Kevin Underhill, the brilliant satirist whose day job is lawyering, reports on delicious and often nutsy but always actual legal tales and takes in his newsletter, LoweringtheBar.
Often, he gives us a collection of short items which he calls Assorted Stupidities. The lead Stupidity in his last report is this, on AI:
In preparation for what it hopes will be a ginormous IPO, AI company Anthropic filed a ginormous SEC registration statement this week. A major function of such things is to disclose “risk factors” to potential investors. I haven’t read Anthropic’s disclosures in detail, but I did notice a couple interesting facts. For example, Anthropic’s net losses have quintupled every year so far, and it projects a net loss of $160 billion for 2026. It also warns that it may never become profitable, although that warning I think is pretty common. Less common, I would guess, is the warning that first appears on page 46, where Anthropic notes that its products may “pose catastrophic or existential risks to humanity….” (Emphasis added.) This warning is repeated several times in the following pages. So I think the takeaway is pretty clear: if AI destroys the entire human species, don’t bother suing Anthropic for that. It warned you. You will get nowhere.