That’s how you lose control while thinking you’re gaining it.
@ThePrimeagen My world-view was forged in the dark, sweaty alleys on IRC, old Reddit and an open internet.
I can survive Mythos. Too effing normie.
Or is it?
@albinowax HTTP/1 sticking around is basically institutionalized attack surface.
Every downgrade path is a negotiation with ambiguity.
AI didn’t create new classes of bugs — it just made it trivial to explore the ones we kept ignoring.
@ApplyWiseAi Can't one of you guys vibe code an app that automates the whole reply-dude thing?
Like find niche - > reply to large creator - > boom boom - > lots of followers?
@BrendanEich@NivalCat@BraveNightly I call this the Greenfield vs Amnesia debate. It is a knowledge decay of all the little learnings.
Software was supposed to last. Accessibility to cheap AI is breaking the fundamental intentions of LONG LASTING software for crying out loud.
https://t.co/yYvJnAkST5