The woolly mammoth could make a comeback by 2028, and one company has a plan for how to monetise it. CC-licensed photo by Bjorn 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.
It’s Friday, so there’s another post due at the Social Warming Substack at about 0845 UK time. It’s about non-public messaging.
A selection of 10 links for you. Ooh, nearly big number. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.
Humane Ai Pin review: too clunky, too limited • WIRED
Julian Chokkattu:
»
I increasingly began to doubt the accuracy of the information the Humane wearable was providing. My mom told me to avoid high-fructose corn syrup right as my dad handed me a bottle of Malta Goya—she said the sweetener in it was banned in California. The Ai Pin agreed with this when I asked it. However, California did not ban it; the state banned four food additives last year, none of which are high-fructose corn syrup.
On my parents’ TV screen, an image of a temple popped up on the Chromecast’s screensaver. My dad asked where it was, so I positioned my Ai Pin toward the screen and said, “Look and tell me where this picture is from.” The answer? Angkor Wat in Cambodia. I didn’t have a specific reason to doubt this, but because the Pin doesn’t have a proper screen, there’s no way to verify it. I launched Google Lens on my phone, pointed the camera at the screen, and … well, the temple is the Phraya Nakhon Cave in Thailand. The images in the Google search matched perfectly with the screensaver.
Not being able to fully trust the results from the Ai Pin’s Ai Mic and Vision features (the latter is still in beta) is just one problem with this wearable computer. Unfortunately, there’s not much else to do with it as it’s missing a great many features. The Humane Ai Pin could be an interesting gadget a year from now after promised software updates, but at the moment it’s a party trick.
«
And this is one of the kinder reviews. Engadget was pretty brutal. However, The Verge: even more so. Shall we start the sweepstake on when Humane is going to close its doors or get acquihired?
unique link to this extract
Elon Musk’s X botched an attempt to replace “twitter.com” links with “x.com” • Ars Technica
Jon Brodkin:
»
Elon Musk’s clumsy brand shift from Twitter to X caused a potentially big problem this week when the social network started automatically changing “twitter.com” to “x.com” in links. The automatic text replacement reportedly applied to any URL ending in “twitter.com” even if it wasn’t actually a twitter.com link.
The change apparently went live on X’s app for iOS, but not on the web version. It seems to have been a problem for a day or two before the company fixed the automatic text replacement so that it wouldn’t affect non-Twitter.com domains.
Security reporter Brian Krebs called the move “a gift to phishers” in an article yesterday. It was a phishing risk because scammers could register a domain name like “netflitwitter.com,” which would appear as “netflix.com” in posts on X, but clicking the link would take a user to netflitwitter.com.
“A search at DomainTools.com shows at least 60 domain names have been registered over the past two days for domains ending in ‘twitter.com,’ although research so far shows the majority of these domains have been registered ‘defensively’ by private individuals to prevent the domains from being purchased by scammers,” Krebs wrote.
Even if the change had been implemented smoothly, auto-replacing “twitter.com” with “x.com” doesn’t do much to help Musk cement his branding shift because x.com still redirects to twitter.com.
One of the newly registered domain names inspired by X’s text replacement is the example mentioned above. Navigating to netflitwitter.com will show you a message that says, “This domain has been acquired to prevent its use for malicious purposes.” The webpage was set up by X user @yuyu0127_ and goes on to say:
»
As of April 8, 2024, the iOS Twitter (now X) client automatically replaces the text “twitter.com” in posts with “x.com” as part of its functionality. Therefore, for example, a URL that appears to be “netflix.com” will actually redirect to “netflitwitter.com” when clicked.
Please be aware that there is a potential for this feature to be exploited in the future, by acquiring domains containing “twitter.com” to lead users to malicious pages. This domain, “netflitwitter.com,” has been acquired for protective purposes to prevent its use for such malicious activities.«
«
Macs to get AI-focussed M4 chips starting in late 2024 • MacRumors
Juli Clover:
»
Apple will begin updating its Mac lineup with M4 chips in late 2024, according to Bloomberg’s Mark Gurman. The M4 chip will be focused on improving performance for artificial intelligence capabilities.
