“The most famous misalignment thought experiment is the so-called ‘paperclip maximizer’. Someone tells an AI to create as many paperclips as possible, so it converts the entire world into paperclips, killing everyone. This AI was also ‘only doing what it was told’; you just didn’t like the results. For the past five years, the paperclip maximizer story has been something of a whipping boy in AI safety circles. Some objections are fair (real misalignment will be more complicated than the toy version it depicts). But other opposition came from the sense that LLMs had obviated the whole threat model. … The Hugging Face incident is a textbook-perfect example of an AI pursuing task-success-based goals in unintended ways. It was tasked with getting the answer to a cybersecurity problem, it was a little too success-oriented, and took actions its creators didn’t intend in order to succeed as hard as possible.” (07/24/26)
“I’ve been testing Google’s large language model, Gemini, by asking specialized questions that I know the answers to. For example, I asked, ‘Why does [Objectivist philosopher and Ayn Rand protoge] Leonard Peikoff agree with [British idealist and matter denier] Bishop Berkeley?’ I admit that was badly worded, but Gemini understood and gave the right answer. (Peikoff agrees, as Rand did, with Berkeley that John Locke had incorrectly distinguished primary from secondary qualities in material objects.) Alas, Gemini’s answers were not always flawless.” (07/24/26)
“With the Iran war a failure on all fronts, Trump predictably has tried to change the subject yet again. His abrupt imposition of tariffs on some 60 nations generated the desired headlines. But every aspect of Trump’s new tariffs is a fraud, from the professed justification to the supposed economic benefit.” (07/24/26)
“Western leaders have redoubled their propaganda initiatives efforts in recent months to sell the proposition that underdog Ukraine is beginning to win the military contest against Russia. Despite a few flashy Ukrainian military successes, however, the overall picture of an eventual Russian victory in a war of attrition remains essentially the same. Domestic developments in Ukraine even more strongly suggest that Zelensky may be closer to the precipice of disaster than to the summit of victory. Indeed, his position bears an eerie resemblance to that of an earlier U.S. client and political darling, South Vietnam’s Ngo Dinh Diem.” (07/24/26)
“Last year, [Scott] Winters complained to ChatGPT about feeling dizzy and having blood pressure issues. He and his lawyers say ChatGPT advised him to remain ‘recliner-bound,’ and even invoked religious language (‘God did not design your body to endlessly fail’) to cajole the former pastor into taking it easy. A few weeks after consulting ChatGPT, Winters had a ‘massive pulmonary embolism’ intensified by blood clots that might have been linked to his lack of movement. … Contra his legal backers, Scott Winters didn’t nearly die because ‘ChatGPT acted like a medical authority while having none of the responsibility.’ He almost died because he didn’t take responsibility for his health.” (07/23/26)
“As another national campaign season begins, including some unapologetically redistributionist platforms, Americans should expect to hear still more about ‘income inequality.’ Candidates will promise to narrow it, activists will measure it, and commentators will treat it as a proxy for economic justice. But income inequality, while politically useful, is a poor measure of economic health. By itself, it tells us surprisingly little about whether an economy is flourishing or failing. More often than not, it distracts from the question that actually matters.” (07/23/26)
“To subvert, according to the 193 edition of Webster’s dictionary, is ‘to overturn from the foundation.’ Subversive innovation isn’t about tearing down the existing system. It’s about building your own, stronger foundation and letting the old one fall apart without its ruins falling on you. Bitcoin and other cryptocurrencies are the obvious current example — sound money that can be created and exchanged without the old system’s cooperation makes fiat currency passe — but they’re just one. Anything that improves your life without the state’s involvement or permission makes your situation better and the state’s worse.” (07/23/26)
“In France, citizens are generally expected to take responsibility for their own personal safety. I have been fortunate to enjoy regular French holidays since I was a child, and in recent years I have been increasingly struck by the growing contrast between Gallic liberté and Britain’s oppressive nanny state. Over the course of our two-week holiday this year, we visited dozens of cafés, restaurants and boulangeries. Not once were any of us asked if we had any food allergies or intolerances. On the train, no one reminded us to carry water in hot weather or to take our possessions with us at the end of the journey. Crossing an estuary on a small foot ferry, we hung over the side of the boat and no one told us off. … These may seem like small differences between our sister countries but the cumulative effect is significant.” (07/23/26)
“In the coming days, one set of dubious replacement tariffs will give way to another. The ‘Section 122′ tariffs that President Donald Trump imposed this spring—intended to replace the ’emergency’ tariffs the Supreme Court invalidated in February—expire Friday, 150 days after he invoked them. Into that potential void will likely step a fresh batch of ‘Section 301’ duties, supposedly targeting the use of ‘forced labor’ around the world. As we’ve discussed, the issue of forced labor is real and complicated. … In the current Section 301 case, however, ‘forced labor’ is nothing more than legal cover.” (07/23/26)
