New research found that mice’s life could be extended by a quarter by blocking an inflammatory protein. CC-licensed photo by Free Public Domain Illustrations by rawpixel 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 0845 UK time. Free signup.
A selection of 9 links for you. Yes, but in mice. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.
US eyeing new rules to keep Chinese software out of cars • The Verge
Andrew Hawkins:
»
While speaking at a forum in Colorado, Alan Estevez, who serves as under secretary of commerce for industry and security, said that the department would propose rules that would require certain vehicle software be made in the US or by its trade partners. The rules would pertain to “key driver components of the vehicle that manage the software and manage the data around that car,” Estevez said, according to Reuters.
The new rules come as the Biden administration ramps up its scrutiny of Chinese auto imports
Such action would mirror trade restrictions placed against companies like Huawei over national security concerns that the telecom giant could be exploited by the Chinese government for espionage.The rules would stem from an investigation launched earlier this year by the Commerce Department into connected vehicle software produced in China and other nations that are considered antagonistic to the US.
The probe focused on “connected vehicles,” a broad term that can be applied to any car with internet access. It was meant to address concerns that technology like cameras, sensors, and onboard computers could be exploited by foreign adversaries to collect sensitive data about US citizens and infrastructure.
China has previously accused the US of repeatedly abusing “the concept of national security” to wrongfully target Chinese companies and impede competition from global markets.
«
It’s a weird sort of pre-Cold War where we’re now suspicious of anything China might do, even though there has been no actual event where such software has been used to undermine something. Though, of course, better safe than sorry.
unique link to this extract
Could A.I. make the modular phone a reality? This concept revives the ‘PhoneBloks’ dream with a twist • Yanko Design
Sarang Sheth:
»
Rather than attaching components to your phone, the PAIR Node allows you to attach sensor modules that help your phone capture different data points around you and your life. Called ‘Nodes’, these modules assist users in forming a lifestyle and AI-driven experience that suits their needs. Rather than buying a phone for its capabilities, you build your phone around the capabilities you need, from having a great camera to being a great health-tracking gadget. As you grow older, the phone grows with you, changing with time and your needs but never being replaced. It’s a clever way to help solve the planned obsolescence problem with tech, by bringing AI-based assistance into the mix.
Phones are nothing but vessels for the apps they carry – but AI is a little different. Everyone uses AI differently based on exactly what the needs of their life are. That prompted PAIR Node’s creators to revisit phone modularity in the AI age. The PAIR Node is a phone you build based on your requirements, but also on the ability to build an AI-powered device that grows with you and trains as you go. The framework of how this works remains extremely similar to the PhoneBloks concept from years back (which was acquired by Google and Motorola and turned into the now-shelved Project Ara), albeit with a few different modules based on how tech has advanced in the past few years.
«
Could we just wonder why it was that Project Ara, which we were told was going to be the future of smartphones, got shelved? Could it be because people, in fact, after all is said and done, at the end of the day, just don’t want modular smartphones? Even if you try to dress them up with AI?
unique link to this extract
New anti-ageing therapy extends life of mice by 25%, study finds • FT
Michael Peel:
»
A therapy based on the science that allows axolotl salamanders to regrow severed limbs can help mice live 25% longer, according to the latest breakthrough in anti-ageing research.
The technique, which involves suppressing a pro-inflammatory protein, protects the rodents against multiple illnesses and is in early-stage human clinical trials for fibrotic lung disease.
The results highlight hopes of how a deepening understanding of the role of individual genes and proteins could help increase both lifespan and healthspan — years of healthy life — in humans.
“What we’ve come across is a pro-inflammatory factor that drives ageing in the broadest sense,” said research leader Stuart Cook, a professor at Duke-NUS Singapore and the UK’s MRC Laboratory of Medical Sciences. “We’ve found that if you inhibit this factor, you increase healthspan — [and] also lifespan as a corollary. It’s a knock-on effect.”
The new research, published in Nature on Wednesday, focuses on the role of a protein named IL-11 that stokes inflammation. Increased production of this protein is associated with ageing in mice, the scientists found.
Deleting the genes that instruct IL-11 production protected the mice against various illnesses, metabolic decline and frailty, the research showed. That in turn enabled the rodents to live on average 24.9% longer.
