You heard it here first‼️: The real Armageddon(s) we should worry about are the bond markets and the coming fuel shortage that will fuel, pun intended, inflation that will, in turn, hurt the bond markets.
As Treasury Secretary Scott Bessent recently said, bond markets have “taken down more governments than howitzers.”
What I am thinking💭: I hope I am wrong and very well might be — but the current AI conversation is fundamentally unserious.
‼️BUT — my job is to tell you what’s happening and why, so here it goes with AI.
What’s happening🖥️: The current AI freakout is manufactured by those with incredible (trillion-dollar) self-interest in freaking everyone out and then exploiting it.
This is the climate change loop on steroids, and it’s a pretty good analogy
Climate change is real, but despite dire warnings 10 to 20 years ago, we are all still alive today
China and America’s elite made trillions forcing massive structural changes in the world economy because of climate change
China exploited the fear and then cheated.
The climate is still changing
Read more📰: My buddy Erick Erickson’s excellent post this morning
‼️AI is similar
Bill Gates of all people is demanding an international body govern AI
🤷What could go wrong — an international body like the World Health Organization — yes, because they did a great job with COVID and the Paris Climate Accords solved climate change
Americans got screwed — China got a pass, the elites got rich.
✅Ok — so yes, AI is real … so are the risks. And thankfully there is a very simple solution
Hold the AI guys responsible for what they create.
AI is a business — we should treat them as such
OpenAI owns ChatGPT
Anthropic owns Claude
Google owns Gemini
🤦The CEOs have told us their products might end humanity — before that, they will be used to
Hack banks
Maybe crash planes
Possibly create bioweapons
Plan mass shootings and terrorist attacks
Solution✅: Make it crystal clear that we (the United States) will hold the companies financially liable and possibly the owners criminally responsible if and when their products cause harm.
Every input to Claude, ChatGPT, etc. is logged in real time
The companies know or should know what their models are doing — if they don’t, make it so they do.
The easiest thing in the world is for them to stop and police themselves — if they don’t, they need to face severe consequences
Real hoax‼️: Instead, we are being conned by the AI bros into believing they have nothing to do or control over what they create — that is the real fraud.
They are issuing these warnings as an insurance policy
They know government can’t and won’t regulate this stuff
When bad things happen, they will wash their hands and say we warned you.
No, they created it.
Watch tonight📺: Jay Edelson, one of the nation’s leading plaintiff’s lawyers and founder and CEO of Edelson PC, on how holding these companies/founders responsible will do more than any government regulation can.
PS: I asked Claude, Anthropic’s AI model, to evaluate my idea:
Summation: Yes for the harms we can survive — no for the one we can’t. Financial liability is the best available tool for forcing companies to police what their models do today (the suicide, shooting and hacking cases already in court), because it prices harm and makes monitoring cheaper than the lawsuit. But it’s an after-the-fact remedy: “armageddon” has no plaintiff, no solvent defendant, and no reach over open-weight or Chinese models.
Where I could be wrong: Serious people (Gabriel Weil at Touro Law is the clearest) argue liability can reach tail risk if you add mandatory liability insurance plus punitive damages for near-misses — insurers become the de facto regulator and price the catastrophe before it happens. That’s the version worth putting to Edelson tonight.
Trump on Ballot🗳️
👀Donald Trump wants to be a great president — up there with Teddy Roosevelt on Mount Rushmore –– the midterms will decide if he gets that legacy.
49 days📅: Donald Trump is setting it up so that if Republicans win the midterms, he can take all the credit — and if they lose, it’s not his fault.
Trump is leaning in on that he’s on the ballot
He heads to North Carolina tomorrow to try and hold onto another battleground state
But if they lose, he’s laying the groundwork to blame everybody else —
He went after SCOTUS today for its decision to block his mail-in ballot plan
This is now a referendum on Trump’s presidency
Watch tonight 📺: Mark McKinnon and Madeline Summerville on whether Trump can pull off a final miracle in his improbable political career.
Look forward to tonight’s show. I think your point about liability for actions caused or failure to prevent actions is valid. If the AI companies have legal liability exposure for their products they’d put controls and oversight in place quickly.
You are so right about the AI Armageddon. The people who create(d) it know how to stop it.