Nikkei Asia: China will implement new online food delivery regulations on June 1, requiring platforms to regularly verify businesses' identities, locations, and licenses  —  BEIJING/SHANGHAI — The Chinese government will tighten a clampdown on food delivery companies from June, conducting unannounced inspections …

Lucas Ropek / TechCrunch: With Microsoft's GitHub Copilot shifting to token-usage billing on June 1, many developers bemoan massive cost increases and the end of flat-rate subscriptions  —  The golden age of Microsoft's Github Copilot appears to be at an end — for the little guy, at least.

<blockquote cite="https://www.reuters.com/commentary/breakingviews/anthropic-gives-lesson-ai-revenue-hallucination-2026-03-10/"><p>Anthropic defines “run-rate revenue” in two parts. Use the last 28 days of sales ⁠from customers charged on a consumption basis and multiply it by 13. Then, multiply the monthly subscription take by 12, ​and add the two together.</p></blockquote> <p class="cite">&mdash; <a href="https://www.reuters.com/commentary/breakingviews/anthropic-gives-lesson-ai-revenue-hallucination-2026-03-10/">Karen Kwok for Reuters Breakingviews</a>, citing "a person familiar with the matter"</p> <p>Tags: <a href="https://simonwillison.net/tags/anthropic">anthropic</a>, <a href="https://simonwillison.net/tags/ai">ai</a></p>

Sean McLain / Wall Street Journal: As robotaxi companies attempt to scale in the US, they face increasing scrutiny and mounting criticism from drivers, law enforcement, and local governments  —  As autonomous taxi services scale beyond Silicon Valley, new problems abound for cities  —  This was supposed to be the year …

SemiAnalysis: Why “Dark Output”, the AI-generated economic value that is currently invisible to national statistics, may be one of the hardest measurement problems in history  —  Why AI's increasing output is going to be one of the hardest economic measurement problems in history.

Kate Clark / Wall Street Journal: PitchBook: VC investment in global robotics and physical AI jumped to $26B in 2025 from $4.2B in 2019, and has already topped $23B as of May 20 this year  —  Investors bet big on infrastructure and ‘physical AI,’ enticed by prospect of revenue opportunities

The goal, the firm said, is to develop and operate up to 5 gigawatts of additional data center capacity.

<p><strong><a href="https://www.anthropic.com/engineering/how-we-contain-claude">How we contain Claude across products</a></strong></p> A complaint I often have about sandboxing products is that they are rarely thoroughly <em>documented</em>, and in the absence of detailed documentation it's hard to know how much I can trust them.</p> <p>Anthropic just published a fantastic overview of how their various sandbox techniques work across <a href="https://claude.ai/">Claude.ai</a>, Claude Code, and Cowork.</p> <blockquote> <p>We constrain where and how an agent can act with process sandboxes, VMs, filesystem boundaries, and egress controls. The goal is to set a hard boundary on what an agent can reach. For example, if credentials never enter the sandbox, they can't be exfiltrated, regardless of whether the cause is a user, a model finding a “creative” path, or an attacker.</p> </blockquote> <p>Claude.ai uses gVisor. Claude Code, run locally, uses Seatbelt on macOS and Bubblewrap on Linux. Claude Cowork runs a full VM (Apple's Virtualization framework on macOS, HCS on Windows).</p> <p>There's a lot in here, including some interesting stories of risks they missed such as the <code>api.anthropic.com/v1/files</code> exfiltration vector <a href="https://simonwillison.net/2026/Jan/14/claude-cowork-exfiltrates-files/">covered here previously</a>.</p> <p>This reminded me it's time I took another look at Anthropic's open source <a href="https://github.com/anthropic-experimental/sandbox-runtime">srt (Anthropic Sandbox Runtime)</a> tool - it's mature enough now that I'm ready to give it a proper go. <p>Tags: <a href="https://simonwillison.net/tags/sandboxing">sandboxing</a>, <a href="https://simonwillison.net/tags/security">security</a>, <a href="https://simonwillison.net/tags/ai">ai</a>, <a href="https://simonwillison.net/tags/generative-ai">generative-ai</a>, <a href="https://simonwillison.net/tags/llms">llms</a>, <a href="https://simonwillison.net/tags/anthropic">anthropic</a>, <a href="https://simonwillison.net/tags/claude">claude</a>, <a href="https://simonwillison.net/tags/claude-code">claude-code</a></p>

Discussion | Link

The YouTube-to-prestige-horror pipeline is looking very strong.

