Start Up No.1941: how US Marines beat an AI camera, renewables far cheaper than coal, spotting the corporate fraudsters, and more


They certainly look like Airpods – but with counterfeiting Apple’s products being big business, do you know how to spot the fakes? CC-licensed photo by Maurizio Pesce 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.


There’s another post coming this week at the Social Warming Substack on Friday at about 0845 UK time. Free signup.


A selection of 10 links for you. Can you hear this? I’m @charlesarthur on Twitter. Observations and links welcome.


US Marines outsmart AI security cameras by hiding in a cardboard box (and other tricks) • PetaPixel

Pesala Bandara:

»

In the book, which will be released on February 28, Scharre recounts how the US Army was testing AI monitoring systems and decided to use the Marines to help build the algorithms that the security cameras would use.

They then attempted to put the AI system to the test and see if the squad of Marines could find new ways to avoid detection and evade the cameras.

To train the AI, the security cameras, which were developed by Defense Advanced Research Projects Agency’s (DARPA) Squad X program, required data in the form of a squad of Marines spending six days walking around in front of them. After six days spent training the algorithm, the Marines decided to put the AI security cameras to the test.

“If any Marines could get all the way in and touch this robot without being detected, they would win. I wanted to see, game on, what would happen,” DARPA deputy director Phil Root tells Scharre in the book.

Within a single day, the Marines had worked out the best way to sneak around an AI monitoring system and avoid detection by the cameras.

Root says: “Eight Marines — not a single one got detected.”

According to Scharre’s book, a pair of marines “somersaulted for 300 meters” to approach the sensor and “never got detected” by the camera. Meanwhile, two marines successfully evaded the camera by hiding in a cardboard box, a strategy they may have tried based on the video game series Metal Gear Solid. “You could hear them giggling the whole time,” Root says.

Meanwhile another smiling marine “field-stripped a fir tree and walked like a fir tree” toward the AI security camera and succeeded, too. Root tells Scharre that the marines were able to easily fool the cameras because “the AI system had been trained to detect humans walking, not humans somersaulting, hiding in a cardboard box, or disguised as a tree.

“So these simple tricks, which a human would have easily seen through, were sufficient to break the algorithm.”

«

So, there’s still going to be jobs for night shift guards. By the way, you’ve got the 4am-noon shift.
unique link to this extract


$50 AirPods Pro? Nope. Here’s how to spot fake Apple earbuds • WSJ

Dalvin Brown:

»

Caroline Ballard, 26 years old, was perusing Facebook Marketplace when she came across a pair of new AirPods Pro for $80, less than a third of the $249 list price.

She arranged to meet the seller at a nearby Walgreens parking lot the same day, a move that isn’t unusual for transactions made via Meta Platforms’s online market. The man pulled up and handed her a small plastic-wrapped white box with a photo of two AirPods on the top. The seller took her $80, got back in his car and drove off.

“I felt like something was off almost immediately,” she said. She noticed some of Apple’s branding missing from the packaging and quickly realized she had been scammed.

…Hunter Andrews, 32, paid a third-party vendor on Walmart’s website $199 for what he thought were AirPods Pro last January. The earbuds have been on sale for a similar price in the past from sellers such as Verizon Communications Inc.

Nothing seemed out of the ordinary at first. The box looked right and the AirPods paired with his iPhone, Mr. Andrews said. The U.S. Army sergeant, who lives in Knightdale, N.C., didn’t find out until he took them to an Apple Store for a replacement six months later.

“They were making a rattling noise,” Mr. Andrews said. “The guy at the Apple Store opens the case, and points to the Chinese writing on the inside. He goes, ‘These are fake.’” Mr. Andrews said he immediately reported the counterfeit AirPods to Walmart, which told him to contact the seller. The seller never refunded his money.

«

Surprise. Huge business in China making fakes. Apple stuff is a great target because it’s pricey, so people jump at cheap knockoffs.
unique link to this extract


US renewable energy farms outstrip 99% of coal plants economically – study • The Guardian

Oliver Milman:

»

Coal in the US is now being economically outmatched by renewables to such an extent that it’s more expensive for 99% of the country’s coal-fired power plants to keep running than it is to build an entirely new solar or wind energy operation nearby, a new analysis has found.

The plummeting cost of renewable energy, which has been supercharged by last year’s Inflation Reduction Act, means that it is cheaper to build an array of solar panels or a cluster of new wind turbines and connect them to the grid than it is to keep operating all of the 210 coal plants in the contiguous US, bar one, according to the study.

“Coal is unequivocally more expensive than wind and solar resources, it’s just no longer cost competitive with renewables,” said Michelle Solomon, a policy analyst at Energy Innovation, which undertook the analysis. “This report certainly challenges the narrative that coal is here to stay.”

The new analysis, conducted in the wake of the $370bn in tax credits and other support for clean energy passed by Democrats in last summer’s Inflation Reduction Act, compared the fuel, running and maintenance cost of America’s coal fleet with the building of new solar or wind from scratch in the same utility region.

