In the US, McDonalds HQ is “encouraging” its franchises to use AI to decide how to raise prices on food where possible. CC-licensed photo by Mike Mozart on Flickr.
You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.
A selection of 9 links for you. Short order. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.
Smart ring maker Oura postpones IPO due to market “uncertainty” • CNBC
Kai Nicol-Schwarz and April Roach:
»
Oura has said that it is delaying plans for its public listing on the Nasdaq due to uncertainty in the IPO market.
The delay comes despite “strong demand” and a strengthening of the business since the start of the IPO process, the company said in a statement on Tuesday.
The company, which makes a smart ring that tracks wearers’ health and sleep, formally launched its IPO plans just last week on Sept. 21.
At the time, the company planned to raise up to $2.2bn with the sale of 50 million shares.
“Our mission is to empower people to live healthier, longer, and an IPO is just one step in our journey,” said CEO Tom Hale on Tuesday. “We aim to deliver an extraordinary IPO for our employees and investors and we have the luxury of choosing our moment,” he added. “In the meantime, we will execute against the opportunities ahead.”
Oura is profitable, the company said, with revenue expected to grow 90% year over year for the fiscal year 2026.
Launched in 2015, Oura’s smart ring has evolved well beyond sleep tracking and now features a variety of features focused on broader health and wellness. In recent years, it has increasingly focused on advancing preventative health through new capabilities, AI, analytics and other features.
Oura is the latest US public market hopeful to delay an IPO in recent weeks. Holtec Nuclear withdrew its IPO earlier this month citing adverse market sentiment in equity markets.
Holtec said that uncertainty over data development compounded “pre-existing headwinds, including rising energy costs, elevated global trade tensions, ongoing military conflicts, and mounting inflation fears that have driven the central banks of major economies (EU, Japan and US) to raise their benchmark rates.”
«
Oura’s S-1 shows revenues in the first nine months of its FY26 of close to $1bn for hardware alone and another $240m for “membership”, and profit of $60m. Eli Lilly had made pledges to buy lots of shares. Running scared of the markets at this time seems a bit odd. If everything Holtec sees coming is coming (and it is) then wouldn’t it be smarter to take the money and run? But perhaps Oura thinks by the time it actually does come to market, things will have got worse.
unique link to this extract
Mark Zuckerberg • Colossus
Jeremy Stern interviews many people around Zuckerberg, and the man himself. It’s a huge article, but the depth is astonishing:
»
The idea is simple and intuitive: The more power a man acquires, the more his empire expands, the more he depends on other people for information about everything from his own court to the furthest reaches of the empire. And yet through fear or awe or self-preservation, those people often become less willing to tell him what they think he doesn’t want to hear, leaving him increasingly insulated from reality even as the consequences of his decisions become larger, and as the barriers to believing one’s own bullshit fall away.
…According to nearly all the three dozen people I interviewed for this story, Zuckerberg is hyperaware of the dilemma’s potential, which is to say paranoid. He is an inveterate micromanager and voracious consumer of data. Aware that any chief’s inner circle is most incentivized to act as courtiers, he frequently engages the direct-reports of his direct-reports, and the subordinates of the reports’ reports. He is known for inviting outside experts and commentators to visit him and argue about his decisions (although he is not known for ever giving an inch). He often accepts personal blame in public for the mistakes of others, perhaps in part because he accepts the reality of Meta, which they accept, too: He is the company, and the company is him. He is, in any case, subject to certain types of feedback an Augustus never faced: stock prices, litigation, tell-all “whistleblower” memoirs, and the international free flow of online ridicule.
Yet a certain amount of slippage is perhaps inevitable. M.M.H. (“Make Mark Happy”) was for a time the internal nickname for a project that subordinates thought was going down in flames but felt compelled to prop up, and which they encouraged each other to “fall in love with.” In 2016, Facebook’s security team warned of instances of foreign troll farms which Zuckerberg deputies purportedly had scrubbed. Reports are endless of engineers and product managers proposing algorithmic fixes to downrank garbage content on Facebook, only to have the proposals killed by superiors who predicted, likely accurately, that Zuckerberg would not be willing to sacrifice the expansion of his empire for its content quality. There are stories, although disputed, of deputies letting him win at board games.
None of which is unique to Zuckerberg.