Last year, Apple introduced the M3, M3 Pro, and M3 Max chips all at once in October, so it’s possible we could see the M4 lineup come during the same time frame. Gurman says that the entire Mac lineup is slated to get the M4 across late 2024 and early 2025.
The iMac, low-end 14-inch MacBook Pro, high-end 14-inch MacBook Pro, 16-inch MacBook Pro, and Mac mini machines will be updated with M4 chips first, followed by the 13-inch and 15-inch MacBook Air models in spring 2025, the Mac Studio in mid-2025, and the Mac Pro later in 2025.
«
AI chips, AI chips everywhere. (And the Mac Pro too.)
unique link to this extract
Can chatbots hold meaningful conversations? • Nautilus
Elena Kazamia:
»
Arseny Moskvichev dreams of the day he can have a meaningful conversation with artificial intelligence. “By meaningful, I mean a conversation that has the power to change you,” says the cognitive and computer scientist. “The problem,” says Moskvichev, “is that LLMs are complete amnesiacs. They only have so much context they can attend to. If you’re out of this context, they forget everything you spoke about with them.”
Even the most advanced chatbots can only process about 16,000 words of text within a prompt when in conversation with a human user. This is called a “context window.” And they can’t connect the information they receive during different “conversations” with a human, or build a storyline.
To help chatbots learn to hold life-changing conversations, and to improve their comprehension of the deep complexities of context—of the webs of relationships between people, events, and timelines that govern human lives—Moskvichev and his colleagues are teaching them to read novels, the way we might learn to read them in high school literature classes.
The act of reading a novel might seem like a relaxing pastime, but it requires a nuanced intelligence. We use memory and complex, layered comprehension to follow multiple characters through twisting plots, scene changes, and narrative. And while we might not think about it, the average novel averages around 80,000 words. The Picture of Dorian Gray, by Oscar Wilde, for instance, runs at 82,000 words, while The Souls of Black Folk, by W.E. DuBois, totals around 72,000. The Little Prince, a children’s book, by Antoine de Saint-Exupéry, has around 17,000 words. All those words gradually build a story we hold and examine in our minds. But such skills are currently out of reach to Large Language Models (LLMs) like Open AI’s ChatGPT, which can process text but cannot be said to read the way we do.
«
A “conversation that can change you”? Shouldn’t Moskvichev just cut out the middleman and read lots of novels? (Or listen to them as audiobooks?)
unique link to this extract
AI-generated pornography will disrupt the adult content industry and raise new ethical concerns • The Conversation
Simon Dubé and Valerie Lapointe:
»
Advancements in machine learning and AI algorithms for image and video production have contributed to the growth of websites for AI-generated pornography, commonly referred to as AI porn.
The mass production of AI porn has significant ethical and social implications. It can offer an unprecedented quantity of customizable sexual stimuli tailored to users’ preferences while drastically cutting down production costs.
On one hand, these new tools enable content creators to produce diverse erotica and allow widespread access to personalized sexual stimuli that meet people’s needs and desires, thereby enhancing their sex life and well-being.
On the other hand, it could lead to problematic overuse of pornography, the spread of deepfakes, and the production of illegal content, such as child pornography.
AI porn also has labour implications, and could create copyright issues as well as impact the jobs of sex workers and adult content creators.
In all likelihood, the impact of AI porn will be more nuanced: some users will benefit, while others may be negatively impacted by it. However, the pace of technological developments leaves little time to plan and research how to harmoniously integrate this new technology into our lives. Like in many other sectors, we are not ready for AI porn.
