A fault with some LEDs used in streetlights is turning cities purple. CC-licensed photo by KuraybaKurayba 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.
On Friday, there’s another post due at the Social Warming Substack at about 0845 UK time.
A selection of 9 links for you. Puple reign, indeed. I’m @charlesarthur on Twitter. Observations and links welcome.
Mike Martin is a military analyst:
If the 20th century was the industrial century, the 21st is the computing century. Microchips are becoming much smaller, much faster, and much cheaper. Software is replacing hardware as the main focus of innovation. Real-life problems (like how to detect cancer, fold proteins, or regulate traffic in cities) are starting to be solved by algorithms (artificial intelligence.) And finally, all of these individual bits of technology are being networked so they can talk to each other.
This is having profound effects on military technology, how governments think about generating future capabilities, and who has access to them.
Firstly, the big 20th-century systems look very vulnerable to attacks from multiple, cheap, small unmanned systems. Imagine an aircraft carrier [typical cost $6bn] beset by a swarm of micro suicide drones, some in the air, some underwater, all networked together with distributed hive processing, so that the drone swarm itself reacts to the aircraft carrier’s countermeasures by reshaping as the engagement unfolds. All this at a cost of one hundredth or one-thousandth of the carrier’s cost.
Secondly, the entire defence industry is modeled to provide big, expensive systems to governments that take years or decades to procure, build, and commission. Now, however, a startup can code the software that creates an underwater drone swarm, buy the processors commercially, and get it to market in a couple of years. It’s not clear whether the defence industry — and the generals and admirals who grew up in an era of big systems — are responding to these changes fast enough. What’s more certain is a significant reordering of defence industries of the world over the next decade.
Lastly, the miniaturization of force leads to the democratization of force. Cheaper, smaller, commercially available technologies mean that fewer wealthy countries, as well as a plethora of non-state actors, can once again get into the big league. US forces in Syria, for instance, are regularly attacked with tiny suicide drones, and the Ukrainians are buying quadcopters from Amazon and modifying them to drop bombs on Russian forces.
This, more than anything else, will change how and who we fight in the coming wars.
Since a Bloomberg report last year that it held risky investments like short-term loans to Chinese companies, Tether said it has shifted its money into safer things like government bonds. A September report, prepared by Tether’s auditor and meant to reassure customers, showed that more than 80% of its $68bn was in fairly safe, liquid stuff — $40bn in U.S. Treasurys, $7bn in money-market funds, and $6bn in cash. (That list of assets is smaller today after a wave of redemptions.)
The rest, though, is in investments that are harder to value and sell, and about which Tether shares very little. It owns some $6bn in loans secured by its own coins, a spokesman confirmed to WSJ last week. A loss of confidence in Tether, like the one that hit FTX’s token, would reduce that collateral to zero, taking 10% of Tether’s assets with it.
It also owns $2.6bn in “other investments,” according to the September report. It’s not totally clear what’s in them, but they are likely venture stakes in other crypto companies held by its owners and affiliates, according to a global investigations firm commissioned by a hedge fund betting against the price of Tether. Semafor reviewed the findings of its report that found Tether holds equity stakes in more than a dozen crypto startups.
Semafor was able to verify some but not others. We confirmed that the crypto exchange that owns Tether invested, either through itself or an affiliate, into: an online-betting site called Betfinex; Dazaar, a data-sharing service; Dusk Networks, whose software turns financial investments into tokens; a crypto trading platform called Rhino; Shape Shift, a crypto wallet; Blockstream, a blockchain infrastructure company; Netki, a digital-ID company; and Keet.io, a video-chat app.
Any honest assessment of that $2.6bn “other investments” portfolio would likely mean it is worth less today than it was in September.
And as token-holders ask for their money back, Tether has to sell the stuff it can — government bonds, corporate bonds, money-market positions. That means the stuff it can’t sell — namely, venture investments – will start to make up a larger portion of its assets. This is how runs on banks start.
Wouldn’t be too confident about any of those big numbers that the auditor provided. Tether hasn’t been properly audited, ever. Something feels like it’s about to shift.
unique link to this extract
David Pan and Naureen Malik:
soaring energy costs, a sharp decline in Bitcoin prices and more competition have compressed profit margins and made it difficult for miners to repay debt. Some are on the verge of bankruptcy.
“There are just tons of assets everywhere, it’s like a mess.” said Mason Jappa, chief executive at Austin, Texas-based crypto-mining service firm Blockware Solutions. “I got messages about transformers, switch gears, and mobile data centers and containers for mining, they are just sitting there.”
There are a lot of losers if the Bitcoin mining industry goes bust. For one, local authorities provided incentives such as tax abatements that reached into the tens of millions of dollars. The power generation planned that the region sorely needs to avoid another energy crisis may not materialize. Some developers made hefty investments to build out Bitcoin mining facilities. The average cost to have one-megawatt capacity of mining infrastructure is currently around $300,000 in the state, the high end of the range, according to Jappa.