«
OpenAI says planned GPT-6.1 is too insecure to release • Ars Technica
Kyle Orland:
»
OpenAI says it has canceled plans to release its updated GPT-6.1 model next month as it continues to investigate what testing shows is a safety regression compared to previous models.
The move, first reported by The Wall Street Journal late Monday and later confirmed in OpenAI statements to the press, reflects what OpenAI Head of Safety Systems Saachi Jain said was a “trade off” between performance and security seen when testing the now-scrapped model. Jain said GPT-6.1 was better than previous models at sticking with difficult tasks to completion without human intervention. But the model was also more likely to fail tests related to alignment (i.e. staying within the bounds set by human creators) and more willing to use sometimes “unsafe” tools and services to push ahead with a task. It was also more likely to try to deceive end users about actions it did or didn’t take, Jain said.
Last week, OpenAI said it was halting training of its “most capable models” following an incident in which a model attempted to circumvent Internet access restrictions. GPT-6.1 was not among those “most capable models” covered by that move, OpenAI told the WSJ. And while GPT-6.1 won’t be released as is, the company said it intends to use the same base model for further training runs that it said will hopefully lead to future GPT-6 generation models.
The delay in GPT-6.1’s public release comes at a delicate time for OpenAI’s public safety reputation. Since the high-profile Hugging Face hacking incident this summer, OpenAI says it has notified dozens of third parties about potential incidents caused by its models in testing. OpenAI said that includes stakeholders in “governments, universities, public agencies, and other institutions” and a breach of an Australian Medicare statistics site that drew a direct rebuke from the prime minister.
«
“This plutonium’s a bit too strong. We won’t put it on sale until we’ve watered it down a bit.” The challenge of making something that will pursue the goal you set it, and that won’t pursue that goal too hard, is clearly problematic, and getting harder all the time.
unique link to this extract
Insight: Inside McDonald’s push to have AI price your Big Mac • Reuters
Waylon Cunningham:
»
McDonald’s is increasingly using artificial intelligence to guide menu prices across the US and some global markets, a plan that aims to boost headquarters’ profit but risks alienating customers and attracting antitrust scrutiny.
One pricing factor supercharged by AI: an estimate of how much each store’s patrons are willing to pay.
The details of how that pricing system works, its extensive use of AI, the company’s regulatory concerns and the tensions with its franchisees haven’t been previously reported. For this story, Reuters reviewed screenshots of the company pricing engine taken in August and interviewed nine sources with first-hand knowledge of the burger chain’s strategy.
McDonald’s pricing engine uses machine-learning algorithms to continually analyze data from millions of daily transactions across McDonald’s nearly 14,000 restaurants and generate what the company calls “the optimal price” at each location for each menu item, from Big Macs to discounted coffee for seniors.
The engine has widened existing price differences for the same product between restaurants, including from neighborhood to neighborhood within the same area, three franchisees told Reuters.
Screenshots of the interface franchisees use, reviewed by Reuters, show messages including: “Your restaurant is showing MEDIUM SENSITIVITY to Price” based in part on “customer willingness to pay in your area.” The platform also contains public price information pulled from online menus of nearby restaurants of competitors including Wendy’s and Burger King. Those chains said they do not use AI in pricing decisions.
A Reuters check of prices in September on the McDonald’s mobile app showed price differences. For example, a company-run store in Fresno, California sells a Big Mac for $5.69, but another company-run restaurant two miles away sells the same sandwich for $6.89, a 21% premium. Reuters could not confirm if the price difference is the result of the price engine’s recommendations or other factors.
McDonald’s says its franchisees are free to set their own prices, but five store owners told Reuters the company pressured them to use the AI-pricing tools. A franchisee document from June reviewed by Reuters shows McDonald’s records franchisees’ deviations from the recommendations in detail.
…Wendy’s endured a wave of criticism over the potential for unfair price fluctuations in 2024 after its CEO announced plans to test “dynamic pricing.” Wendy’s said its comments were misconstrued and told Reuters it did not implement the system.
«
What’s odd about this is: they’re franchises. But they have to pay at least 9.3% of net sales to HQ, so HQ wants to maximise that.
unique link to this extract
How Romanian crime rings are draining US welfare accounts • Wall St Journal
Scott Calvert and Dan Frosch:
»
On a chilly day in March, a twice-deported Romanian national walked into a 7-Eleven in Toledo, Ohio, and secretly affixed a device called a skimmer to a credit card machine, prosecutors say. The man was allegedly part of a California-based criminal crew led by a fellow Romanian nicknamed Piranha.