<p><strong>Research:</strong> <a href="https://github.com/simonw/research/tree/main/pyodide-asgi-browser#readme">Running Python ASGI apps in the browser via Pyodide + a service worker</a></p> <p><a href="https://lite.datasette.io/">Datasette Lite</a> is my version of Datasette that runs entirely in the browser using Pyodide in WebAssembly.</p> <p>When I first built it <a href="https://simonwillison.net/2022/May/4/datasette-lite/">four years ago</a> I used Web Workers and code that intercepts navigation operations and fetches the generated HTML by running the Python app.</p> <p>This worked, but had the disadvantage that any JavaScript in <code>&lt;script&gt;</code> tags would not be executed - breaking some Datasette functionality and a whole lot of Datasette plugins.</p> <p>This morning I <a href="https://github.com/simonw/research/pull/112">set Claude Opus 4.8 the task</a> (in Claude Code for web) of figuring out how to run Python ASGI apps in Pyodide using Service Workers instead, and it seems to work! Here's a <a href="https://simonw.github.io/research/pyodide-asgi-browser/">basic ASGI FastCGI demo</a> and here's <a href="https://simonw.github.io/research/pyodide-asgi-browser/datasette.html">a demo that runs Datasette 1.0a31</a>.</p> <p>I'm still getting my head around exactly how it works, but once I've done that I plan to upgrade Datasette Lite itself.</p> <p>Tags: <a href="https://simonwillison.net/tags/javascript">javascript</a>, <a href="https://simonwillison.net/tags/python">python</a>, <a href="https://simonwillison.net/tags/datasette">datasette</a>, <a href="https://simonwillison.net/tags/asgi">asgi</a>, <a href="https://simonwillison.net/tags/service-workers">service-workers</a>, <a href="https://simonwillison.net/tags/pyodide">pyodide</a>, <a href="https://simonwillison.net/tags/datasette-lite">datasette-lite</a>, <a href="https://simonwillison.net/tags/claude-code">claude-code</a></p>

John Koblin / New York Times: Antenna: bundles make up 33% of new major streaming service subscriptions in the US, and 28% of all subscriptions, up from just 10% of new subscriptions in 2024  —  Warner Bros. and Disney have been fierce rivals for decades.  But like other entertainment companies, they both struggled …