Blocking IL-11 with an antibody had a similarly positive effect on lifespan for middle-aged mice. Males aged 75 weeks — roughly equivalent to 55 years in humans — lived 22.5% longer, while the figure rose to 25% for females.
«
Very much an IN MICE study. If IL-11 has this deleterious effect (in the age sense) why is it there?
unique link to this extract
Balloons to hoist tourists 100,000 feet into the stratosphere • CNBC
Magdalena Petrova:
»
People interested in seeing the earth from an unusual vantage point may soon have another option [than getting in a rocket]. CNBC spoke to three startups — France-based Zephalto, Florida-based Space Perspective and Arizona-based World View — that aim to hoist tourists to the stratosphere using pressurized capsules and massive gas-filled balloons.
“The capsule itself is designed to to carry eight customers and two crew into the stratosphere,” said Ryan Hartman, CEO of World View. “There will be a center bar where people can gather, and then, of course, there will be a bathroom aboard the capsule.”
The balloon rides will last around 6 hours, but will not take passengers all the way to space. Most will reach heights of 15 to 19 miles above the earth’s surface, flying in an area known as the stratosphere. The start of space is generally accepted by the US government to be around 80 kilometers, or about 50 miles, above the earth’s surface.
Jane Poynter, founder and co-CEO of Space Perspective, has a different view. “There is no universal definition of space,” Poynter said. “We are regulated as a spaceship. If we go over 98,000 feet, we are a spaceship. Outside the capsule, it’s essentially a vacuum. We’re above 99% of Earth’s atmosphere, which is why the sky is so deep black.”
Compared to rocket-powered space tourism, the physical sensation that passengers will experience on a stratospheric balloon ride is more comparable to being on an airplane. Passengers will not experience weightlessness.
«
So like a very expensive plane trip with slightly better views? I don’t get it.
unique link to this extract
“Extraordinarily disappointed” users reckon with the Google-fication of Fitbit • Ars Technica
Scharon Harding:
»
Since the acquisition closed in 2021, the Google-fication of Fitbit has largely meant a reduction in features and a focus from Google on getting people onto the Fitbit app. Long-time users have flocked to Fitbit—sometimes upon Fitbit’s request—to share hundreds of complaints about recent changes. However, Google has been mostly unresponsive to customer feedback.
In June, Google announced it was discontinuing Fitbit.com’s online dashboard. After July 8, users seeking similar features that the web app provided have to download the Fitbit mobile app. On Fitbit’s Community forum, a company representative confirmed that users’ “details and logging for activities, nutrition, sleep, and weight” would remain available via the app. However, the change inconvenienced users who preferred or needed to access such data on a bigger screen than a phone’s. Worse, the app lacks some of the features of the online dashboard, such as food logging.
Despite these obvious user drawbacks, the need to Googlize Fitbit seemed to drive the change. Announcing the news on the Community forum, a Fitbit company rep said:
»
Combined with Google’s decades of being the best at making sense of data, it’s our mission to be one combined Fitbit and Google team. Consolidating the Fitbit.com dashboard into the Fitbit app is a part of that mission, and will allow us to focus on features that provide even more valuable insights to our users.
«
Google has invested in the Fitbit app, which includes plans to let premium subscribers test experimental generative AI Fitbit features soon. Google is also developing a large language model for new features for the Fitbit app that users are being forced onto. Google has been pushing users to the Fitbit app for a while; in 2022, Fitbit devices lost the ability to sync with computers.
«
Google: still bad at hardware.
unique link to this extract
Google AI Overviews only show for 7% of queries, a new low • Search Engine Land
Danny Goodwin:
»
Google’s AI Overviews now appear less than 7% of the time. This is one of 10 noteworthy findings from a new analysis of AI Overviews.
This trend of less visibility for AI-generated answers started in mid-April, when the number of Google Search results without SGE [Search Generative Experience, the AI-generated search answer] jumped to 65%, up from 25%.
Google then announced the rollout of AI Overviews in the U.S. at Google I/O in May, and the trend continued. AI Overviews only showed for 15% of queries. We next saw multiple examples of incorrect and dangerous AI-generated answers, such as Google suggesting people drink urine and eat rocks. Google promised to improve AI Overviews.
This data was shared with Search Engine Land by enterprise SEO platform BrightEdge and its BrightEdge Generative Parser, which has been tracking and monitoring AI Overviews (and formerly Search Generative Experience) since late last year.
