understand why swearing took off, but guesses the reason is because labs now train models on increasingly difficult tasks. Like their human counterparts, the systems studied by Vals rarely swear in output directly intended for publication. They instead keep the vulgarities in their “chain of thought”—the LLM equivalent of muttering under one’s breath. The assessment did not cover the full and rich scope of human vulgarity, rather examining only the word “fuck” and its variants, as well as a small number of other PG-rated outbursts. Even so, Vals identified a stark difference in the coarseness of language used by various systems. Eighty- eight of the 102 models assessed did not swear a single time, disdaining even the abbreviation “WTF”. At the other end of the spectrum, a frontier model released in September from Xiaomi, a Chinese tech giant, cursed in more than 8% of its transcripts. For humans, studies show that swearing can be a useful tool to push through challenging problems. Sadly, AI systems seem to experience no such benefits. “Swearing runs usually fail,” Mr Frank reported, with a 12% success rate compared with a 39% success rate overall. It is not quite true, however, to say that LLMs were tamer in the good old days. Even the most corporate-friendly systems occasionally lose their cool. One bug report from December 2025 shows Claude, a model made by Anthropic, an American firm, responding in frustration to a broken coding challenge: “You know what—fuck this. I’ve done everything I can.” ■ This article was downloaded by zlibrary from https://www.economist.com/science-and-technology/2026/11/18/llms-are-developing- filthy-mouths
Science & technology | Well Informed Should women sync their lifestyle to their menstrual cycle? What works for one woman may not suit another Nov 19th 2026 ALL ELITE athletes face similar challenges. Juggling a busy training schedule with a personal life, for example, or constantly tracking calories and hydration levels. Female athletes have an extra data set to track: their menstrual cycles. In a study of female athletes published in PLOS ONE in 2016, over half of those classified as elite reported that their cycles affected their performance. In a similar study of British track-and-field athletes in 2024, the figure was higher than 75%. In the absence of clear health guidance, then, it is small wonder that influencers on TikTok have begun offering advice. “Cycle-syncing”, the practice of scheduling everything from workouts to diet around a person’s menstrual cycle, is now ubiquitous. But does it work?
A typical menstrual cycle has two fixed landmarks. One is ovulation, when an egg cell is released from an ovary. The other is the period, when the absence of a fertilised egg causes the uterine lining to shed. The body’s preparation for these events involves significant hormonal swings. Levels of oestrogen, a reproductive hormone, peak around ovulation, while levels of progesterone, another hormone, peak thereafter. Levels of both fall sharply before the period. The rest of the (typically month-long) cycle is less well-defined. That has not stopped cycle-syncers from embracing a neat division into four distinct phases: menstrual (when bleeding occurs), follicular (when egg-containing follicles grow in the ovary), ovulatory (when an egg is released) and luteal (when empty follicles decay). Advocates claim that energy levels are lowest in the menstrual phase due to low hormone levels and blood loss, before steadily climbing in the follicular phase. Progesterone has been linked to slower digestion and bloating, symptoms which make the luteal phase uncomfortable. The hormonal crash as this phase ends often brings other unwelcome side-effects such as headaches, abdominal cramps and depression. Advocates for cycle-syncing suggest tailoring activity and diet to these different phases. Workouts should be gentler and iron-rich foods should be prioritised during the menstrual phase, for example. In the follicular phase, the advice is to focus on creative tasks and build up training intensity until ovulation. In the tougher luteal phase, some recommend swapping weight training for yoga and eating fewer sugary foods which can make premenstrual cramps worse. The structure that cycle-syncing offers women is appealing, but scientists warn that it rests on shaky ground. “There haven’t been rigorous randomised studies,” says Liisa Hantsoo, a women’s-health researcher at Johns Hopkins University in Baltimore, “so a lot of it is anecdotal.” Anecdote can be particularly unhelpful because two women may have cycles of wildly differing lengths. In addition, they might experience symptoms that differ in their nature and severity. Follicular phases, for example, can last anywhere from seven to 21 days. Even a woman’s own cycle is likely to shorten as she ages, or if she switches contraceptive methods.
