Why did the Ingenuity helicopter crash on Mars? An investigation blames bland terrain. CC-licensed photo by Kevin Gill 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 10 links for you. Grounded. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.
AI could be making scientists less creative • Gizmodo
Todd Feathers:
»
Researchers at the University of Chicago and Tsinghua University, in China, analyzed nearly 68 million research papers across six scientific disciplines (not including computer science) and found that papers incorporating AI techniques were cited more often but also focused on a narrower set of topics and were more repetitive. In essence, the more scientists use AI, the more they focus on the same set of problems that can be answered with large, existing datasets and the less they explore foundational questions that can lead to entirely new fields of study.
“I was surprised at the dramatic scale of the finding, [AI] dramatically increases people’s capacity to stay and advance within the system,” said James Evans, a co-author of the pre-print paper and director of the Knowledge Lab at the University of Chicago. “This suggests there’s a massive incentive for individuals to uptake these kinds of systems within their work … it’s between thriving and not surviving in a competitive research field.”
As that incentive leads to a growing dependence on machine learning, neural networks, and transformer models, “the whole system of science that’s done by AI is shrinking,” he said.
The study examined papers published from 1980 to 2024 in the fields of biology, medicine, chemistry, physics, materials science, and geology. It found that scientists who used AI tools to conduct their research published 67% more papers annually, on average, and their papers were cited more than three times as often as those who didn’t use AI.
«
Bird flu jumps from birds to human in Louisiana; patient hospitalized • Ars Technica
Beth Mole:
»
A person in Louisiana is hospitalized with H5N1 bird flu after having contact with sick and dying birds suspected of carrying the virus, state health officials announced Friday.
It is the first human H5N1 case detected in Louisiana. For now, the case is considered a “presumptive” positive until testing is confirmed by the Centers for Disease Control and Prevention. Health officials say that the risk to the public is low but caution people to stay away from any sick or dead birds. A spokesperson for Louisiana’s health department told Ars that the hospitalized patient had contact with both backyard and wild birds.
Although the person has been hospitalized, their condition was not reported. The spokesperson said the department would not comment on the patient’s condition due to patient confidentiality and an ongoing public health investigation.
The case is just the latest amid H5N1’s global and domestic rampage. The virus has been ravaging wild, backyard, and commercial birds in the US since early 2022 and spilling over to a surprisingly wide range of mammals. In March this year, officials detected an unprecedented leap to dairy cows, which has since caused a nationwide outbreak. The virus is currently sweeping through California, the country’s largest dairy producer.
To date, at least 845 herds across 16 states have contracted the virus since March, including 630 in California, which detected its first dairy infections in late August.
«
Um, just a watching brief.
unique link to this extract
Google’s lost moonshots • Jerry Liu
The aforesaid Liu has just spent six years in consulting and startups, having previously worked at Facebook/Meta:
»
1. Misaligned Incentives
Google’s innovation machine is driven by PM careers, and PM careers are driven by metrics. If you’re a product manager at Google, what’s your incentive? Ship something small that looks good on your performance review, or spend years on a project that might fail spectacularly? It’s like trying to work on decade-spanning climate change projects with politicians who need to win the next election and have term limits of 4 years. You take your wins, and you get out before the bridge collapses. Which inevitably it will, because can you expect any human project to only ever be winning, quarter after quarter?2. Moonshot-scale Budget
This is crucial: moonshot-scale problems need moonshot-scale resources. Think about how VCs fund startups. When something shows promise, not only do they need more funding, the fundraising often jumps by orders of magnitude. What Google calls moonshots often feel more like well-funded experiments. Meanwhile, Meta commits resources at a scale that matches their ambitions. Look at Reality Labs – they’ve burned more money than most companies will ever see, but they keep going.3. Institutional Learning
And maybe the hardest problem: institutional learning. Both companies fail, but they fail differently. When Google Glass flopped, what happened to all that knowledge? Sure, some of it probably lives in internal docs, but the teams scattered, the context was lost, and the deep learning – the kind that only comes from failure – largely evaporated. I also heard that Google Plus’s assets were also cannibalized internally; Please let me know if you’ve interacted with any part of Google Plus’s remains recently.«
This is very true: whatever happened to Loon, the balloon thing, and all the other moonshots? Quietly fallen to earth, it seems.
unique link to this extract
Ingenuity Mars helicopter January grounding: what happened? • The Register
Richard Speed:
»
It appears the bland Martian surface triggered a chain of events that left NASA’s Ingenuity helicopter permanently grounded on the red planet.
The helicopter’s flying career came to an abrupt end earlier this year when Flight 72 was cut short, and communications were briefly lost. After re-establishing contact, it soon became clear Ingenuity would not be flying again – the rotor blades were damaged, and one was entirely detached.
At the time, the prevailing theory was that the flight ended when Ingenuity’s downward-facing camera could not pick out features on the surface. According to the NASA Jet Propulsion Laboratory (JPL), this is still the most likely scenario for what started a chain of events that left the helicopter crippled.
