A new motorcycle helmet promises all-round vision thanks to AI. Will that reduce accidents? CC-licensed photo by sprklg 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. Envisioned. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

---

This AI motorcycle helmet promises 100% blind-spot elimination • New Atlas

Utkarsh Sood:

»

AI has now made its way to motorcycle helmets. Gone are the days of cautious shoulder checks. Intelligent Cranium Helmets promises to provide riders with the highest level of protection by combining improved visibility, crash detection, and seamless connectivity into a single, high-tech helmet.

After showing a design concept back in 2015, the company has finally been able to produce an actual mass-produced, retail-ready product that is full as a tick with futuristic functionality. It’s called the iC-R, which is actually a range of four smart motorcycle helmet models, each with an added set of features to tickle your fancy. It was unveiled at CES 2025, and proved successful at garnering a fair amount of attention in the moto world.

But first, who is Intelligent Cranium Helmets? The Virginia-based company started in 2015 with a safety-first smart helmet tech approach that put a prime focus on integrating artificial intelligence into its product range.

“While driving to work one day, I noticed a number of motorcyclists traveling in the same direction I was,” says CEO and co-founder, Ambrose Dodson. “I observed the motorcyclists repeatedly turning their heads whenever they needed to change lanes, and I said to myself – there has to be a safer way for these riders to ride.”

The focal point of the AI helmet is its 240º field of reverse camera vision, which stitches together footage from two rear-mounted cameras into a wrapraound rear view video feed. This feed can be displayed right above the rider’s eyeline in a heads-up display (HUD).

The rider’s own visual field of view covers somewhere around a further 120º, adding up to what Intelligent Cranium claims is a full 360º of visual awareness. What’s more, there’s a 1080p/60fps action camera at the front of the helmet that covers a 152º angle.

«

I’m reminded of XKCD 538, “Security”, which overlooks how incredible technology is foiled by simple force. And this is nice, but won’t stop drivers pulling out of side roads or affect any of the most common motorcycle crash causes.
unique link to this extract

---

Cutting-edge Chinese “reasoning” model rivals OpenAI o1—and it’s free to download • Ars Technica

Benj Edwards:

»

Alongside the release of the main DeepSeek-R1-Zero and DeepSeek-R1 models, DeepSeek published six smaller “DeepSeek-R1-Distill” versions ranging from 1.5 billion to 70 billion parameters. These distilled models are based on existing open source architectures like Qwen and Llama, trained using data generated from the full R1 model. The smallest version can run on a laptop, while the full model requires far more substantial computing resources.

The releases immediately caught the attention of the AI community because most existing open-weights models—which can often be run and fine-tuned on local hardware—have lagged behind proprietary models like OpenAI’s o1 in so-called reasoning benchmarks. Having these capabilities available in an MIT-licensed model that anyone can study, modify, or use commercially potentially marks a shift in what’s possible with publicly available AI models.

“They are SO much fun to run, watching them think is hilarious,” independent AI researcher Simon Willison told Ars in a text message. Willison tested one of the smaller models and described his experience in a post on his blog: “Each response starts with a … pseudo-XML tag containing the chain of thought used to help generate the response,” noting that even for simple prompts, the model produces extensive internal reasoning before output.

«

Ben Thompson at Stratechery, who follows this stuff in a manner that is a bit more reader-friendly than Willison, is very excited at the prospect of a locally-run LLM that fits in a suitably enabled home computer.

It feels like this stuff is rapidly becoming commoditised; the only thing you’ll need is a computer with LOTS more RAM. Which takes us back to the good old 1990s/2000s, when getting more RAM was almost more important than having a faster CPU.
unique link to this extract

---

Gaming on Apple Vision Pro could see huge growth soon, suggest game makers • 9to5Mac

Ryan Christoffel:

»

As published by GamesBeat, a recent Game Developers Conference survey sought to gauge engagement from developers with Apple’s new spatial computing platform.

The data points to big growth coming soon. Here’s what it says:

»

Although only 8% of VR/AR developers are currently making games for Apple vision OS, the platform looks to be growing its foothold. Almost one-fifth (18%) of respondents say their next games will be on the platform, and one-fourth are interested in Apple’s VR headset.

«

These three data points are all interesting in their own ways.

First, it’s a bit surprising that 8% of VR/AR game developers are currently working on a visionOS project. We certainly haven’t seen much from their efforts yet, as Vision Pro gaming options have been scarce so far. That number is encouraging to hear.

Second, the fact that 18% of respondents confirmed their next game would be on Vision Pro shows how much growth we should see soon.

The final question is especially relevant too, especially when comparing visionOS interest to that of other platforms.

