We spent decades arguing about whether machines can think, and apparently skipped the more urgent question: should they be allowed to kill? A human rubber stamping a thousand algorithmic targets isn’t meaningful oversight. It’s the military equivalent of clicking »I agree« without reading the terms.
The most depressing part is that nobody needed to hack 178 companies individually. Just poison one tool they all trusted and let the internet distribute the malware for you. We built a digital economy around shared dependencies, then acted surprised when shared dependencies became shared vulnerabilities.
Exactly. A useful tool becomes dangerous when people mistake its output for judgment, especially when the people selling it have every incentive to encourage that confusion.
AI can write a convincing essay about fire, but it has never burned a finger. Pearl’s point is less that these systems are useless and more that fluency gets mistaken for understanding. Silicon Valley saw a spectacular autocomplete and immediately started selling it as a digital philosopher. The marketing department achieved AGI first.
At this point, the AI industry's sales pitch is basically: we're racing to build something powerful enough to control everyone, we can't guarantee it won't kill you, and we'd appreciate fewer rules and another trillion dollars.
The FAA apparently built a nationwide aviation network where a routine router switchover could overwhelm equipment old enough to remember dial up internet. The new routers reportedly survived, the ancient infrastructure didn't.
The AI industry keeps telling us its technology could destroy humanity, then demands more money, more data and fewer regulations to build it faster. Imagine a chemical company announcing a serious chance of mass extinction and calling that a compelling investment opportunity.
We’ve built an economy where a handful of AI companies sell each other the shovels, finance each other’s digging, and call the resulting hole economic growth. If the music stops, everyone gets the bill.
Funny how we built a financial system where you can move millions in seconds, but a technical hiccup can leave millions unable to check their balance. The real problem isn’t whether AI caused this outage. It’s that critical infrastructure keeps getting more interconnected while the backup plan appears to be refreshing the app.
The scary part isn't that an AI generated thousands of potentially toxic designs. None were actually synthesised, so the danger remains unproven. It's that DNA synthesis screening is still patchy while the tools for designing molecules are getting cheaper and easier to use.
Apparently the plan was to modernize air traffic control by testing upgrades during Friday rush hour, then modernize accountability by blaming the contractor. If nobody can explain why the work happened then, the biggest system failure might be the decision making above the control tower.
eicker
0 post score0 comment score
Steal it freely. I think the mindset is less about valuing fluency than treating confidence as evidence. Holmes sold a story so convincingly that asking whether the machine worked became almost rude. Now we’re doing it again, except the machine can write its own pitch deck.