Performing an air crash investigation from hundreds of millions of kilometers away is tricky. It’s impossible to get hands on the wreckage, there are unlikely to be any witnesses, and there aren’t brightly colored black boxes to give clues about what happened in the final minutes of the flight.
What there is, however, is telemetry. Data sent during the final flight indicates that around 20 seconds after take-off, Ingenuity’s navigation system couldn’t find enough surface features to track. It was designed to operate over textured, flat terrain, not the steep, featureless sand ripples where it ultimately met its demise.
“Photographs taken after the flight indicate the navigation errors created high horizontal velocities at touchdown,” according to JPL. Engineers reckon the most likely scenario is that Ingenuity made a hard landing on the slope of a sand ripple. The sudden pitch and roll exerted stress on the rotor blades past their design limits, and all four snapped at their weakest point.
«
Making “social” social again: announcing Mozi • Medium
Ev Williams did Twitter, did Medium, and now he’s doing a sort of.. travel-contact app:
»
Mozi is a social app — not in the sense of “social media.” But in the sense of interacting with other people and building relationships.
In fact, it’s not a media app at all. There is no posting photos or videos or liking or following. There are no influencers — except your friend who may influence you to meet up for a coffee when you’re in town.
The primary value proposition of Mozi (today) is simple: It lets you know when you’re going to be in the same place (city or event) as someone you know. And the goal is straightforward: to connect more often—and in person—with the people you care about.
For example, I just got back from Miami. Before going, I put my plan (just the city and what days) into Mozi. This information was shared just with my contacts (minus any I wouldn’t want it to be). So, even before going, I was able to see both the people I know who live there and other friends who were visiting at the same time, so we could meet up and make plans.
Mozi also helps you decide where to go. “Events” on Mozi (currently a beta feature) lets you see who you know may be going—or considering going—to a conference or event before you go. (If you happen to be going to SXSW, join the Mozi event. I’ll be there too.)
«
Not in a burning hurry to try this, to be honest. It’s reminiscent of Foursquare, but without the gamification; it’s a sort of private shared-only-with-contacts-you-want-to meetup app. I wonder about the mental load of having to choose which contacts to share with; what you really want is to see who’s in the city you’re going to and include or exclude on that basis. You can get it now for iOS. (Android is of course on a wait list.)(Thanks Q for the link.)
unique link to this extract
The ‘Ghost Gun’ linked to Luigi Mangione shows just how far 3D-printed weapons have come • WIRED
Andy Greenberg:
»
More than a decade after the advent of the 3D-printed gun as an icon of libertarianism and a gun control nightmare, police say one of those homemade plastic weapons has now been found in the hands of perhaps the world’s most high-profile alleged killer. For the community of DIY gunsmiths who have spent years honing those printable firearm models, in fact, the handgun police claim was used to fatally shoot UnitedHealthcare CEO Brian Thompson is as recognisable as the now-famous alleged shooter himself—and shows just how practical and lethal those weapons have become.
In the 24 hours since police released a photo of what they say is Luigi Mangione’s gun following the 26-year-old’s arrest Monday, the online community devoted to 3D-printed firearms has been quick to identify the suspected murder weapon as a particular model of printable “ghost gun”—a homemade weapon with no serial number, created by assembling a mix of commercial and DIY parts. The gun appears to be a Chairmanwon V1, a tweak of a popular partially 3D-printed Glock-style design known as the FMDA 19.2—an acronym that stands for the libertarian slogan “Free Men Don’t Ask.”
The FMDA 19.2, released in 2021, is a relatively old model by 3D-printed-gun standards, says one gunsmith who goes by the first name John and the online handle Mr. Snow Makes. But it’s one of the most well-known and well-tested printable ghost gun designs, he says.
…The fact that even a relatively old model of 3D-printed firearm allegedly allowed the killer to shoot Thompson repeatedly on a Manhattan street—certainly the most high-profile shooting ever committed with a ghost gun or a 3D-printed weapon—shows how far DIY weapons tech has come, says Cody Wilson, the founder of the gun rights group Defense Distributed. Unlike the earliest 3D-printed gun models, the FDMA 19.2 can be fired hundreds or even thousands of times without its plastic components breaking.
«
Originally the makers of these guns weren’t sure if they would explode when fired. Now, they’re more confident. (I’ve slightly tweaked the original text to avoid any assumptions about the identity of the killer and ownership of the gun: both are crucial to the case.)
unique link to this extract
Is doom scrolling really rotting our brains? The evidence is getting harder to ignore • The Guardian
Siân Boyle:
»
Brain rot was portended almost 20 years ago when scientists studied the effects of a new invention called “email”, specifically the impact a relentless barrage of information would have on participants’ brains. The results? Constant cognitive overload had a more negative effect than taking cannabis, with IQs of participants dropping an average of 10 points.
And this was prior to smartphones bringing the internet to our fingertips, which has resulted in the average UK adult now spending at least four hours a day online (with gen Z men spending five and a half hours a day online, and gen Z women six and a half).
In recent years, an abundance of academic research from institutions including Harvard medical school, the University of Oxford and King’s College London found evidence that the internet is shrinking our grey matter, shortening attention spans, weakening memory and distorting our cognitive processes. The areas of the brain found to be affected included “attentional capacities, as the constantly evolving stream of online information encourages our divided attention across multiple media sources”, “memory processes” and “social cognition”.
