Month: October 2023

Enlarge (credit: Thomas Trutschel / Contributor | Photothek)

The United States Department of Justice is investigating Tesla after a Reuters report revealed in July that the EV maker secretly created a team to divert customer complaints because it had grossly exaggerated its vehicles' driving range. Reuters' source confirmed that "the directive to present the optimistic range estimates came from Tesla Chief Executive Elon Musk."

The driving range was so far below company estimates that many customers assumed their cars were defective. Three customers launched a class-action suit, alleging fraud and false advertising. This mounting backlash over Tesla's overly optimistic driving range estimates came at a tense time for Tesla following an unsuccessful launch of Tesla's Full Self Driving (FSD) Beta—a feature deemed so dangerous that Tesla had to recall 362,758 cars—and a criminal investigation into its Autopilot claims.

The DOJ appears to now be probing all of Tesla's recent missteps. According to a US Securities and Exchange Commission (SEC) filing from Tesla on Monday, the DOJ has sent requests for information and subpoenaed Tesla for documents related to Tesla’s Autopilot and FSD features, as well as documents "regarding certain matters associated with personal benefits, related parties, vehicle range, and personnel decisions."

Enlarge (credit: Getty Images / Benj Edwards)

On Wednesday, Stanford University researchers issued a report on major AI models and found them greatly lacking in transparency, reports Reuters. The report, called "The Foundation Model Transparency Index," examined models (such as GPT-4) created by OpenAI, Google, Meta, Anthropic, and others. It aims to shed light on the data and human labor used in training the models, calling for increased disclosure from companies.

Foundation models refer to AI systems trained on large datasets capable of performing tasks, from writing to generating images. They've become key to the rise of generative AI technology, particularly since the launch of OpenAI's ChatGPT in November 2022. As businesses and organizations increasingly incorporate these models into their operations, fine-tuning them for their own needs, the researchers argue that understanding their limitations and biases has become essential.

"Less transparency makes it harder for other businesses to know if they can safely build applications that rely on commercial foundation models; for academics to rely on commercial foundation models for research; for policymakers to design meaningful policies to rein in this powerful technology; and for consumers to understand model limitations or seek redress for harms caused," writes Stanford in a news release.

Enlarge (credit: Getty)

Reddit ignited a war this year. Dramatic changes in API access pricing (from free to unaffordable) was one of its most polarizing moves ever. It resulted in apps beloved by long-time Reddit users, including moderators and people with accessibility needs, closing shop. Community trust was sacrificed, too. Disgusted with Reddit for how it handled third-party apps, abruptly ushered in pricing changes, and treated moderators who protested, numerous valuable, knowledgeable users quit the platform.

Originally, Reddit framed its API pricing changes as a way to prevent generative AI companies from using Reddit data to train large language models (LLMs) without Reddit getting anything in return. With Reddit no longer dealing with small third-party developers—all of which are now either paying Reddit or getting some sort of exemption—Reddit is reportedly taking the fight to where it should have been focused the entire time: generative AI firms.

Can Reddit survive without search?

On Friday, The Washington Post, as spotted by The Verge, said Reddit "has met with top generative AI companies about being paid for its data," citing an anonymous source.

• 

  The guts of the Pixel 8 Pro, once the cooling pad has been removed. Check out that sweet Google logo! [credit: iFixit ]

The Pixel 8 Pro has been out for a few days now, and iFixit got hold of one for a teardown on its YouTube channel.

Opening it up doesn't seem much different from last year. Step one is getting past all the phone glue, which involves heating up the screen, pulling the screen away from the body with a suction cup, and cutting the glue around the edge with a soft pick. iFixit says the adhesive was "easier to get into [than that in the iPhone] 15 Pro Max." Unlike the iPhone 15 design, though, the Pixel 8 Pro only opens from one side, so you'll most likely have to remove every single part in the phone to swap out the back.

The guts of the phone look just like the Pixel 7 Pro, but Google has started to pick up on a few of Apple's tendencies to make the phone's insides look nice. This year the battery is cleanly branded "Google," whereas last year it was just covered in manufacturer information and warnings. Of course, you'll only see this after you peel off the graphite cooling pad, which, for the second year in a row, has seen a big increase in area; it now covers about 60 percent of the battery. The Pixel phone runs a "Tensor" SoC made in partnership with Samsung, so more cooling will definitely help.

Enlarge / Gillian Anderson's likeness has been promoting Star Citizen since 2015. (credit: Cloud Imperium Games)

Eleven years ago, Wing Commander designer Chris Roberts announced Star Citizen, an online multiplayer game that he said would "change the way people perceive games for the PC." Roberts told Ars' Kyle Orland soon after that he didn't enjoy the four-year development of another hit, Freelancer, because "spending that many years disconnected from your audience, sort of working off by yourself, wasn't creatively fun for me." With Star Citizen, Roberts said he could keep development from dragging on by engaging fans and using a pre-built engine, as opposed to what Roberts said would be "two years" building his own.