On average, the marginal cost for the coal plants is $36 each megawatt hour, while new solar is about $24 each megawatt hour, or about a third cheaper. Only one coal plant – Dry Fork in Wyoming – is cost competitive with the new renewables. “It was a bit surprising to find this,” said Solomon. “It shows that not only have renewables dropped in cost, the Inflation Reduction Act is accelerating this trend.”

«

Actually, this article is worth clicking through to for the amendation at its end. It’s one of those “well, it depends which direction you’re going” sort of things.
unique link to this extract


The war over gas stoves hides easier ways to fix indoor air pollution • Vox

Kelsey Piper:

»

It’s absolutely worth trying to reduce indoor air pollution. But the cheapest, easiest way to do so, for most Americans, is to run your stove’s hood fan, or keep your windows open while cooking. Next on the list is to get a large air filter and run it continuously (we use Coway and BlueAir, based on a Wirecutter recommendation).

Air filters appear to improve respiratory health, improve heart health in the elderly, and reduce pollutants significantly, with an effect size that looks a lot larger than that associated with replacing a gas stove. (One drawback: air filters can’t completely filter the nitrogen oxides produced by gas stoves, which may make replacing your stove worth it for parents of children with asthma.)

For most of us, replacing your stove is an expensive step compared to the benefits you’ll get in cleaner air. And cost does matter: if we want to improve indoor air quality broadly, we should focus on the cheapest, most convenient interventions. Cooking with your hood fan on or the windows open costs nothing. Getting and continuously running a good air purifier in your home is relatively cheap, and it genuinely can make a difference in your health and especially the health of your small children, regardless of how you cook your food.

If you want to go ahead and swap out your gas stove for an induction stove, go for it. But if you’re freaked out about the possibility the air in your home is making your kids sick, start with the easy steps — and relax about the gas.

From a climate perspective, while gas stoves can leak methane, they’re a tiny fraction of methane emissions — only 3% of household gas emissions, and those household emissions are a small share of overall emissions. Trying to scare people about gas stoves for the sake of the climate means picking what is likely to be a politically unpopular fight, while passing up easier progress on more significant issues.

«

(Thanks G for the link.)
unique link to this extract


Link your Mastodon and Twitter accounts • Moa

»

Link your Mastodon account to Twitter with Moa Bridge. No bamboozle.

Enter your Mastodon ID and join the Moa party!

This Moa instance is run as a public utility. If you’re interested in getting involved, finding out more, reporting issues, and keeping track of status updates, visit the moaparty.com website »

«

Now, this is probably going to require you to give up write access to both accounts, and it’s going to be up to you to figure out how safe it is by reading the (open source) code.
unique link to this extract


Most Shared content from your Twitter friends • TweetShelf

Now that Nuzzel is dead – bought by Twitter, incorporated into Twitter Blue, then killed – it’s good to know there’s a different aggregator that will find the most-shared stories/podcasts/whatever from Twitter. (Apologies, I can’t find who sent me the link!)
unique link to this extract


Google nukes 50,000 accounts pushing Chinese disinformation • Bleeping Computer

Sergiu Gatlan:

»

Google’s Threat Analysis Group terminated tens of thousands of accounts linked to a group known as “Dragonbridge” or “Spamouflage Dragon” that is disseminating pro-Chinese disinformation across multiple online platforms.

According to Google, Dragonbridge gets new Google Accounts from bulk account sellers, and, in some instances, they’ve even switched to accounts previously used by financially motivated actors repurposed for posting disinformation videos and blogs.

Last year, the company took down more than 50,000 accounts used by Dragonbridge across its platforms, including YouTube, Blogger, and AdSense. In total, 100,960 accounts have been shut down since the influence network was first spotted.

This reflects Google’s focus on this coordinated information operation linked to China, described as “the most prolific IO actor TAG tracks.”

However, despite the Chinese influence operation’s large size and high volume of content production, it has minimal to no engagement from real viewers.

For instance, the vast majority of its YouTube channels had no subscribers when they were taken down last year, and more than 80% of videos had fewer than 100 views.

Dragonbridge blogs on Blogger also had a very low engagement, with less than 10 views per post for almost 95% of posts when they were terminated in December.

«

The disinformation business just ain’t what it used to be.
unique link to this extract


I’m a corporate fraud investigator. You wouldn’t believe the hubris of the super-rich • The Guardian

Freya Berry:

»

I work in corporate investigation, and it’s my job to spot fraud and corruption. Sometimes it’s as subtle as manipulating your free cash flow. Sometimes it’s as shameless as photoshopping your bank statements.

The job is varied. Generally, I’m looking for red flags: accounting abuses, yes, but also egregious business practices, such as failure to pay suppliers. We want to answer the question: is the business doing what it says it’s doing, to the standard it should be? And if not, can that company really be worth what the market believes it is?

I use a number of methods to do this. Some you might expect – financial reports, local filings, data sets – but others you might not, including social media. Recently I was deep in the Instagram account of a CEO whose company we believe is a house of cards. I gazed for a long time at a snap of a family member, posing on the steps of a private jet with champagne and a gun. If I had to paint a portrait called “hubris”, it would look something like that.

