Article: AI is helping us to patch more vulnerabilities than ever before!

Reality: AI is causing us to have more vulnerabilities to patch than ever before!

It would actually be interesting to see the distribution of patches over the age of the code it was applied to.

It’s both, really. AI is identifying vulns, writing exploits, and also generating new vulns at the same time.

The final stage will be a Windows that resembles more The Thing than an actual OS.

That’s why you use AI to find the vulnerabilities, not patch them

100% this. Their update stability has gotten even worse, I didn’t think that was possible.

I don’t want to be funny, but that’s not mutually exclusive.

Patch Tuesday was last week. Keep up.

Only ∞ more to go!

974 security flaws.

Srs. Stahp microslofting.

And broke 900 functions...

How many flaws did it create though?

I’m looking forward to the next update when they fix the 9,740 security flaws they created this week. It’s a flawless strategy, numbers always going up, just the way they want!

Going off the below… ᶜᵃʳʳʸ ᵗʰᵉ ᵒⁿᵉ‘‘‘‘ a lot

I would not want to work in IT these days; even debating if I really want to do cybersecurity or just stay a career college student and work a part time job.

I re-enteted IT after “retiring”.

Dang it. Stoopid mortgage.

I’m only going for it for taking care of family and well my body is starting to break down.

Look at it this way: all the fuckups mean job security.

That’s true, it’s just there’s so many… I kinda like to watch it burn a little bit; but I guess that increases the job value.

No wonder my work laptop was freaking out today.

Can’t wait to see what they broke next week.

Copy text from Excel now fails 50% of times.

I am willing to use Windows 98 over this shit.

Slop.

What’s the use when there are exploits they’re not aware of and therefore aren’t fixing? It’s a matter of time before an llm exploits or and destroys it all.

1. Given known empirical numbers of bugs per per one thousand lines of code (which is between a few and about 25) … and the size of Microsoft products, how many bugs do they probably have in total?

2. Given how “productive” LLMs are at adding lines of code, how many new bugs are there likely added per month?

1. Given known empirical numbers of bugs per per one thousand lines of code (which is between a few and about 25) … and the size of Microsoft products, how many bugs do they probably have in total?

Is this the number for human written code only or does it include boilerplate/ generated code too?

The big question is: How many things are they breaking? If they have to fix nearly a thousand bugs in a real hurry, they’ll probably break at least as many other things.

Story says llms are finding vulnerabilities, which would account for increased numbers.

Crazy how none of those fixes are uninstalling windows.