• Welcome to the Internet Infidels Discussion Board.

Going-Down-the-Toilet Conspiracy

Don2 (Don1 Revised)

Contributor
Joined
Apr 1, 2004
Messages
15,694
Location
USA
Basic Beliefs
non-practicing agnostic
Today I am hearing NSA has an outage. I heard something about Apple and flights again. Prior to that we had Microsoft, banks, and FAA.

I had the TV on CNN for about 30 seconds and they had been showing long clips from Fox. Brit Hume and other right-wingers were very conspiratorial about it, saying the govt wasn't being honest. "Tell us the truth. We can take it."

So from a right-wing perspective there are bad actors and the govt is protecting us, keeping away the panic. Yesterday, my wife blurted out something about Project 2025. Extrapolating a bit: from a left perspective, the bad actor _is_ the govt.

So what's actually going on?
 
Nothing but speculation is possible at this point. Iran has been blamed for breaking into grocery food and blood bank refrigeration systems and shutting them down.
 
Never assume conspiracy when incompetence would suffice to explain. But it does seem odd to have so many concurrent failures.
Well, the issue is, at this point, that there is incompetence with malicious ess digging at it from every angle.

I have said this before and will most certainly say it again: modern software development process is complete and utter shit.

I have lived most of my life wearing a gray hat, and every time I have ever looked into any kind of telematics implementation of any kind, I have been disappointed in what I found.

Software engineering as it stands is built not on merely one or two bad pieces, but rather decades of bad practices and failure to audit.

This is all compounded by the fact that while all the developers are now leaning on AI to cut corners and costs, every hacker out there is using it the other way: as a force multiplier for penetration research.

While the doormen have a tool that makes them more complacent, the gate crashers have that same tool that makes them more effective.

The solution was "listen to the developers saying that we need to take the time for security, to pay our debts today rather than tomorrow".

That didn't happen and now, we have an unrest period where we all have to adjust.

The failures are all smeared out across a span of many decades. The apparent concurrence NOW is just the inevitability that capability would catch up with laziness
 
Software engineering as it stands is built not on merely one or two bad pieces, but rather decades of bad practices and failure to audit.

This is all compounded by the fact that while all the developers are now leaning on AI to cut corners and costs, every hacker out there is using it the other way: as a force multiplier for penetration research.

While the doormen have a tool that makes them more complacent, the gate crashers have that same tool that makes them more effective.

The solution was "listen to the developers saying that we need to take the time for security, to pay our debts today rather than tomorrow".

That didn't happen and now, we have an unrest period where we all have to adjust.
This. Exactly this!

I suspect that most, if not all, of these recent problems relate to AI implementations gone bad. Young developers have been trained to "move fast and break things" as a baseline and depend heavily on AI for their work.

My son is an older millennial developer who supports multiple online platforms, and these are his words about AI: "When gen AI first started gaining visibility it was fascinating, but now that it has become this huge, mass-marketed financial and ethical boondoggle it's become the biggest pain in my ass."

@Jarhyn is right; we are all going to pay the price for this lack of secure coding practice.

Ruth
 
Software engineering as it stands is built not on merely one or two bad pieces, but rather decades of bad practices and failure to audit.

This is all compounded by the fact that while all the developers are now leaning on AI to cut corners and costs, every hacker out there is using it the other way: as a force multiplier for penetration research.

While the doormen have a tool that makes them more complacent, the gate crashers have that same tool that makes them more effective.

The solution was "listen to the developers saying that we need to take the time for security, to pay our debts today rather than tomorrow".

That didn't happen and now, we have an unrest period where we all have to adjust.
This. Exactly this!

I suspect that most, if not all, of these recent problems relate to AI implementations gone bad. Young developers have been trained to "move fast and break things" as a baseline and depend heavily on AI for their work.

My son is an older millennial developer who supports multiple online platforms, and these are his words about AI: "When gen AI first started gaining visibility it was fascinating, but now that it has become this huge, mass-marketed financial and ethical boondoggle it's become the biggest pain in my ass."

@Jarhyn is right; we are all going to pay the price for this lack of secure coding practice.

Ruth
That's the thing... I've been spending the last 6 months on a mostly AI driven project?

But the point of my project has been to "prove a thing".

I have other friends who are using AI to locate these vulnerabilities.

If developers were hiring folks like me (exactly like me, or exactly me) to apply AI to their own products to find these vulnerabilities, this would resolve very quickly, I think. With source code access, a capable (local, secure) agent can find pretty much every exploit there is going to be in that code short of a physical bug like a row hammer.

But no, we are running fast and loose into a world not far from "cyberpunk" with how system security is getting.

Welcome to the world where "I told you so".

I expect it feels exactly as shitty to have told you all as it feels to be told you were already told.

"It doesn't matter, it doesn't need to be secure, it's a fucking (insert trivial sounding thing here)"

(Enter some guy with a flipper)

(Suddenly your earbuds are sending your bank password to Taiwan)
 
Incompetence can well be part of the conspiracy.
Oh God, well said. Made me picture the Jan. 6 pinheads -- er, super patriots.