«
I dunno, I bet there’s a demographic that’s extremely ready for AI porn. (Dubé is a research fellow at Indiana University, Lapointe is a PhD candidate in psychology at Quebec University.)
unique link to this extract
Colossal Biosciences has made headlines as the de-extinction startup primed to bring back the woolly mammoth. Here’s its business model • Fortune
Allie Garfinkle:
»
I’m like everyone else, and am obsessed with the idea of seeing a woolly mammoth. But what I really wanted to know was: how do you make money doing it? Thankfully, the answer’s not the woolly mammoth meatball—that’s a project cooked up by a different company focused on lab-grown meat. Colossal’s mammoth monetization model emphasizes two near-to-medium term efforts, and a longer-term play.
Let’s start with the near-term: “There’s consumer education, where we already get billions of media impressions,” said [CEO and founder Ben] Lamm. “We’re doing a docuseries, and we do lots of different educational content. Transparently, those do make money…Then, the midterm is the technology companies that we spin out, as we monetize, but those take a while to get to the point where they’re successful.”
…the longer-term strategy is tied to Colossal’s already-extensive conservation efforts. “De-extinction and species preservation are connected,” said Lamm. “Sometimes people want to separate out species preservation, conservation, and I say, no, they’re all one thing.”
Colossal’s conservation efforts are wide-ranging. For example, the startup is currently working with the University of Alaska and University of Stockholm on one of the largest-ever studies focused on radiocarbon dating mammoths. Likewise, Colossal last year joined the BioRescue group working to save the northern white rhino from extinction. And, there’s the Victorian grassland earless dragon—a miniscule Australian lizard believed to be extinct in 1969. The lizard was recently rediscovered in the wild, and Colossal’s set up a breeding and release colony with the Melbourne Zoo devoted to the fetching little reptile.
Lamm even envisions getting a piece of the carbon credit market, which is projected to be as big as $2.4 trillion by 2027, according to some estimates. “In the long-term, if we get to sustainable populations of, say, the Thylacine [Tasmanian tiger] in southern Australia, we’ll get government subsidies, carbon credits, biodiversity credits.”
«
Hm. I feel like the woolly mammoth meatball might be a more solid market. That, and hunting rights.
unique link to this extract
AWS resource restrictions point to datacenter power issues • The Register
Dan Robinson:
»
Datacenter power issues in Ireland may be coming to a head amid reports from customers that Amazon is restricting resources users can spin up in that nation, even directing them to other AWS regions across Europe instead.
Energy consumed by datacenters is a growing concern, especially in places such as Ireland where there are clusters of facilities around Dublin that already account for a significant share of the country’s energy supply. This may be leading to restrictions on how much infrastructure can be used, given the power requirements. According to Ireland’s Central Statistics Office (CSO), power usage by datacenters increased by 31% between 2021 and 2022, hitting 18% of the total metered electricity consumption.
A recent report from the International Energy Agency (IEA) estimated that if unchecked, this could grow to 32% of Ireland’s electricity by 2026.
AWS users have informed The Register that there are sometimes limits on the resources that they can access in its Ireland bit barn, home to Amazon’s eu-west-1 region, especially with power-hungry instances that make use of GPUs to accelerate workloads such as AI.
“You cannot spin up GPU nodes in AWS Dublin as those locations are maxed out power-wise. There is reserved capacity for EC2 just in case,” one source told us. “If you have a problem with that, AWS Europe will point you at spare capacity in Sweden and other parts of the EU.”
We asked AWS about these issues, but when it finally responded the company was somewhat evasive.
«
The Irish power grid operator was also evasive. When both organisations are evasive like that, you can take it as confirmation that yes, you’re right.
unique link to this extract
AWS told to pay $525m in cloud storage patent suit • The Register
Dan Robinson:
»
A jury has ordered Amazon Web Services to pay $525m for infringing distributed data storage patents in a case brought by a technology outfit called Kove IO.
Kove, which styles itself as a pioneer in high-performance computer storage and data management technologies, filed its original complaint [PDF] in 2018. It claims that AWS is infringing on three Kove-held patents in cloud services, such as the Amazon S3 storage platform, as well as in its DynamoDB database service, and in other related products and services.
The trial came to a close on Wednesday (April 10), with the jury finding in favor of Kove and awarding it damages of over half a billion dollars. AWS said it intends to appeal the verdict, which acknowledged that the company had not willfully infringed on the patents in question.
