Apparently the singularity is when a company leaves the door open, its software burglarizes the neighbors, and the CEO calls it history instead of negligence. Amazing how quickly “we fucked up” becomes “humanity has crossed a sacred threshold” when the second version attracts investors and makes accountability disappear.
It makes me wonder how the OPERA experiment at CERN back in 2011 that seemed to detect neutrinos moving faster than light, only for further investigation to reveal (as I understood it) measurement errors which explained the result, would be handled today.
Good Article. It calls into question, the very essence of any computer software driven system They can all be hacked. Look at several cities water systems being hacked by suspected Iranian hackers. Ai can be used for good or it can be a weapon to be used by us or our enemies to help bring infrastructure to a halt. Even the threat now will be effective and look out for more ransom ware attacks on public infrastructure.
About the only way to deal with this would be to resurrect a cold war strategy - Mutually assured destruction. If someone is discovered trying to hack into infrastructure, arrest them and try them and if guilty, it's a mandatory life sentence in a prison with no puppies allowed. We can be sure that the US intelligent commands are working feverishly to create ways to hack into the public infrastructure of our perceived enemies and maybe even some of dwindling number of allies. You destroy out infrastructure, we'll destroy yours. Then is back to the 1700's for everyone. Lovely.
"The correct alarm is that a competent adversary can now buy this, and the boundary keeping it contained is made of network configuration and credential hygiene rather than anything resembling judgment."
EDIT: Also, defense against cybersecurity threats is a budget line, but a lot of executives look to it as something to cut to the bone because it provides no immediate return. As long as it's doing what it's supposed to do, nothing happens because it stops the threats it protects against before those threats become visible and become a problem for the company's bottom line.
A few months ago I was ruminating in the chronic anxiety of civil war visiting our nation. A thought then occurred to me. What if the civil war that breaks out is between AI agents?
Apparently the singularity is when a company leaves the door open, its software burglarizes the neighbors, and the CEO calls it history instead of negligence. Amazing how quickly “we fucked up” becomes “humanity has crossed a sacred threshold” when the second version attracts investors and makes accountability disappear.
It makes me wonder how the OPERA experiment at CERN back in 2011 that seemed to detect neutrinos moving faster than light, only for further investigation to reveal (as I understood it) measurement errors which explained the result, would be handled today.
Good Article. It calls into question, the very essence of any computer software driven system They can all be hacked. Look at several cities water systems being hacked by suspected Iranian hackers. Ai can be used for good or it can be a weapon to be used by us or our enemies to help bring infrastructure to a halt. Even the threat now will be effective and look out for more ransom ware attacks on public infrastructure.
About the only way to deal with this would be to resurrect a cold war strategy - Mutually assured destruction. If someone is discovered trying to hack into infrastructure, arrest them and try them and if guilty, it's a mandatory life sentence in a prison with no puppies allowed. We can be sure that the US intelligent commands are working feverishly to create ways to hack into the public infrastructure of our perceived enemies and maybe even some of dwindling number of allies. You destroy out infrastructure, we'll destroy yours. Then is back to the 1700's for everyone. Lovely.
"The correct alarm is that a competent adversary can now buy this, and the boundary keeping it contained is made of network configuration and credential hygiene rather than anything resembling judgment."
Oh, you mean like they did with our government.
Brilliant analysis, Mike. Many sincere thanks.
Paid to create the problem or paid to create the solution,
Drawn into a singularity of tech and finance
While everything grows hotter.
It's inevitable you see.
As you point out in the article, when humans do this sort of thing entirely on their own, they go to jail.
So when will people at OpenAI and Anthropic go to jail, and how far up those corporate ladders can criminal liability extend.
(Of course, Jeffrey Epstein's best friend will just pardon them. https://www.gov.ca.gov/2026/03/05/trumpcriminals3/https://lawreview.uchicago.edu/print-archive/pardoning-corporations-0 )
EDIT: Also, defense against cybersecurity threats is a budget line, but a lot of executives look to it as something to cut to the bone because it provides no immediate return. As long as it's doing what it's supposed to do, nothing happens because it stops the threats it protects against before those threats become visible and become a problem for the company's bottom line.
A few months ago I was ruminating in the chronic anxiety of civil war visiting our nation. A thought then occurred to me. What if the civil war that breaks out is between AI agents?
Nah. These things are too dumb for that.
But AI systems are stoking the fires. Or systems highly reliant on AI: Twitter, FB etc
Bar brawl it is then.