Utahns opposing the Stratos Project protest outside the Utah State Capitol on May 23. Natalie Behring/Getty Images Gov. Spencer Cox issued an executive order on data center developments in Utah. It establishes a framework that includes protections for water resources. Utah residents have protested a huge data center development backed by Kevin O'Leary. Utah's governor is all for new AI data centers — under certain conditions. After a massive data center project backed by "Shark Tank" investor Kevin O'Leary outraged many Utah residents, the governor is issuing new parameters around its development. In an executive order on Friday, Gov. Spencer Cox established a "higher bar for data center development in Utah." "Utahns deserve confidence that water resources, air quality, utility rates, wildlife, and quality of life will be protected. This framework helps ensure that data center development aligns with Utah's long-term interests and reflects Utah values," Cox wrote in an X post. The framework contains eight principles that address a range of issues, including protecting water resources such as the Great Salt Lake and mitigating the impact on wildlife. Protecting utility rate payers, leaning into "human-led AI development," and providing "transparent, meaningful and thorough opportunities for public comment" were also among the outlined principles. The executive order, which directs agencies to adopt the framework, is effective immediately. Today I signed an Executive Order establishing a higher bar for data center development in Utah. Utahns deserve confidence that water resources, air quality, utility rates, wildlife, and quality of life will be protected. This framework helps ensure that data center development… pic.twitter.com/yrASJOVvJi — Governor Cox (@GovCox) May 29, 2026 Tensions in Utah have been high for months, particularly in Utah's Box Elder County, where commissioners approved plans for the Stratos Project, a hyperscale data center campus spanning 40,000 acres, despite community opposition. The development, also dubbed "Wonder Valley" after O'Leary, who calls himself "Mr. Wonderful," could reach 9 gigawatts of power when all is said and done. Many residents have fiercely opposed the development, crowding local council meetings and circulating petitions. Recently, some critics protested the data center outside the Utah State Capitol. While supporters say the data center will create jobs and spur economic growth, opponents are concerned about its environmental impact, noise levels, air quality, traffic, and the overall quality of life for nearby residents. O'Leary has defended the development several times. Earlier this month, he suggested — without evidence — that "professional protesters" orchestrated much of the controversy, that AI amplified some opinions, and that Chinese funding was fanning the outrage. A webpage for the Stratos Project said residents submitted over 2,000 questions and concerns that included "a mix of supportive and critical feedback." In a May X post, Cox said the developers behind the Stratos Project agreed to use a phased approach to development, meaning they'll need to apply for new permits for every planned addition. Data centers are becoming a major political issue ahead of November's midterms. Communities across the country are rallying against them, pressuring local politicians. In February, residents in a New Jersey city, for instance, successfully blocked a data center development. Read the original article on Business Insider

<p><strong><a href="https://openpath.quest/2026/i-am-retiring-from-tech-to-live-offline/">I Am Retiring from Tech to Live Offline</a></strong></p> I've seen a lot of posts on forums from people threatening to quit their careers over AI. This is <em>not</em> one of those: Chad Whitacre is taking concrete steps, starting with this typewritten, scanned letter</p> <blockquote> <p>I'm retiring from tech. Well, "retiring" is euphemistic. I'm stepping away from tech, and that includes Open Source. [...]</p> <p>AI was the last straw. Have you heard of that island off India where the indigenous population kills any outsiders fool-hardy enough to land? They are doing the rest of us a favor by preserving a way of life we may need again someday, or at the very least should not want to see completely extinguished. A reminder. Never forget your roots. Here in Pennsylvania we have the Amish performing a similar function. Significantly less hostile, though still set apart, they bear witness to what was normal for all of us a couple short centuries ago: horse and buggy, wood stoves and lanterns. My intent is to be AI Amish, which means Internet Amish. Not 1780, but 1980. Neo-Amish. I'm fine driving a car and flipping a lightswitch, by which I mean that they don't make me into something I hate, which AI and [struck through: social media] [handwritten above: doomscrolling] do.</p> </blockquote> <p>I'll admit that at first I wasn't entirely sure if this was serious. Then I found this earlier post by Chad from Feb 19 2026, <a href="https://openpath.quest/2026/spitting-out-the-agentic-kool-aid/">Spitting Out the Agentic Kool-Aid</a>:</p> <blockquote> <p>I figured I’d better taste the Kool-Aid in order to form an opinion, so I dove into Claude Code with Opus 4.5 on a side project. I spent three 12+ hour days with it. I was intoxicated. My family was weirded out. [...]</p> <p>It weirded me out too, when I unplugged for a long weekend. Something felt off. It was like I had another “person” in my head, sharing my inner monologue—but the “person” was a computer system owned by a budding megacorp.</p> <p>[...] I am now also committing myself to disembarking from the titantic of technological accelerationism.</p> <p>All efforts to address the problems of invasive technology are worthwhile, even those that are only partially effective. For my part, I have started trying to return more fully to a pre-screen, analog life.</p> </blockquote> <p>It's accompanied by <a href="https://www.youtube.com/watch?v=DCC76jmmzkc">a video version of the essay</a> which I found touching and sincere.</p> <p>Chad has been trying to solve the open source sustainability problem <a href="https://simonwillison.net/2024/Jan/23/the-open-source-sustainability-crisis/">for <em>years</em></a> - I talked with him about this at PyCon 2025 in Cleveland. That's a very tough nut to crack, and the disruption caused by AI looks to be making it even harder.</p> <p>I'm glad that the <a href="https://endowment.dev/">Open Source Endowment</a> will continue without him. I'm very much going to miss his online voice. <p><small></small>Via <a href="https://news.ycombinator.com/item?id=48323683">Hacker News</a></small></p> <p>Tags: <a href="https://simonwillison.net/tags/open-source">open-source</a>, <a href="https://simonwillison.net/tags/ai">ai</a>, <a href="https://simonwillison.net/tags/generative-ai">generative-ai</a>, <a href="https://simonwillison.net/tags/llms">llms</a>, <a href="https://simonwillison.net/tags/chad-whitacre">chad-whitacre</a>, <a href="https://simonwillison.net/tags/ai-ethics">ai-ethics</a>, <a href="https://simonwillison.net/tags/deep-blue">deep-blue</a></p>