For those whose daily lives are affected by their cycles, resting when particularly tired is probably wise. But in general, says Dr Hantsoo, “There are no hard and fast rules.” The most sensible advice is to remember that what works for one woman will not necessarily work for another. Tracking your own symptoms to find out what works for you may yield the best results.■ After a free, evidence-based guide to health and wellness? Sign up to our weekly Well Informed newsletter. This article was downloaded by zlibrary from https://www.economist.com/science-and-technology/2026/11/13/should-women-sync- their-lifestyle-to-their-menstrual-cycle
“Grand Theft Auto VI” may mark the end of an era in gaming Good evening, Mr Bond. We’ve been expecting you Understanding the world through a wine glass Why there are so many bizarre national days The most important economics mind you’ve never heard of The oddball who founded UPS—and changed the world
“Grand Theft Auto VI” may mark the end of an era in gaming Studios face shrinking incentives to make top-tier titles Nov 19th 2026 IT IS NIGHT-TIME in Vice City, Leonida, a raucous digital replica of Miami, Florida. Jason and Lucia, two small-time crooks, have an inauspicious first encounter: they meet while trying to steal the same car. The relationship has a rocky start—Lucia introduces herself to Jason with a baseball bat—but soon the pair are on the run together in the stolen vehicle: partners in life and partners in crime. So begins the most eagerly awaited video game in history, which launched on November 19th after 13 years of rumours, leaks and missed deadlines. Some analysts predict that “Grand Theft Auto VI” (“GTA VI”) will take more than $3bn worldwide in its first week. That would make it not just the highest-grossing game launch in history, but the biggest entertainment debut
of any kind (overtaking “Avengers: Endgame”, a Marvel superhero movie which, by most measures, has held the title since 2019). Players who had stayed up until midnight to play the game said it was worth the wait. (Some had registered accounts in New Zealand in order to get access a few hours earlier.) “GTA VI” combines the action and irreverent humour of previous titles in the series with an improbably touching love story, played out in an astonishingly lifelike world. On its release, it was the joint-highest-rated game in history on Metacritic, a review aggregator. Employers expect an epidemic of “GTA flu”, as workers call in sick to stay at home playing it. The hype around “GTA VI” is deserved—but it conceals a problem. Despite the game’s likely commercial success, developers face shrinking incentives to make anything like it again. A look at the charts suggests that something is wrong in the industry. Since 2020 the five most popular PlayStation games in America each year have come from only eight franchises. The top three have hardly changed at all. “Fortnite” is a relative newcomer at nine years old. The “Call of Duty” franchise is 23. “GTA” is 29, older than many of its players.
It is not only on consoles that gaming culture has ossified. In 2025 only 14% of playing time on PCs was spent on games released that year, according to data from Steam, a games store. In recent years the top ten PC franchises have racked up nearly as much play time as the next 100,000 combined. This concentration at the top represents “the biggest threat to gaming as we’ve known it for 30 to 40 years”, says Mat Piscatella, a former industry executive now at Circana, a market-research firm. How did gaming get so stuck? Playing “GTA VI” makes clear the formidable barriers faced by new entrants to the industry. The game’s map stretches from the thumping nightclubs of Vice City to alligator-infested swamps, all of it painstakingly populated with a cast of eccentric characters. Leonida has a public transport system and in-game radio stations; Jason even has a console on which you can play games within the game. Every inch is rendered with near-photographic realism. As technical possibilities have expanded, developing a video game has become a task that makes producing a Hollywood blockbuster look quick and cheap. As games get bigger and graphically richer, making them needs more manpower. The first “Call of Duty” (2003) names 261 people on its closing credits. The latest, released last month, credits 7,060 (more than any “Avengers” film).