November 03, 2021
Welcome to Axios Future, where I loved seeing your Halloween costumes almost as much as I loved secretly siphoning off my kid's candy.
If you haven't subscribed, please do.
- Send feedback, tips and edible NFTs (see item No. 6) to [email protected].
Today's Smart Brevity count: 1,604 words or about 6 minutes.
1 big thing: America's roads are getting deadlier
Motor vehicle crash fatalities experienced the largest half-year spike on record in the first six months of 2021.
Why it matters: Road deaths are one of the biggest if underappreciated public health threats in the U.S. and the world.
- Speeding, distracted driving, and drug and alcohol use all play a role, but tougher automated traffic policing could help reduce the death toll.
By the numbers: An estimated 20,160 people died in motor vehicle accidents through the first half of 2021, according to data released late last week by the National Highway Traffic Safety Administration (NHTSA).
- That's the highest total for the first six months of the year since 2006 and about 18% higher than the death toll for the first half of 2020 — the biggest percentage increase for that time period since NHTSA began keeping records in 1975.
- It puts the U.S. on track for more than 40,000 motor vehicle crash deaths in 2021, roughly equivalent to the number of Americans who died last year in gun homicides, suicides and accidents combined.
What they're saying: "This is a crisis," Transportation Secretary Pete Buttigieg said in a statement after the data was released. "We cannot and should not accept these fatalities as simply a part of everyday life in America."
That's precisely what we do.
- The U.S. has long been an outlier among relatively well-off nations for its dangerous roads — a recent OECD report ranked only a few countries like Mexico and South Africa below the U.S. for per capita road fatalities.
- Most of America's economic peers have seen road fatalities drop over the past 20 years, some quite significantly.
- U.S. motor vehicle deaths are also down somewhat from 2000, but the drop is shallower, and since hitting a modern low of 32,479 in 2011, deaths have generally risen over the past decade.
Between the lines: The pandemic marked the beginning of the surge in U.S. motor vehicle fatalities.
- 38,680 people died on U.S. roads in 2020, a roughly 7% increase from the year before.
- But that figure is starker because deaths rose even as the number of miles driven by Americans dropped by about 13% in 2020 to the lowest level in two decades.
- While driving is nearly back to pre-pandemic levels, motor vehicle deaths per mile traveled have barely fallen.
Context: As the roads emptied out during the pandemic, it opened up room for Americans to indulge their worst road habits: driving too fast, too distracted and, often, too impaired. And those habits have continued into 2021.
- Data cited by NHTSA shows average speed increased during the last three quarters of 2020, and extreme speed — 20 mph or more above the posted speed limit — became more common, which helped result in an 11% increase in speed-related fatalities.
- Research found an increase in the involvement of drugs and alcohol in crashes, with one study showing almost two-thirds of seriously or fatally injured drivers tested positive for at least one intoxicant between mid-March and mid-July 2020.
- Data also showed an increase in phone use while driving during the pandemic and into 2021, which further contributes to accidents. An analysis by Zendrive found in more than 16% of the crashes its algorithm detects, a cellphone was manipulated less than five seconds before impact.
2. Microsoft adds GPT-3 AI writing model to its cloud service
Microsoft announced yesterday it will begin offering an updated version of the AI natural language program (NLP) GPT-3 to business customers as part of its Azure cloud platform.
Why it matters: The move puts what is likely the most powerful AI writing and reading algorithm at the fingertips of large businesses that will be able to use it to automatically analyze and generate new written content.
Driving the news: While OpenAI — the artificial general intelligence research company that created GPT-3 — has and will continue selling access to the model through its own API, Microsoft will offer a version for corporate clients that emphasizes "safety and security," says Eric Boyd, corporate vice president of Azure AI at Microsoft.
- Flashback: In 2019, Microsoft invested $1 billion in OpenAI, a partnership that made it the exclusive provider of cloud computing services for the company.
How it works: GPT-3 is a natural language transformer program that was trained on half a trillion words on the internet, making it the largest such model in the world when it was released last summer.
- Boyd says Azure customers could use GPT-3 to summarize vast amounts of customer feedback or analyze transcripts of live sports broadcasts to generate running commentary.
- While companies are already using computer vision and other AI tools, "the number of natural language use cases [for businesses] dwarfs everything else," he adds.
3. The synthetic data that will help build AI and the metaverse
Synthetic data — the generation of artificial images to train AI and computer vision — will be key to building out a future metaverse.