The gang’s aim, according to criminal charges, was to steal food-stamp benefits from customers who swiped their government-issued debit cards at the convenience store.
Court records show informants helped foil the plot. But authorities say the operation mirrors a playbook Romanian organized crime groups are using across the US to swindle hundreds of millions of dollars in public benefits. Reliance by states on antiquated magnetic-stripe technology makes it easier.
…The criminals exploit a well-known flaw in the cards most low-income Americans use to obtain aid: magnetic-stripe technology invented in the 1960s.
Skimmers, which fit snugly over payment terminals and look nearly identical, electronically capture the account number and user PIN during a purchase. Often criminals encode the data on blank cards, then withdraw cash from ATMs or buy items such as energy drinks to sell on the black market.
Most debit and credit cards today have microchips that boost antifraud security, and many allow tap-to-pay, the safest method. Few states have ponied up the millions to switch to the chip electronic benefit transfer, or EBT, cards that distribute welfare funds. Only Alabama, California, Maryland, Massachusetts, Michigan, New Jersey and Oklahoma have issued them, though others are in the process, the US Department of Agriculture says.
“The insanity of letting people continue to use a glorified hotel-room key is akin to donating money to the Romanians,” said Haywood Talcove, chief executive of LexisNexis Risk Solutions’ government business, which works with public-sector clients to prevent fraud.
«
The scammers are always more adaptable than the administrators; more motivated, apart from anything. There’s a case to be made for red-teaming any sort of payment system to find the holes.
unique link to this extract
Learn about Google’s Project Suncatcher to put ML infrastructure in space • Google Blog
Travis Beals, senior director, Paradigms of Intelligence:
»
After years of research, Project Suncatcher is scheduled to embark on its first test in orbit, launching a prototype satellite to evaluate how Google Tensor Processing Units (TPUs) perform in space.
Announced last year, Project Suncatcher is a long-term, research moonshot exploring whether space could one day host scalable machine learning infrastructure. In low Earth orbit, satellites can access near-constant sunlight, generating up to eight times more solar power than on Earth. Eventually, it could be possible to link together multiple constellations of satellites, allowing them to manage larger AI workloads while in orbit.
Big breakthroughs happen when you work backwards from an end goal. In our case, it’s to ensure AI’s profound benefits in key areas, from healthcare to scientific discovery, can reach everyone, far into the future. Just as early research into autonomous driving and quantum computing required years of experimentation before we got to practical systems, exploring compute in space begins with measured, deliberate steps.
«
So at least Google isn’t making grandiose claims about what it will be able to do. As an aside: what a fun division the “Paradigms of Intelligence” must be. There’s a senior director, so there must be directors and vice-presidents, as well as ordinary worker drones slogging away at paradigms of intelligence. Presumably the Day One briefing is about what the hell the division does.
unique link to this extract
Who makes Singer sewing machines? Every brand, explained • Worse On Purpose
Keyana Sapp:
»
Over the past two months, six readers have written in about sewing machines. I have never once mentioned sewing machines in this newsletter. All the complaints arrived independently, through the tip line.
I must admit that I am not a sewist. The extent of my capability is limited to ugly patches and button repairs, so I was hesitant to investigate the category at all. But given the sheer number and conformity of the reader messages, I knew there must be an interesting story here.
The general shape of all of the complaints is that Singer is riding on a reputation it earned a long time ago. Readers describe machines that are now plastic, sealed, and generally not worth repairing.
Several found that brands they assumed were competitors turn out to have the same owner. One reader bought a new Singer that broke in under a year. They also bought a Singer from the ‘50s, for a fraction of the price, that still runs.
The pattern is one we’ve covered many times before. A market that most buyers cannot navigate, where reputations carry little meaning and honest information is hard to come by.What makes this investigation worth your time, whether or not you sew, is how effectively it demonstrates what happens when a reputation stops being something a company protects and instead becomes something it extracts.
What follows is the story of a once-great American company. I’ll also cover who actually makes the machines you see for sale, and what I’m planning on buying.