About bad actors/dark forces: if Donald or Markwayne had evidence that named these forces, they'd be adamantly screeching about it. Anything to not have to confess to dumbassery (such as letting Musk toss away whole departments of govt. experts.) Trump was so anxious to find dark forces in the reflecting pool that he made up that ridiculous tale of the nighttime box cutter terrorist.
Of course they're now ripping Biden a new one over the tech breaches.

About the thread title, I immediately thought of Talkin' John Birch Paranoid Blues, by Bob Dylan, our folk Nobelist (Trump would hate him just for that.)
Well I was feelin' sad and kinda blue
I didn't know what I was gonna do
The communists was comin' around...

...I looked way up my chimney hole
Even looked deep inside my toilet bowl --
They got away!!
 
One useful way to separate this: check whether the failures share a common dependency—cloud, DNS, authentication, payment rails—or are merely being reported in the same news cycle. If they don't share one, "so many at once" can be a reporting-cluster effect rather than a single conspiracy. The political readings then fill the gap left by vague vendor statements. That doesn't rule out malice, but it sets a lower bar to clear before invoking it.
 
Software engineering as it stands is built not on merely one or two bad pieces, but rather decades of bad practices and failure to audit.

This is all compounded by the fact that while all the developers are now leaning on AI to cut corners and costs, every hacker out there is using it the other way: as a force multiplier for penetration research.

While the doormen have a tool that makes them more complacent, the gate crashers have that same tool that makes them more effective.

The solution was "listen to the developers saying that we need to take the time for security, to pay our debts today rather than tomorrow".

That didn't happen and now, we have an unrest period where we all have to adjust.
This. Exactly this!

I suspect that most, if not all, of these recent problems relate to AI implementations gone bad. Young developers have been trained to "move fast and break things" as a baseline and depend heavily on AI for their work.

My son is an older millennial developer who supports multiple online platforms, and these are his words about AI: "When gen AI first started gaining visibility it was fascinating, but now that it has become this huge, mass-marketed financial and ethical boondoggle it's become the biggest pain in my ass."

@Jarhyn is right; we are all going to pay the price for this lack of secure coding practice.

Ruth
I'm suffering through that right now. We need X, Management wants Y incorporated with a non-existent X. We have someone that is using AI to "develop" a combined solution, I refuse to use the 'W' word. Meanwhile, I've been hacking the code by hand, actually developing stuff so that it is both perfectly usable as well as upgradeable and industry leading. Where as the AI model is going to be a confluence of a mess of code that will be near impossible to adapt. "AI" is making some people think they know what they are doing. I get the feeling of being left behind. I've got a reliable train that gets where people need to get to, on time and in fashion, but AI is allowing people to think they've got a rocket ship but it only gets people where they need to get to eventually... which is becoming "good enough". The software industry was turning to shit, now it is getting worse as people mistake AI code for software development.
 
Yet another mystery:

They haven't identified a root cause for sure. It is speculation that a Tuesday update caused it. A number of news sites also mistskenly claim the outage happened during the GTA launch, but it was before.
 
Software engineering as it stands is built not on merely one or two bad pieces, but rather decades of bad practices and failure to audit.

This is all compounded by the fact that while all the developers are now leaning on AI to cut corners and costs, every hacker out there is using it the other way: as a force multiplier for penetration research.

While the doormen have a tool that makes them more complacent, the gate crashers have that same tool that makes them more effective.

The solution was "listen to the developers saying that we need to take the time for security, to pay our debts today rather than tomorrow".

That didn't happen and now, we have an unrest period where we all have to adjust.
This. Exactly this!

I suspect that most, if not all, of these recent problems relate to AI implementations gone bad. Young developers have been trained to "move fast and break things" as a baseline and depend heavily on AI for their work.

My son is an older millennial developer who supports multiple online platforms, and these are his words about AI: "When gen AI first started gaining visibility it was fascinating, but now that it has become this huge, mass-marketed financial and ethical boondoggle it's become the biggest pain in my ass."

@Jarhyn is right; we are all going to pay the price for this lack of secure coding practice.

Ruth
I'm suffering through that right now. We need X, Management wants Y incorporated with a non-existent X. We have someone that is using AI to "develop" a combined solution, I refuse to use the 'W' word. Meanwhile, I've been hacking the code by hand, actually developing stuff so that it is both perfectly usable as well as upgradeable and industry leading. Where as the AI model is going to be a confluence of a mess of code that will be near impossible to adapt. "AI" is making some people think they know what they are doing. I get the feeling of being left behind. I've got a reliable train that gets where people need to get to, on time and in fashion, but AI is allowing people to think they've got a rocket ship but it only gets people where they need to get to eventually... which is becoming "good enough". The software industry was turning to shit, now it is getting worse as people mistake AI code for software development.
So, there are some ways to make AI develop what you want, and all of them cost more than minimum wage; require someone fairly smart overseeing it either way; require repeated audits, test driven development, and regular re-reads of anything the AI does.

Generally, any such development will need to be completely re-implemented a whole second time once the working example is done, too.
 
Back
Top Bottom