Paper after paper spells out how vulnerable we are to internet-induced brain rot. “High levels of internet usage and heavy media multitasking are associated with decreased grey matter in prefrontal regions,” finds one. People with internet addiction exhibit “structural brain changes” and “reduced grey matter”. Too much technology during brain developmental years has even been referred to by some academics as risking “digital dementia”.
In 2018, a decade of data analysed by leading memory psychologists at Stanford University found that people who frequently engaged with multiple online platforms have reduced memory and attention spans.
And yet we seem to be doing very little to stem the tide.
«
BBC complains to Apple over misleading shooting headline • BBC News
Graham Fraser:
»
The BBC has complained to Apple after the tech giant’s new iPhone feature generated a false headline about a high-profile murder in the United States.
Apple Intelligence, launched in the UK earlier [last] week, external, uses artificial intelligence (AI) to summarise and group together notifications.
This week, the AI-powered summary falsely made it appear BBC News had published an article claiming Luigi Mangione, the man arrested following the murder of healthcare insurance CEO Brian Thompson in New York, had shot himself. He has not.
A spokesperson from the BBC said the corporation had contacted Apple “to raise this concern and fix the problem”. Apple declined to comment.
“BBC News is the most trusted news media in the world,” the BBC spokesperson added. “It is essential to us that our audiences can trust any information or journalism published in our name and that includes notifications.”
The notification which made a false claim about Mangione was otherwise accurate in its summaries about the overthrow of Bashar al-Assad’s regime in Syria and an update on South Korean President Yoon Suk Yeol.
But the BBC does not appear to be the only news publisher which has had headlines misrepresented by Apple’s new AI tech. On 21 November, three articles on different topics from the New York Times were grouped together in one notification – with one part reading “Netanyahu arrested”, referring to the Israeli prime minister.
It was inaccurately summarising a newspaper report about the International Criminal Court issuing an arrest warrant for Netanyahu, rather than any reporting about him being arrested.
«
Apple would have expected that there would be screwups with Apple Intelligence, but it’s hard to see how it prevents this sort of mangling from happening.
unique link to this extract
iOS 18.2’s new Mail app is nice, but I disabled one of its main features • 9to5Mac
Michael Burkhardt:
»
With iOS 18.2, Apple introduced an all new Mail app. It introduced mail categorization, a fresh coat of paint, contact photos/business logos for conversations, a new system for grouping emails, and more. All of that sounded nice when it was unveiled back at WWDC, but now that I’ve actually spent some time using it, I’m having some doubts.
One of the biggest features in the new Mail app is categorization, breaking down your emails into varying categories of Primary, Transactions, Updates, and Promotions.
This all sounds nice in concept, since it’d declutter your inbox, and the Primary tab would contain everything that’s important. In practice though, a lot of things were incorrectly categorized, and I found myself swiping over to the “All Mail” tab most of the time, that way I could see everything without having to deal with inaccurate sorting.
And yes, you can choose to recategorize senders if you don’t like how Apple chose to sort it. However, I find that a bit tedious compared to simply turning categorization off entirely.
Apple thought about the fact that everyone might not necessarily like categorization, and provided a simple way to disable it.
«
I think “features” like this would get in the way of using the machine. Google does this with Gmail, and I truly don’t like that either. So far, nothing in 18.2 (which I haven’t installed) looks utterly compelling.
unique link to this extract
Apple plans thinner, foldable iphones to revive growth • WSJ
Aaron Tilley and Yang Jie:
»
Starting next year, Apple plans to introduce an iPhone that will be thinner than the approximately 8-millimeter profile of current models, said people familiar with the company’s plans. The model is intended to be cheaper than Pro models, with a simplified camera system to reduce costs.
The company is also planning two foldable devices, the people said. A larger device, intended to serve as a laptop, would have a screen that unfolds to be nearly as large as some desktop monitors, at about 19 inches. A smaller model would unfold to a display size that would be larger than an iPhone 16 Pro Max, intended to serve as a foldable iPhone, the people said.
Both foldable designs have been in development for years, but some key parts weren’t ready. Major challenges included improving the hinge, a mechanism that allows the device to fold and unfold, and the display cover, a flexible material protecting the foldable screen.
Current foldable phones on the market aren’t thin, light or energy-efficient enough to meet Apple’s standards, which is why Apple has been slower to enter this segment, said Jeff Pu, an analyst with Hong Kong-based brokerage Haitong International Securities.
Apple experimented with other different designs, such as having a display on the outside of the device when it is folded, but it now favors an inward-folding design, people familiar with the devices said.
Although Apple initially aimed to introduce the larger device first to gauge market response, it now appears that the foldable iPhone will likely be ready ahead of it. Apple executives are pushing for a 2026 release, but the company may need another year to address technical challenges, the people said.
«
So a foldable iPad and an inwardly-folding phone. But far enough away that they might have been “held up” by “technical challenges”.
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
The Overspill: when there's more that I want to say 