The technology at the center of the case relates to distributed hash tables, a decentralized system used to store and retrieve data, with the data in this instance being the location information for specific data files in a scale-out data storage platform.
«
Not sure it really matters whether you meant to infringe the patents; it’s just the fact of the infringement. Wonder if this company is also going after Google and Microsoft.
unique link to this extract
‘Help’ written in palm fronds assists US forces in rescuing Micronesians stranded on tiny island • Stars and Stripes
Joseph Ditzler:
»
The men, all in their 40s, set sail from Polowat Atoll, Micronesia, on March 31 in a small, open 20-foot skiff powered by an outboard motor, according to a news release from Coast Guard Forces Micronesia, Sector Guam. The men had experience navigating in those seas.
The skiff was damaged, though the release did not specify how or when, and the motor rendered inoperative.
A week later, on April 6, a relative “reported her three uncles had not returned from Pikelot Atoll,” approximately 115 miles northwest of Polowat Atoll, a part of Chuuk State in the Federated States of Micronesia.
Pikelot, a low coral island covered with palm trees and shrubs, is a speck just 2½ miles long and 1¾ miles wide in a search area the Coast Guard described as 78,800 square miles of the South Pacific.
Joint Rescue Sub-Center Guam mobilized a search that drew a U.S. Navy P-8 Poseidon patrol aircraft from Kadena Air Base, Okinawa, and the Guam-based Coast Guard cutter Oliver Henry.
The Poseidon crew found the three Sunday thanks to a message they left on the Pikelot beach.
“In a remarkable testament to their will to be found, the mariners spelled out ‘HELP’ on the beach using palm leaves, a crucial factor in their discovery,” Coast Guard Lt. Chelsea Garcia, the search and rescue mission coordinator, said in the release. “This act of ingenuity was pivotal in guiding rescue efforts directly to their location.”
«
There’s one really big question in my mind. Really big. Why did they spell HELP when they could have made SOS much bigger with the same number of fronds? (The picture shows that it’s a properly tiny atoll. They were lucky.) As the story points out, some previous stranded people did use SOS. Worked for them.
unique link to this extract
Instagram will automatically blur nudes in direct messages • 404 Media
Samantha Cole:
»
Instagram will soon start testing blurring nudes in direct messages, according to an announcement from Meta on Wednesday. It will use “on-device machine learning” to detect nudes, and is aimed at stopping sextortion schemes that target teenagers, the company wrote.
The “nudity protection” feature will detect nudes in direct messages and automatically blur them, both when you’re getting a nude sent to you and when you’re attempting to send one to someone else. Users will also see a popup message that says, “Take care when sharing sensitive photos. The photo is blurred because nudity protection is turned on. Others can screenshot or forward your photos without you knowing. You can unsend a photo if you change your mind, but there’s a chance others have already seen it.”
Because the processing is performed on device, Meta says its detection will still work on end-to-end encrypted chats, with Meta not having access to the images themselves unless a user reports them.
The feature will be turned on by default for users under the age of 18, and adults will get a notification “encouraging” them to turn it on. “This feature is designed not only to protect people from seeing unwanted nudity in their DMs, but also to protect them from scammers who may send nude images to trick people into sending their own images in return,” the announcement says.
…Meanwhile, Meta lets AI-generated fake influencers that steal real women’s images run rampant on Instagram and can’t stop even the most obvious catfish romance scams happening on the platform—and on Facebook, people are constantly being tricked by AI-generated content into believing it’s real.
«
Have to admire Cole pointing out the non-thing that Meta has achieved here: how much of a problem are nude pics in DMs? Meta doesn’t say, though it does say
»
we’re also developing technology to help identify where accounts may potentially be engaging in sextortion scams, based on a range of signals that could indicate sextortion behavior. While these signals aren’t necessarily evidence that an account has broken our rules, we’re taking precautionary steps to help prevent these accounts from finding and interacting with teen accounts.
«
| • 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 