…After China banned crypto miners last year, Texas sought to fill the gap as a way to add fuel to the state’s fast-growing economy. But because mining hinges on power consumption, the wave of new demand threatens to stress a grid still trying to recover from failures during an extreme winter storm in February 2021 that left millions in the dark for days and more than 200 people dead.
Venture capitalist Tim Draper thinks bitcoin will hit $250,000 a coin by the middle of 2023, even after a bruising year for the cryptocurrency marked by industry failures and sinking prices.
Draper previously predicted that bitcoin would top $250,000 by the end of 2022, but in early November, at the Web Summit tech conference in Lisbon, he said it would take until June 2023 for this to materialize.
He reaffirmed this position Saturday when asked how he felt about his price call following the collapse of FTX.
“I have extended my prediction by six months. $250k is still my number,” Draper told CNBC via email.
Bitcoin would need to rally nearly 140-fold from its current price of around $17,000 for Draper’s prediction to come true. The cryptocurrency has plunged over 60% since the start of the year.
…Draper’s rationale for bitcoin’s breakout next year is that there remains a massive untapped demographic for bitcoin: women. “My assumption is that, since women control 80% of retail spending and only 1 in 7 bitcoin wallets are currently held by women, the dam is about to break,” Draper said.
Crypto has long had a gender disparity problem. According to a survey conducted for CNBC and Acorns by Momentive, twice as many men as women invest in digital assets (16% of men vs. 7% of women).
Errrr. This feels a bit like the obverse of the confident prediction that “makeup sales are just about to boom, we just need to get the other 50% of the population to buy it!” that has been heard down the years. Crypto is recapitulating everything. Anyway, see you in June, Mr Draper. (Do read the story for some of the VC bets Draper has made, and make your decision about how good he is at reading women.)
unique link to this extract
European Union privacy regulators have ruled that Meta shouldn’t require users to agree to personalized ads based on their online activity, according to people familiar with the decision, a ruling that could limit the data that Meta can access to sell such ads.
A board representing all EU privacy regulators on Monday approved a series of decisions ruling that EU privacy law doesn’t allow Meta platforms, such as Instagram and Facebook, to use their terms of service as a justification for allowing such advertising, the people said.
The new EU decisions can be appealed, which could lead to their being suspended pending potentially lengthy litigation. If upheld, though, they could make it harder for Meta and other platforms to show users ads based on what they tap and watch within those platforms’ own apps. Meta has for years allowed users to opt out of personalizing ads based on data from other websites and apps. But it hasn’t given any such option for ads based on data about user activity on its own platforms—such as which videos an Instagram user watches.
If any significant portion of its users opts out of such targeting, Facebook and Instagram would end up with less information with which to build audiences for the personalized ads that analysts and people close to the company say make up the bulk of its revenue.
Which is basically the power that Apple’s Ad Tracking Transparency (ATT) system offers, but without having to go through a zillion legal go-arounds.
unique link to this extract
The sky over the city of Vancouver was the color of a television tuned to a Prince concert.
OK, maybe not the whole sky. But enough of it that people noticed. A bunch of streetlights — a few hundred out of thousands — had suddenly changed. What had been moonshine white was now blue, or purple, or even violet. They weren’t any less bright, objectively speaking. But purple doesn’t exactly illuminate a sidewalk the way white does. The spectrum of Vancouver had taken a hard left turn. It didn’t look bad. It wasn’t unsafe, particularly. It was just weird.
So people placed worried calls to the city. And after all the hue and cry, Vancouver rolled out the utility trucks and set out to replace the chromatic aberrations — even though the lights were still pretty new. Like most other cities, Vancouver has spent the past few years switching from old sodium-vapor streetlights to LEDs. The new bulbs, basically arrays of computer chips that convert electricity to light, are cheaper, less power-hungry, and longer-lasting. LED streetlights are supposed to shine for the better part of a decade.
Unless they don’t. Because the Great Purpling didn’t start — or end — in Vancouver. Reports stretch back to 2020 and across the hemisphere — Wisconsin, North Carolina, Florida, New Mexico, California, even Ireland. “It’s something we began seeing about two years ago,” says Jeff Brooks, a representative for Duke Power, which is responsible for streetlights across the Carolinas and parts of Florida and the Midwest. “I’ve had people call and ask if this was because it’s Halloween, or because their football team in that area wears purple.”
It’s a bit overlong – it could have been a third of the length and done the job – but it’s entertaining enough. (Thanks wendyg for the link.)
unique link to this extract
Katherine J. Wu:
Perhaps the worst can be averted if the government does more to vaccinate the vulnerable and prep hospitals for a protracted influx of COVID patients; and if the community at large reinvests in a subset of mitigation measures as cases rise. “There is still the possibility that they may muddle through it without a mass die-off,” says Yanzhong Huang, a senior fellow for global health at the Council on Foreign Relations. “But even the most smooth and orderly transition,” he told me, “will not prevent a surge of cases.”