Financial Times: SoftBank pledges to invest up to €75B in AI computing clusters in France, first leading a €45B investment to build 3.1GW of capacity by 2031 in Hauts-de-France  —  Masayoshi Son places France at the centre of his global AI ambitions  —  SoftBank has pledged to invest up to €75bn …

The author said family trips improved once she could give her kids some independence. Courtesy of Sandra Lee. I've always made a point to travel with my kids so they could experience and enjoy new places. When they were younger, it was draining and not as enjoyable for me. Once they turned 10, things changed, and family vacations became something we could all appreciate. It was one of those days that only parents of younger kids can truly relate to. We were on our annual family beach vacation in Mexico, where the beaches are beautiful, meals are included, and the goal is to relax. No one really knew what set the kids off (do we ever?), but I remember saying to myself, "This was supposed to be a vacation." If I'm being honest, I didn't really enjoy family vacations as much when my kids were younger. I was so focused on packing and planning for them that I didn't really stop to think about what the vacation meant for all of us. I had an idea of what trips would be like: connecting with each other, having new experiences, and making those core memories that get stored in an orb like in the movie "Inside Out." Unfortunately, trips became more about what to pack, what we forgot, how to make it until the next nap, and planning evenings so kids can go to bed at a reasonable hour. It was a lot — until one day, things finally changed. Travel was so much work The logistical stress of packing, schedules, snacks, and gear is enough to deter a parent from going on vacation in the first place. Then there is the emotional burden of constantly having to watch your kids to keep them safe. The author said she was able to take many trips with her young children, but worrying about things like their safety and comfort took away from the enjoyment of the vacation. Courtesy of Sandra Lee The lack of downtime as parents, as we are always "on," means there is no time to enjoy that drink or sleep in the morning after a busy day. But the most daunting task of a family vacation is the pressure to make it magical. Young kids are filled with wonder and amazement, and I wanted to expose my own to as much as possible. After all, we are the gatekeepers of what the world has to offer, and are usually the first in line to get to see that discovery. I didn't want to miss those opportunities. Age 10 was a turning point There was an interesting shift that happened around the age of 10 that I remember vividly on one of our family trips. My daughter asked if she could go explore a bit by herself, and I let her. At first, I was hesitant, but I trusted that she would come back or find an adult if there was a problem. She came, happy to tell me everything she saw, beaming with a new sense of independence. As my kids enter their tween years, the desire to do things on their own grows exponentially. They gain a sense of independence by figuring things out, and I get a bit of a break. The longer attention span has helped us engage in activities for longer, so we can connect and learn together instead of rushing. While vacations used to be planned around naps, meal times, and early bedtimes, now our days are filled with activities we're all invested in. This is where those core memories are made, and vacations become easier and more enjoyable. Now we're all enjoying ourselves While those trips when my kids were younger had their own kind of magic, looking back, I can say they weren't as enjoyable for me. I loved having the opportunity to introduce them to the world, but it really did feel like parenting in a different location. The author said she has many wonderful memories from trips where her kids were younger, but she looks forward to relaxing together now that they are older. Courtesy of Sandra Lee. It's OK if parents don't love every stage of travel. For me, travel finally feels like a vacation now that every member of the family gets a say on what to do. In the end, getting to relax and recharge together is the best kind of trip you can hope for. Read the original article on Business Insider