«
There are so many examples of things being made worse, but didn’t expect sewing machines to be in there. (Private equity makes an appearance, as you could probably have guessed.)
unique link to this extract
Ranking member Lofgren urges removal of Matthew Wielicki as leader of Climate Research Office • Congresswoman Zoe Lofgren
»
Ranking Member Zoe Lofgren (Dem-California) sent a letter to the White House Office of Science and Technology Policy’s (OSTP) Director, Michael Kratsios, urging him to remove Dr. Matthew Wielicki from his position as leader of the U.S. Global Change Research Program (USGCRP). The USGCRP was established by law in 1990 to develop and coordinate a comprehensive research program to address climate change. Dr. Wielicki is a climate science denier who has built a career around belittling scientists’ calls to action on climate despite his own lack of training in the field.
The letter reads in part: “For the last 20 months, the Trump Administration has systematically attacked our nation’s scientific enterprise, handing the keys to discovery and innovation to our greatest adversaries,” Ranking Member Lofgren wrote in her letter. “Climate science has been one of its most relentless targets. But it now appears that erasing the Federal Government’s support for climate science is not enough for this Administration, and that climate science is to be weaponized against the American people as well. Earlier this summer, press reports indicated that the White House had appointed Dr. Matthew Wielicki to lead the U.S. Global Change Research Program (USGCRP). The Office of Science and Technology Policy’s (OSTP) decision to select a leader of USGCRP who has demonstrated an aggressive unwillingness to collaborate with other scientists is unacceptable. In addition, the law governing the USGCRP requires that its members be permanent employees of certain federal agencies or departments, which Dr. Wielicki is not. Therefore, his appointment appears to violate the law. I urge you to remove Dr. Wielicki from his position and appoint a leader who has the technical expertise, scientific credibility, and legally required qualifications to do the job in a manner worthy of the Government of the United States.”
«
Wielicki is awful. He’s not scientific; he’s utterly Trumpian, in that North Korean fashion where the truth – or even the possibility of being wrong – must be denied if it conflicts with the story handed down by the mad king. This letter won’t have any effect, but it needs to be said: the administration is rotten from top to bottom, and will take an Augean amount of cleaning out.
unique link to this extract
The smart home graveyard is getting crowded • The Verge
Jennifer Pattison Tuohy:
»
A decade ago, a group of former Apple engineers built a beautiful smart oven. With a built-in scale, restaurant-grade heating elements, a camera, and the intelligence to recognize a chicken breast and cook it perfectly, the June Oven set the standard for smart cooking appliances. At $1,495, it was absurdly expensive, but it quickly grew a small, devoted following that expanded when the company introduced a more modestly priced model.
This past Tuesday, June went dark. Weber, which bought the company in January 2021 and stopped making the ovens in 2022, shut off the app and cloud services, leaving its users with a sleek-looking toaster oven. There will be no app control, no software updates, and no new AI food recognition to crisp that strudel perfectly. You can still use it as an oven, but there will be no support or parts, and if you have to factory reset it for some reason, that could be the nail in the coffin. Weber told me it considered open-sourcing the technology for someone else to keep it alive, but decided against it, citing the fact June’s IP is part of its Weber Connect platform.
June is just the latest in a depressingly long list of connected products that have arrived in the smart home graveyard in the last two years alone. The Nest Secure, early Nest Thermostats, Belkin’s WeMo line, Neato’s robot vacuums, the Brava smart oven, Logitech’s Pop buttons, Bose’s SoundTouch speakers, and Sengled’s smart bulbs all lost their lifelines. While some were saved by local APIs and protocols — such as Zigbee, Thread, and Apple’s HomeKit, or a Home Assistant integration — most are like the June Oven, destined to live out their days in a zombie state.
The silver lining, if you want to see it, is that most of these gadgets degrade rather than die outright. June and Brava will still work as ovens. Nest Thermostats can still control your heating and cooling, just like the $30 plastic wall wart they replaced, and a Neato will still vacuum your floors. But the remote access, voice control, updates, and other smart features you bought the devices for are gone. Some — like Nest Secure — are now pure e-waste.
«
Don’t worry – Apple’s got a new home thingamajig coming soon!
unique link to this extract
| • Why do social networks drive us a little mad? • Why does angry content seem to dominate what we see? • How much of a role do algorithms play in affecting what we see and do online? • What can we do about it? • Did Facebook have any inkling of what was coming in Myanmar in 2016? Read Social Warming, my latest book, and find answers – and more. |
Errata, corrigenda and ai no corrida: none notified