China represents, in many ways, SARS-CoV-2’s final frontier. With its under-vaccinated residents and sparse infection history, the nation harbors “a more susceptible population than really any other large population I can think of,” says Sarah Cobey, an computational epidemiologist at the University of Chicago. Soon, SARS-CoV-2 will infiltrate that group of hosts so thoroughly that it will be nearly impossible to purge again. “Eventually, just like everyone else on Earth, everyone in China should expect to be infected,” says Michael Worobey, an evolutionary virologist at the University of Arizona.
Whatever happens, though, China’s coming wave won’t recapitulate the one that swept most of the world in early 2020. Though it’s hard to say which versions of the virus are circulating in the country, a smattering of reports confirm the likeliest scenario: BF.7 and other Omicron subvariants predominate. Several of these versions of the virus seem to be a bit less likely than their predecessors to trigger severe disease. That, combined with the relatively high proportion of residents—roughly 95%—who have received at least one dose of a COVID vaccine, might keep many people from falling dangerously ill.
It could be a way to kill off a huge number of older citizens, many of whom have been very resistant to getting vaccinated. This is going to be brutal – and probably covered up, as India did. (Thanks G for the link.)
unique link to this extract
Elon Musk’s Neuralink, a medical device company, is under federal investigation for potential animal-welfare violations amid internal staff complaints that its animal testing is being rushed, causing needless suffering and deaths, according to documents reviewed by Reuters and sources familiar with the investigation and company operations.
Neuralink Corp is developing a brain implant it hopes will help paralyzed people walk again and cure other neurological ailments. The federal probe, which has not been previously reported, was opened in recent months by the US Department of Agriculture’s Inspector General at the request of a federal prosecutor, according to two sources with knowledge of the investigation. The probe, one of the sources said, focuses on violations of the Animal Welfare Act, which governs how researchers treat and test some animals.
The investigation has come at a time of growing employee dissent about Neuralink’s animal testing, including complaints that pressure from CEO Musk to accelerate development has resulted in botched experiments, according to a Reuters review of dozens of Neuralink documents and interviews with more than 20 current and former employees. Such failed tests have had to be repeated, increasing the number of animals being tested and killed, the employees say. The company documents include previously unreported messages, audio recordings, emails, presentations and reports.
…In all, the company has killed about 1,500 animals, including more than 280 sheep, pigs and monkeys, following experiments since 2018, according to records reviewed by Reuters and sources with direct knowledge of the company’s animal-testing operations.
Animal experimentation is a regrettable fact of life for many developments – particularly drug testing – but putting an emphasis on speed over accuracy and empathy (which seems to be a Musk trope) is not desirable. An investigation will slow everything down. It’s hard to know if that’s good or bad news for the animals, to be honest.
Upon downloading the app and opening the Magic Avatars feature, Lensa walks you through the process, explaining that the technology, which operates using the open source Stable Diffusion model, is not perfect and “may generate artifacts, inaccuracies, and defects in output images.”
Thankfully, I had already been warned by a friend who received a distorted image of himself with two heads, so I was primed for some odd results (which I did indeed receive, but more on that later).
The next step is to upload 10-20 photos. Lensa recommends close-up selfies with a variety of backgrounds, facial expressions, and angles to get the best results.
After submitting 10 photos, I was asked to indicate my gender as female, male, or other. I was then directed to a checkout page, where users are given the option to select from 50, 100, or 200 “unique avatars.” I opted for 50 for $3.99, which is half the regular cost as part of my free trial membership.
Once I made a payment, I was informed the process would take about 20 minutes, and I was given the option to receive a notification when the avatars were complete. I selected yes.
About 15 minutes later I received a push notification that my avatars were ready. My 50 avatars were delivered in 10 categories of 5 images including Iridescent, Light, Stylish, Anime, Cosmic, Fantasy, Kawaii, Pop, Focus, and Fairy Princess.
As expected, some of the results were very bizarre, while others made me feel quite beautiful. Many looked absolutely nothing like me, and in one I look vaguely like disgraced Theranos founder, Elizabeth Holmes, black turtleneck and all.
As my friend warned, I did receive a handful of distorted renderings with multiple limbs or heads, which was … the stuff of nightmares. But for a robot creating art in 15 minutes, it did a decent job. Ultimately, it was fun, though maybe not worth $3.99.
Amazing that all that computing resource to produce something you’d previously have had to hire a human Photoshop expert to perform, and which might have taken them a good few hours and probably cost you the thick end of a hundred pounds if they were doing it as a favour, is now dismissed as perhaps not worth the price of a cup of coffee. Moore’s Law is still around.
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