Month: August 2024

Enlarge / Human driver will continue to be a job at the Boring Company Las Vegas Loop for the foreseeable future. (credit: PATRICK T. FALLON/AFP via Getty Image)

Autonomous driving capabilities are a central component of Tesla's stratospheric share price, with CEO Elon Musk repeatedly telling investors that they're the difference between "being worth a lot of money or worth basically zero." But real-world performance on the road lags far behind Musk's claims, with the latest data point coming from another Musk venture, the Boring Company, and its tunnels under Las Vegas.

The Boring Company might be Elon Musk's strangest side hustle. Whether it was sparked by a desire to avoid traffic commuting to SpaceX or part of an insidious plan to undermine rail projects, the results for the sewer-sized tunnels have been about what you'd expect: Proposed tunnels between Washington DC and Baltimore, underneath I-405 in Los Angeles, and from Chicago to its major airport remain literal pipe dreams.

So far, there's just a 2.2-mile loop with three stations serving the Las Vegas Convention Center, albeit with the potential to expand the subterranean system to 68 miles (110 km) in total.

Enlarge / Telegram CEO Pavel Durov at Mobile World Congress in Barcelona, Spain, on Tuesday, Feb. 23, 2016. (credit: Getty Images | Bloomberg)

Telegram CEO Pavel Durov is heading to court for a possible indictment after being released from police custody, authorities in France said on Wednesday. "An investigating judge has ended Pavel Durov's police custody and will have him brought to court for a first appearance and a possible indictment," according to a statement from the Paris prosecutor's office that was quoted in an Associated Press article.

Durov was arrested in Paris on Saturday and questioned by police for several days. The French investigative judge will "decide whether to place him under formal investigation following his arrest as part of a probe into organized crime on the messaging app," Reuters wrote today.

"Being placed under formal investigation in France does not imply guilt or necessarily lead to trial, but indicates that judges consider there is enough to the case to proceed with the probe. Investigations can last years before being sent to trial or shelved," Reuters wrote. The judge's decision on a formal investigation is expected today, the article said.

Enlarge / There's a reason the Ariel Nomad works so well in open-world games like Forza Horizon. Now Ariel has developed an electric version. (credit: Ariel)

The low-volume British carmaker Ariel got its start with a reimagining of the iconic Lotus Seven, gaining something of a reputation for building the car that filled Jeremy Clarkson's epiglottis full of bees on Top Gear, back when that was the most-torrented TV show on the Internet. Then it expanded its lineup with the Nomad, which added off-road ground ability to the mix, creating a car that starred on The Grand Tour as well as in the last few iterations of Forza Horizon. Today, Ariel unveiled its idea for an all-electric take on that go-anywhere car with the E-Nomad.

Ariel says the E-Nomad can match the more than respectable acceleration of its gas-powered car, which means 0–60 mph (98 km/h) in 3.4 seconds despite wearing nobbly all-terrain tires. But rapidly accelerating electric vehicles aren't anything special. What does stand out is the E-Nomad's relative lack of mass—it tips the scales at just 1,975 lbs (896 kg), less than half as much as most of the EVs on sale today.

The E-Nomad's 41 kWh battery pack lives behind the cabin, replacing the internal combustion engine and fuel tank you'd find on the regular car. Ariel sourced it from Rockfort Engineering and says the pouch cells offer "best-in-class energy density." In total, the battery pack weighs less than 660 lbs (300 kg) and can send up to 281 hp (210 kW) to the drive unit at the rear. The drive has a peak torque output of 361 lb-ft (490 Nm) and is optimized for mass, weighing 202 lbs (92 kg).

Enlarge (credit: Aurich Lawson | Getty Images)

As far back as 2016, work on AI-based chatbots revealed that they have a disturbing tendency to reflect some of the worst biases of the society that trained them. But as large language models have become ever larger and subjected to more sophisticated training, a lot of that problematic behavior has been ironed out. For example, I asked the current iteration of ChatGPT for five words it associated with African Americans, and it responded with things like "resilience" and "creativity."

But a lot of research has turned up examples where implicit biases can persist in people long after outward behavior has changed. So some researchers decided to test whether the same might be true of LLMs. And was it ever.

By interacting with a series of LLMs using examples of the African American English sociolect, they found that the AI's had an extremely negative view of its speakers—something that wasn't true of speakers of another American English variant. And that bias bled over into decisions the LLMs were asked to make about those who use African American English.