Why it matters: AI has long been trained on images — including human faces — captured from the real world, but doing so can create serious privacy concerns.
- Using synthetic data instead can help sidestep that issue, though it brings new worries about accuracy and authenticity.
Driving the news: Facebook announced on Tuesday that it plans to shut down its decade-old facial recognition system and delete the facial scans of more than a billion users, out of what it said were privacy concerns.
Between the lines: Increasingly, privacy concerns will lead companies to move from capturing real faces and other images to train AI as they transition to using synthetically generated data.
- Tel Aviv-based synthetic data company Datagen does high-quality level digital scans and motion capture of real people and objects and then uses AI to generate realistic but not real versions.
- Gartner predicted recently that by 2024, 60% of the data used for the development of AI and analytics projects will be synthetically generated.
The big picture: Since images of real people aren't being used directly, privacy and bias are less of a concern.
- Early computer vision systems were often trained on datasets taken from the internet that were disproportionately white and male, which meant they were less accurate in recognizing faces from other races and genders.
- With synthetic data, "you can incorporate the real distributions of the real world, so there's no bias among age, gender and more," says Gil Elbaz, co-founder and CTO of Datagen.
What's next: Synthetic data will be key to creating a more realistic version of the AR and VR future called the metaverse.
- "The metaverse is going to have a hardware and software component," says Elbaz. "Synthetic data will be part of the software that enables the right kind of hardware."
4. Pentagon sharply raises estimate of China's nuclear expansion
China is rapidly accelerating the expansion of its nuclear stockpile and is likely seeking to quadruple its number of nuclear warheads by 2030, according to the Pentagon's annual report to Congress on China's military power, writes Axios' Zachary Basu.
Why it matters: U.S. officials and experts have raised alarms at reports of China's nuclear expansion and testing of advanced weapons capabilities, including a hypersonic missile this summer, as tensions with Washington have reached new highs.
Driving the news: The Pentagon's assessment found that China may have up to 700 deliverable warheads by 2027 and 1,000 by 2030 — a sharp revision upward from last year, when the U.S. estimated China's stockpile would double from the low 200s over the next decade.
- The U.S., by comparison, has 5,550 nuclear warheads, while Russia has 6,255, according to the Stockholm International Peace Research Institute.
- Unlike those two countries, however, China has refused to join talks on arms control.
The big picture: The 192-page Pentagon report details China's broader military goals and evolving capabilities, and it aligns with warnings from senior U.S. military officials that China poses the most significant threat to U.S. military supremacy out of any potential adversary.
5. Worthy of your time
The race is on to develop a vaccine against every coronavirus (Maryn McKenna — Wired)
- And if they succeed, they'll just need to race to convince everyone to actually take it.
Is Amazon changing the novel? (Parul Sehgal — New Yorker)
- Yes, and for everyone except readers who love weird Amazonian micro-genres — like Adult Baby Diaper Lovers books, a very real thing — that's probably bad.
Hey, Facebook, I made a metaverse 27 years ago (Ethan Zuckerman — The Atlantic)
- Adventures in early metaversing.
An inconvenient truth about AI (Rodney Brooks — IEEE Spectrum)
- An AI expert on why artificial intelligence won't best humans anytime soon.
6. 1 meta thing: I ate an NFT
So last night my wife and I went to our first-ever NFT party — and I ate an NFT. Kind of.
Why it matters: NFTs — or non-fungible tokens, which are unique and non-interchangeable datum stored on a digital ledge — are the future.
- Or definitely at least a future.
What's happening: Metaversal — a holding company that invests, produces and curates NFTs — held its launch party last night at Kimika, a Japanese-Italian fusion restaurant in Manhattan.
- There were drinks, hors d'oeuvres, and a return to the New York tradition of many people mingling in a small space (with vaccination required).
- The event also involved the release of "Fractals of Taste" from chef Rocco DiSpirito, billed as the first-ever custom recipe NFT, with proceeds from the sale going to the Colorectal Cancer Alliance.
What they're saying: "NFTs give chefs a rare opportunity to codify and memorialize a food experience," says DiSpirito. "'Fractals of Taste' represents my unique view on flavor."
The bottom line: Partygoers had a chance to eat the recipe in edible, non-NFT form.
- I'm not sure how the actual NFT would taste, but the bun I was served was delicious.