Notably, 26% of respondents said that visionOS was of high interest to them as a game developer, while only 25% said the same for PlayStation VR2—a platform that might be about to bolster Apple’s Vision Pro efforts.

«

I’ll believe it when I see it, because the Vision Pro just doesn’t have a big enough user base to be attractive. That would be the most sunk of sunk costs.

The Gamesbeat link is worth reading, because it distils an analysis of the current situation and mood in games. Which is: gloomy.
unique link to this extract

---

Trump Admin accused of using AI to draft Executive Orders • Futurism

Maggie Harrison Dupré:

»

Mere hours after being sworn in as the 47th president of the United States on Monday, returning President Donald Trump got to work signing dozens — and counting — of executive orders, which range from commands for the US to withdraw from the Paris climate agreement and the World Health Organization to ordering an end to birthright citizenship and renaming the Gulf of Mexico to the “Gulf of America.”

But while the executive actions range in scope, legal experts have called attention to some curious common threads: bizarre typos, formatting errors and oddities, and stilted language — familiar artifacts that have led to speculation that those who penned them might have turned to AI for help.

“Lots of reporting suggested that, this time around, Trump and his lawyers would avoid the sloppy legal work that plagued his first administration so they’d fare better in the courts,” Slate journalist and legal expert Mark Joseph Stern remarked last night in a Bluesky post. “I see no evidence of that in this round of executive orders.”

“This is poor, slipshod work,” he added, before alleging that the actions were “obviously assisted by AI.”
In another post, Stern pointed to a deeply questionable section of an executive action titled “Unleashing Alaska’s Extraordinary Resource Potential,” which details how the US will take advantage of the state’s “untapped supply of natural resources,” in part by drilling for fossil fuels in regions of previously-protected natural land.

In that section, the order includes a numbered list of several distinct Public Land Orders to be reinstated. Each land order, however, is listed next to the number one — an apparent slip-up, we should point out, that we’ve noticed on seemingly AI-generated content in the past.

«

I wouldn’t rule out it just being slipshod humans, given it’s the Trump team. If they did use AI, the output is probably greatly improved from the original.
unique link to this extract

---

Weight-loss jabs linked to reduced risk of 42 conditions including dementia • The Guardian

Nicola Davis:

»

People with diabetes taking medications found in weight-loss jabs have a reduced risk of 42 conditions, research has found, paving the way for such drugs being used to treat a host of health problems.

The most comprehensive study of its kind showed that psychotic disorders, infections and dementia were among conditions found to be less likely to occur when using GLP-1RAs, which are found in the medications Saxenda, Wegovy and Mounjaro.

The researchers compared health outcomes for people with diabetes who received usual care with those also given drugs such as liraglutide, semaglutide and tirzepatide. While the team revealed the risk of many conditions was lower for the latter group, the risk of other conditions, including arthritic disorders, was increased.

And the scientists say that the benefits are not just restricted to people with diabetes, suggesting they could also be found in other people using the jabs, such as those who take them to fight obesity.

“We only studied people with diabetes but there is no biologic or clinical reason to think that the beneficial and risk profiles would be very different in people without diabetes,” said Dr Ziyad Al-Aly, a co-author of the research from Washington University in St Louis.

However, Aly said it was unlikely people without obesity would experience a similar range of potential benefits. He added that some of the positive associations might be linked to weight loss, while it was also important to consider the risks.

«

Greater weight leads to higher risk of psychotic disorders, infections and dementia? (For the latter, the study only has 3.5 years of data, so that seems an odd conclusion.)
unique link to this extract

---

The making of Community Notes • Asterisk

Asterisk magazine:

»

what falls out of those later studies that compare Community Notes to expert fact checks as well — the trust is higher. 

Lucas [Neumann, product designer at X]: Yes. But one thing to note about that outcome is that we had to put a lot of work into overcoming people’s priors. If you go back to 2021, and someone sees a tweet with a box on it, they immediately think, “Oh, this is a fact check.” They would assume that Twitter wrote it, or that the Twitter CEO decided that it should be there. What we’re taking one hour to tell you here is something we had to explain to them in a split second with just one line of copy. Arriving at that design and what those words are — I don’t think anyone here has ever done so many iterations on one rectangle. Things like, what’s the shade of blue that will make people calmer when they see this? The original design that Keith made was an orange box with “This is misleading information” at the top. Coming from that design to what we have now was a learning process. 

Keith [Coleman, VP of product at X]: That line — “Readers added context they thought people might want to know” — we iterated on that line so many times to find something that could succinctly describe what had happened here, how this came to be, that this was by the people, not by the company, and that it was there for your information, not to tell you what to think. 

