Linked List: January 11, 2023

Wednesday, 11 January 2023

El Pintador – AI Art Generation App for iOS 

One reason a scam like the aforementioned ChatGPT rip-off is annoying is that it distracts from legitimate apps in the same space. El Pintador is a new iOS app built atop Stable Diffusion that is worth your attention:

Simply tell El Pintador what you would like to see and watch as the AI generates beautiful new art right before your eyes.

In addition to a bunch of helpful templates that serve as starting points, El Pintador features lessons to help people fine tune their prompts. It’s fun, and the team behind it (Xibbon, founded by Miguel de Icaza) has been responsive to feedback since launching it just before Christmas. Also: free of charge to download and try.

App Store Scam of the Week: ‘ChatGPT Chat GPT AI With GPT-3’ 

Sami Fathi, reporting for MacRumors:

ChatGPT is free to use on the web for anyone with an OpenAI account, but it has inspired scammers and sketchy developers to take advantage of its popularity for ill-gained profit. One app in particular, named “ChatGPT Chat GPT AI With GPT-3,” gives the impression it is the official app for the ChatGPT bot, but appears to have no affiliation to Open AI, the creators of ChatGPT, or the bot itself.

The app charges users a $7.99 weekly or $49.99 annual subscription to use the bot unlimited times and eliminate intrusive in-app ads. The app and its bot are inconsistent, sometimes providing generic or entirely irrelevant responses to a prompt offered by the user.

The app is currently the second most popular productivity app on the App Store in the United States, indicating it is rather popular. The app has nearly 12,000 ratings, with a number of positive and negative reviews. “This is a fake app,” one review said. “This is just faking openai endorsement and more bad stuff,” another user said. Despite its suspicious activity, presence, and soaring popularity, the app has passed Apple’s App Store review process multiple times since its initial launch three weeks ago.

As noted by Fathi in an update, the app was pulled from the App Store yesterday. But it should have been flagged sooner. I don’t think it’s feasible to expect App Store reviewers to catch every potential scam app. But as I’ve long argued, I do think it’s reasonable to expect Apple to catch every scam app that makes its way onto the list of most popular apps. They can’t catch every scam app but they ought to be able to catch every successful one.

Surveillance Footage of Tesla ‘Full Self-Driving’ Crash on Bay Bridge 

Ken Klippenstein, reporting for The Intercept:

Highway surveillance footage from Thanksgiving Day shows a Tesla Model S vehicle changing lanes and then abruptly braking in the far-left lane of the San Francisco Bay Bridge, resulting in an eight-vehicle crash. The crash injured nine people, including a 2-year-old child, and blocked traffic on the bridge for over an hour.

The video and new photographs of the crash, which were obtained by The Intercept via a California Public Records Act request, provides the first direct look at what happened on November 24, confirming witness accounts at the time. The driver told police that he had been using Tesla’s new “Full Self-Driving” feature, the report notes, before the Tesla’s “left signal activated” and its “brakes activated,” and it moved into the left lane, “slowing to a stop directly in [the second vehicle’s] path of travel.” [...]

Since 2016, the federal agency has investigated a total of 35 crashes in which Tesla’s “Full Self-Driving” or “Autopilot” systems were likely in use. Together, these accidents have killed 19 people. In recent months, a surge of reports have emerged in which Tesla drivers complained of sudden “phantom braking,” causing the vehicle to slam on its brakes at high speeds. More than 100 such complaints were filed with NHTSA in a three-month period, according to the Washington Post.

The footage of the crash is weird. The Tesla just slows to a stop in the passing lane. When AI systems make mistakes, they often are very different from the sort of mistakes people make.