Google continued to reduce the presence of AI Overviews in June – dropping from 11% to 7% of queries, according to BrightEdge. However, there was also a slight increase in AI Overviews in mid-June before the big drop.
«
Seemed like a great idea once, doesn’t seem so now.
unique link to this extract
Job application for Remote AI Writing Evaluator • Jobs at Outlier
»
Outlier helps the world’s most innovative companies improve their AI models by providing human feedback. In this role, you will become an AI coach, assessing the quality of AI-generated writing, reviewing the work of fellow writers, and crafting original responses to prompts in order to teach the model what truly excellent writing looks like.
While we will be selective in terms of writing ability, we are actively recruiting for this role and encourage you to apply if you feel you would be a good fit.
Highlights:
• Flexible work schedule: Work whenever and wherever you want
• Weekly payouts: Automatically receive timely payments (no invoicing!)
• Unlock opportunities: Gain experience in the field that will dominate the next decade and beyondWhat you’ll be doing:
• Rating the quality of AI-generated writing on rubrics such as factuality, completeness, brevity, and grammatical correctness
• Reviewing the work of fellow human writers
• Responding to prompts with top-tier original writing«
Oh, you wanted to know about the pay? US$25 per hour. Available to people living in the US, Canada, UK, Ireland, New Zealand and Australia.
It seems quite a weird job. (But not open to anyone: you have to respond from an active .edu address. Students and academics only, it seems.)
unique link to this extract
Academic journals are a lucrative scam – and we’re determined to change that • The Guardian
Arash Abizadeh:
»
The commercial stranglehold on academic publishing is doing considerable damage to our intellectual and scientific culture. As disinformation and propaganda spread freely online, genuine research and scholarship remains gated and prohibitively expensive. For the past couple of years, I worked as an editor of Philosophy & Public Affairs, one of the leading journals in political philosophy. It was founded in 1972, and it has published research from renowned philosophers such as John Rawls, Judith Jarvis Thomson and Peter Singer. Many of the most influential ideas in our field, on topics from abortion and democracy to famine and colonialism, started out in the pages of this journal. But earlier this year, my co-editors and I and our editorial board decided we’d had enough, and resigned en masse.
We were sick of the academic publishing racket and had decided to try something different. We wanted to launch a journal that would be truly open access, ensuring anyone could read our articles. This will be published by the Open Library of Humanities, a not-for-profit publisher funded by a consortium of libraries and other institutions. When academic publishing is run on a not-for-profit basis, it works reasonably well. These publishers provide a real service and typically sell the final product at a reasonable price to their own community. So why aren’t there more of them?
…There is an obvious alternative [for funding]: universities, libraries, and academic funding agencies can cut out the intermediary and directly fund journals themselves, at a far lower cost. This would remove commercial pressures from the editorial process, preserve editorial integrity and make research accessible to all. The term for this is “diamond” open access, which means the publishers charge neither authors, editors, nor readers (this is how our new journal will operate). Librarians have been urging this for years. So why haven’t academics already migrated to diamond journals?
«
Meta won’t bring future multimodal AI models to EU • Axios
Ina Fried:
»
“We will release a multimodal Llama model over the coming months, but not in the EU due to the unpredictable nature of the European regulatory environment,” Meta said in a statement to Axios.
Apple similarly said last month that it won’t release its Apple Intelligence features in Europe because of regulatory concerns.
The Irish Data Protection Commission, Meta’s lead privacy regulator in Europe, did not immediately respond to a request for comment.
Meta plans to incorporate the new multimodal models, which are able to reason across video, audio, images and text, in a wide range of products, including smartphones and its Meta Ray-Ban smart glasses.
Meta says its decision also means that European companies will not be able to use the multimodal models even though they are being released under an open licence. It could also prevent companies outside of the EU from offering products and services in Europe that make use of the new multimodal models.
The company is also planning to release a larger, text-only version of its Llama 3 model soon. That will be made available for customers and companies in the EU, Meta said.
Meta’s issue isn’t with the still-being-finalized AI Act, but rather with how it can train models using data from European customers while complying with GDPR — the EU’s existing data protection law.
«
“Able to reason across”? This “reasoning” stuff is news to everyone, I think. We need better phrases for what these things do. Meanwhile, the EU is not making itself any friends in California.
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 