Emily [Thai, formerly University of Chicago consultant to the Birdwatch scheme which became Community Notes]: I don’t think you will ever hear any of us — anybody who worked on this project — ever say the word “fact check.” There’s a care to avoid using that phrasing in any of the things we say about the product, any of the language about it, anything on the product surface, because it’s entirely about providing context and information and then letting you make your own decision about how to trust it.

«

This is a long but very interesting interview with this team.
unique link to this extract

---

Netflix UK audience reach overtakes BBC1 for first time in 2024 • Deadline

Jake Kanter:

»

Netflix was the most-watched TV service in the UK for three months last year, overtaking dominant domestic network BBC1 for the first time, according to a Deadline analysis.

The unseating of Britain’s most popular channel may not have been permanent, but represents a possible inflection point in the battle between traditional broadcasters and U.S. streaming giants. The BBC said it was “meaningless” to compare the entirety of Netflix with a single channel and that its portfolio had double the number of viewers of the Squid Game streamer.

Viewing figures from BARB, the UK’s official ratings body, showed that Netflix’s audience reach overtook BBC1 in September, October, and November 2024. For these three months, Netflix’s average audience reach stood at 43.2M, compared with BBC1’s 42.3M viewers.

BBC1 pulled ahead again in December, despite Netflix posting a record reach of 46.4M after streaming series including UK original Black Doves. BBC1’s reach was 48.4M last month, no doubt buoyed by festive hits including Gavin & Stacey and Wallace & Gromit: Vengeance Most Fowl.

«

Yet you never hear anyone complaining about the news on Netflix, or demanding it be defunded. (Nor listening to its radio stations or reading its news website.) A basic ad-free Netflix subscription will cost £132 per year (likely going up presently, see below); the standard BBC licence fee is £169.50. Amazing what that extra £37.50 gets you.
unique link to this extract

---

Netflix is raising prices again – The Verge

Emma Roth:

»

Netflix is raising prices yet again. In its latest earnings report released Tuesday, the streaming service announced that “we are adjusting prices today across most plans” in the US, Canada, Portugal, and Argentina.

Netflix spokesperson MoMo Zhou tells The Verge that the ad-supported tier is increasing from $6.99 to $7.99 per month, while the standard ad-free tier will go from $15.49 to $17.99 per month. Its highest-priced premium tier is also increasing from $22.99 to $24.99 per month. The price hikes will go into effect during subscribers’ next billing cycle, according to Zhou.

“As we continue to invest in programming and deliver more value for our members, we will occasionally ask our members to pay a little more so that we can re-invest to further improve Netflix,” the company’s letter to investors says. Netflix last raised the price of its subscription in October 2023. This is also the first time it’s raising the price of this ad-supported plan, which it rolled out in 2022.

Netflix added 19 million new subscribers over the past few months — the most in its history during a single quarter — bringing its global total to 300 million.

«

A graphic with the story shows that since 2015, prices have risen by 25% (no HD, one screen) to 100% (HD or 4K, two to four screens). Has the content really got that much more expensive to make/license?
unique link to this extract

---

Reeves intervenes in UK car finance mis-selling case to protect lenders • Financial Times

George Parker, Alistair Gray and Akila Quinio:

»

In April the Supreme Court is due to hear an appeal brought by car loan providers challenging an October ruling from the Court of Appeal that sided with consumers who complained about “secret” commissions on car loans.

The judgment that it was unlawful for banks to pay a commission to a car dealer without the customer’s informed consent sent shockwaves through the UK banking system and triggered thousands of pounds in compensation payments from lenders FirstRand Bank and Close Brothers.

HSBC analysts have estimated the total cost of compensation could reach £44bn, echoing the £50bn paid out by banks after the scandal of the mis-selling of payment protection insurance (PPI).

In a submission to the Supreme Court, seen by the Financial Times, the Treasury claims the case has “potential to cause considerable economic harm and could impact the availability and cost of motor finance for consumers”.

The Treasury application said that the case might “generate a perception that regulation in the UK is uncertain”. Last week Reeves called in regulators to push them into sweeping away rules that hinder growth.

It also argues that if liability is established, then the Treasury would seek to persuade the Supreme Court that “any remedy should be proportionate to the loss actually suffered by the consumer and avoid conferring a windfall”.

Treasury insiders argue that rather than taking sides with the banks against wronged consumers, the government wants to maintain the viability of a finance sector vital for the purchase of both new and second-hand cars.

«

OK, but how much did the banks get from this? How much did it push prices up? Why do banks get away with it again and again?
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