Corruption and its consequence, fraud, have many origins. Sometimes the executives like their jets and yachts a little too much. Or they might have connections with the mafia or drug cartels. Or they might just make big promises they can’t deliver. Far from being an easy way out, fraud is a high-wire act, the rewards high, the penalties higher. Companies therefore go to great pains to conceal it, making death threats to whistleblowers or putting tracking devices in short-sellers’ cars.

But while the fraudsters I’ve encountered are generally clever – or at least cunning – sooner or later something interesting happens: they get carried away. Drunk on their own genius, they concoct grander schemes, more flamboyant methods. They give large contracts to family members, who are themselves shareholders in the company. They fire their top-tier auditor and bring in a no-name firm. In the case of Sam Bankman-Fried, they allegedly participate in a Signal group chat entitled “Wirefraud”.

«

I once heard a talk given by a guy who’d defrauded his company of tens of thousands. He worked in the accounting department, and it was the classic story: he approved invoices, and nobody else did. So he began faking them. Then it got so big he couldn’t stop doing it. So many of these things unravel when the people involved go on holiday.

A huge amount of corporate fraud comes down to fake invoicing. It really is that simple.
unique link to this extract


ChatGPT’s mind-boggling, possibly dystopian impact on the media world • Vanity Fair

Joe Pompeo rounds up a week’s worth of ChatGPT beshogglings, and then finds some who are and who aren’t worried:

»

[Veteran newspaperman, Steve] Brill isn’t worried about ChatGPT and its ilk putting skilled reporters out of work. He told me about a final paper he assigns for his journalism students at Yale, in which they have to turn in a magazine-length feature and list “at least 15 people they interviewed and four people who told them to go fuck themselves. There is no way they could do that assignment with ChatGPT or anything like it, because what journalists do is interview people, read documents, get documents leaked to them.”

Still, Brill continued, “One of the assignments I give them on the second or third week is a short essay on how Watergate would have played out differently in the internet age, because Bob Woodward comes in as a guest for that session. I asked ChatGPT to answer that question, and the answer I got was this banal but perfectly coherent exposition. The difference is, you didn’t have to interview or talk to anyone. So maybe it’ll put some op-ed columnists out of work.”

As for [technology journalist Alex] Kantrowitz, getting plagiarized by bots hasn’t turned him into a ChatGPT hater. “I’m still super bullish on generative AI, and I still think it can be useful for journalism,” he said. “Sometimes I’ll use it when I’m stuck on a story, and I never include [the AI-generated text] in the story, but it can get my brain going, and that’s helpful. If you think about how it will impact journalism in next two or three years, the likely answer is, quite minimally. But as this technology gets better at scouring the internet and taking information, as its writing gets better, we’ll start to see a world where it can produce better writing and analysis than most professional reporters. If you’re doing original reporting and unearthing things people don’t already know, you’re probably gonna be okay. But if you’re an analysis person, let’s say, 20 years down the road, you might need to find something else to do.”

«

There’s quite a lot of “my job’s going to be OK, though” commentary. On the latest Exponent.fm podcast, James Allworth (who works at Cloudflare) was a bit apprehensive about how it could probably knock out a lot of tech support jobs.
unique link to this extract


How online mobs act like flocks of birds • Noema Mag

Renee DiResta explains how flocks and “murmurations” occur – each bird responds to what it sees its seven closest neighbours doing – and then applies it to online networks:

»

After the nudges to assemble into flocks come the nudges to engage — “bait,” as the Extremely Online call it. Twitter’s Trending Topics, for example, will show a nascent “trend” to someone inclined to be interested, sometimes even if the purported trend is, at the time, more of a trickle — fewer than, say, 2,000 tweets. But that act, pushing something into the user’s field of view, has consequences: the Trending Topics feature not only surfaces trends, it shapes them. The provocation goes out to a small subset of people inclined to participate. The user who receives the nudge clicks in, perhaps posts their own take — increasing the post count, signaling to the algorithm that the bait was taken and raising the topic’s profile for their followers. Their post is now curated into their friends’ feeds; they are one of the seven birds their followers see. Recurring frenzies take shape among particular flocks, driving the participants mad with rage even as very few people outside of the community have any idea that anything has happened. Marx is trending for you, #ReopenSchools for me, #transwomenaremen for the Libs Of TikTok set. The provocation is delivered, a few more birds react to what’s suddenly in their field of view, and the flock follows, day in and day out.

Eventually, perhaps, an armed man decides to “liberate” a DC pizza parlor, or a violent mob storms a nation’s capitol. Although mainstream tech platforms now act to disrupt the groups most inclined to harassment and violence — as they did by taking down QAnon groups and shutting down tens of thousands of accounts after the January 6th insurrection — the networks they nudged into existence have by this point solidified into online friendships and comradeships spanning several years. The birds scatter when moderation is applied, but quickly re-congregate elsewhere, as flocks do.

Powerful economic incentives determined the current state of affairs. And yet, the individual user is not wholly passive — we have agency and can decide not to take the bait. We often deploy the phrase “it went viral” to describe our online murmurations. It’s a deceptive phrase that eliminates the how and thus [falsely] absolves the participants of all responsibility.

«

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

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.