Roberts has definitely engaged his audience in Star Citizen, to the tune of $616 million raised from more than 4.8 million "Star Citizens." It has just taken a bit longer than two years to give them a true release.

Roughly 11 years after Star Citizen's initial announcement that included it, then nine years after its first potential release date, Squadron 42, the single-player campaign, is now "feature complete" and has "entered its polish phase." Roberts announced this in a video released Sunday as part of an annual CitizenCon for backers, along with footage from the game and details on its development.

Enlarge / Is this "evocative" header image stretching the limits of good taste? (credit: Aurich Lawson | Getty Images)

Here at Ars, we've seen time and again how simple web and/or mobile games can be cloned or outright stolen by unscrupulous developers aiming to cash in on someone else's game concept. But developer Josh Simmons was in a unique position to inflict a particularly rude punishment on websites that were directly stealing and monetizing his web game Sqword without permission.

Since its launch last year, Simmons says he has attracted a "steady group of daily active users" for Sqword, which involves placing letters sequentially in a 5×5 grid to make as many valid words as possible. But as noted on Simmons' blog (and noticed by 404 Media), searching for Sqword also brings up several "game aggregator" sites that simply embed the game content from Sqword.com in an iFrame window, only now surrounded by annoying banner ads.

"This made me angrier than it should have—not because Sqword is a cash cow—we don't run ads on the site and don't make money from it, it's just for fun—but because it was a passion project with friends, something pure and intentionally free to play WITHOUT ads," Simmons writes. "It's against my ethos as a developer, there are banners and popups everywhere. If I build an app, I believe it should either be free or it should be upfront about what the subscription or purchase price is (and then not upsell you). I couldn't abide seeing my code monetized in this way."

Inspiration can come from the most unlikely places, as fantasy author Ursula Vernon, aka T. Kingfisher, clearly knows. Vernon won the 2023 Hugo Award for Best Novel this past weekend for her dark fairy tale, Nettle and Bone, and while she was unable to travel to Chengdu Worldcon in China for the event, she posted the text of her acceptance speech (read at the ceremony by a friend) on her Patreon. After the usual preliminary remarks and thanks, Vernon opted to forego "serious and heavy" commentary for the following revelation:

There is a species of water beetle that regularly gets swallowed whole by frogs. And while there’s a lot of things you can do to keep from being eaten, once you’re inside a frog, your options are severely limited. Generally you get digested. But this particular species of beetle said “You know, I bet there’s another way.” And it started walking. In fact, it walked through the frog’s digestive tract and out the back end.

This is 100 percent true, you can look it up.

Naturally, we did look it up and honestly can't believe we missed covering this fascinating study in 2020. (At least we didn't miss the 2022 study on how certain species of beetle have evolved unusual "back pockets" to safely house symbiotic bacteria during metamorphosis, shuffling the populations out of those pockets via friction to the genital area as they emerge from their pupae.)

Shinji Sugiura of Kobe University in Japan discovered the unusual survival strategy of the aquatic beetle Regimbartia attenuata while looking into how predation pressures can lead to the evolution of innovative escape behavior in prey animals. He fed a bunch of the beetles to a pond frog (Pelophylax nigromaculatus) under laboratory conditions, expecting the frog to spit the beetle out. That's what happened with Sugiura's prior experiments on bombardier beetles (Pheropsophus jessoensis), which spray toxic chemicals (described as an audible "chemical explosion") when they find themselves inside a toad's gut, inducing the toad to invert its own stomach and vomit them back out.

Enlarge (credit: Aurich Lawson)

Bribery. Embezzlement. Terrorism.

What if an AI chatbot accused you of doing something terrible? When bots make mistakes, the false claims can ruin lives, and the legal questions around these issues remain murky.

That's according to several people suing the biggest AI companies. But chatbot makers hope to avoid liability, and a string of legal threats has revealed how easy it might be for companies to wriggle out of responsibility for allegedly defamatory chatbot responses.

Enlarge (credit: Aurich Lawson)

Bribery. Embezzlement. Terrorism.

What if an AI chatbot accused you of doing something terrible? When bots make mistakes, the false claims can ruin lives, and the legal questions around these issues remain murky.

That's according to several people suing the biggest AI companies. But chatbot makers hope to avoid liability, and a string of legal threats has revealed how easy it might be for companies to wriggle out of responsibility for allegedly defamatory chatbot responses.