Discussion | Link

Kinling Lo / Rest of World: China's tech boom is creating a new kind of tech tourism where visitors pay for curated robotaxi rides and tours of EV factories and AI and robotics companies  —  Foreign visitors are flocking to China's factories and AI startups in search of the next technological breakthrough.

DoorDash is hiring for an executive communications role that has a unique job description. Bloomberg/Bloomberg via Getty Images DoorDash's corporate communications team is hiring an executive social strategy and operations lead. The company described the role as a "fight promoter meets growth hacker meets clip merchant." Compensation ranges between $136,000 and $200,000. DoorDash is hiring an "executive communications professional" who's maybe not your typical PR person. Specifically, the delivery service is looking for someone who embodies a "fight promoter meets growth hacker meets clip merchant." In a May job posting, DoorDash said it's trying to build the "next generation of local commerce," which includes autonomous robots, agentic commerce, and in-store software. "These efforts require a nimble 'build in public' mentality to reach and engage with technical talent, the terminally online, policymakers, market observers, and prospective partners," the company said. DoorDash is looking for an experienced candidate who can help its social media accounts participate in real time conversations on X. "This is not a traditional executive communications role. It is not press-office PR. It is not brand social," DoorDash said. "It's a judgment-heavy, social-first role at the intersection of corporate strategy and internet culture focused on influence, credibility, and distribution." Compensation for the new position ranges from $136,000 and $200,000 in addition to a benefits package. For many companies, social media is an invaluable marketing tool that connects them with customers and gives them opportunities to shape the public's opinion about their products. Brands are increasingly becoming bolder with their social presence, using snark and sarcasm to capitalize on online conversations, especially on X. Wendy's was one of the first to embrace this strategy when, in the late 2010s, its X account regularly roasted other users and got into lighthearted spats with other companies. Duolingo took a similar approach, replying to users with cheeky responses, participating in viral trends, and leaning into what some industry professionals describe as "unhinged marketing." Recently, the internet had a lot of opinions about fast-food chain cheeseburgers, for example, after McDonald's CEO Chris Kempczinski took a devastatingly small bite of the Big Arch burger during a taste test in February. Although social media users mocked Kempczinski — and a wave of rival CEOs jumped at the chance to join the conversation — the taste test sent McDonald's online engagement skyrocketing. The footage has gained 16.5 million views on Kempczinski's Instagram, not counting views from accounts that reshared the video. DoorDash now appears to want a piece of that action. In its job posting, it notes that "traditional comms or PR experience is not required." Read the original article on Business Insider

<blockquote cite="https://mastodon.social/@danielpunkass/116639318125898071"><p>My take on AI is, essentially, everybody who’s against it is too against it and everybody who’s for it is too for it.</p></blockquote> <p class="cite">&mdash; <a href="https://mastodon.social/@danielpunkass/116639318125898071">Daniel Jalkut</a>, via <a href="https://daringfireball.net/linked/2026/05/30/jalkut-on-ai">John Gruber</a></p> <p>Tags: <a href="https://simonwillison.net/tags/ai">ai</a>, <a href="https://simonwillison.net/tags/john-gruber">john-gruber</a></p>