It ought to be enough for self-driving systems to be safer, mile-per-mile, than human-driven vehicles. That may even be true for Tesla’s “full self-driving” mode, today. 35 crashes in 6 years isn’t a lot. But that’s not going to be enough. It’s not fair, but the bar for self-driving cars is much higher than having a lower accident rate than human-driven vehicles. Even though human-caused crashes are increasing in frequency, they’re so commonplace that they don’t register as shocking. (Same thing with gun deaths in the U.S.) An accident caused by an AI-driven car, though — that’s jarring because that’s novel. “Dog Bites Man” doesn’t make the front page. “Man Bites Dog” does.

As for Tesla’s system in particular, it strikes me as bizarre that it’s legal for them to enable this when they themselves still describe the feature as “beta” software.

Semafor: Microsoft Looking to Invest $10B in OpenAI, Maker of ChatGPT 

Liz Hoffman and Reed Albergotti, reporting for Semafor:

The $29 billion is a big valuation for OpenAI, a company that hasn’t yet figured out its business model, and $10 billion is a big price tag for Microsoft’s shareholders. [...] If OpenAI figures out how to make money on products like ChatGPT and image creation tool Dall-E, Microsoft will get 75% of the profits until it recoups its initial investment.

Beyond the financial risks and rewards for Microsoft, the bigger prize is that it gets to work alongside OpenAI in developing the technology on Microsoft Cloud, which instantly puts Microsoft at the forefront of what could be the most important consumer technology over the next decade.

That’s a huge coup for Microsoft, especially considering Google, a rival, has helped pioneer some of the technology used by OpenAI. Microsoft was also in talks to incorporate some of those features into its other programs, like Word and Outlook email, The Information reported.

Clippy, but actually smart and useful. That’d be one hell of a feature.

Dare Obasanjo, on Twitter:

Satya Nadella cements his entry in the CEO hall of fame by spotting OpenAI early and locking down access to its fruit for use in Office and Bing.

FTC is worried about Activision which is gaming’s past when Microsoft is locking down the future of AI.

It seems way early to credit this investment as a hall of fame decision, but it does seem forward-looking in a way that the Activision acquisition does not.

Executive Turnover at Rivian 

Sean McLain and Nora Eckert, reporting for The Wall Street Journal:

The departures, confirmed by a Rivian spokeswoman, are the latest developments in what has been a challenging period for Rivian, which has been rolling out its first all-electric models but last year missed a critical milestone of manufacturing 25,000 vehicles. The company said it was off its goal by about 700 vehicles in part because of difficulty getting parts.

Rivian’s stock has also tumbled since its blockbuster initial public offering in November 2021, down roughly 79% through Tuesday’s close.

Rivian’s stock price is what it is — but a miss of 700 vehicles from a goal of 25,000 seems pretty good to me. That’s about 97 percent. But for context, Tesla made a record 1.3 million vehicles last year. Rivian obviously has a long way to go to reach a mainstream level of vehicle production.

The executives who have left were some of Rivian’s longer-tenured employees. Among them is Randy Frank, vice president of body and interior engineering, and Steve Gawronski, the vice president in charge of parts purchasing. Both had departed around the beginning of this year. Mr. Frank joined Rivian in 2019 from Ford Motor Co. Mr. Gawronski joined in 2018 from the autonomous vehicle startup Zoox.

When executives who’ve only been at the company for 3 or 4 years are among the “longer-tenured”, I think it’s more a sign of how dynamic the entire electric vehicle industry is at the moment, rather than a mark against Rivian in particular. There’s a lot of turnover (and effectively, musical chairs, as they tend to go from one company to another) at Tesla and Apple’s secretive Project Titan, too. After leading development of Tesla’s Model 3, Doug Field went to Apple, but then left Apple for that plucky startup Ford. But then Apple hired an executive who’d spent 31 years at Ford. It’s a huge market that’s being disrupted from multiple angles.

The Complexity of Evaluating Modern Phone Cameras 

Marques Brownlee’s title for this video is “What Is Happening With iPhone Camera?”, but it’s a nuanced consideration of just how complex phone cameras have gotten in the era of computational photography.