Tech Brew Ride Home - Mon. 05/24 - Reading The Tea Leaves Of The Epic/Apple Trial

Episode Date: May 24, 2021

After Tim Cook’s testimony, can the judge’s line of questioning give us a hint of how the trial might resolve as arguments close today? Citizen App is considering rent-a-cops as a service. The AI community wants to open source large language models to address recent controversies. And a computer chip that changes its structure to thwart hackers. Sponsors: PorkBun.com/tld/design Cybereason.com Links: Tim Cook plays innocent in Epic v Apple’s culminating testimony (Tech Crunch) Judge in Fortnite case holds Tim Cook's feet to the fire over App Store competition (Protocol) HomePod and HomePod mini to support Apple Music Lossless in a future software update (9to5 Mac) Leaked Emails Show Crime App Citizen Is Testing On-Demand Security Force (Vice) The race to understand the exhilarating, dangerous world of language AI (Technology Review) Morphing computer chip repels hundreds of professional DARPA hackers (New Atlas) Learn more about your ad choices. Visit megaphone.fm/adchoices

On April 4th, 2023, around 2 in the morning, a man was found stabbed multiple times on a sidewalk in downtown San Francisco. Hey, who did this to you? What happened next turned the story into a political firestorm. Reports have identified the victim as Bob Lee, the founder of Cash App. From Bloomberg Podcasts, this is Foundering, the Killing of Bob Lee, beginning April 16. Welcome to the Techmeme right home for Monday, May 24th, 2021. I'm Brian McCullough today. After Tim Cook's testimony, can the judge's line of questioning give us a hint as to how the trial might resolve as arguments are closing today? Citizen app is considering Rennacops as a service. The AI community wants to open source large language models to address recent controversies and a computer chip that changes its structure to thwart hackers. Here's what you miss today in the world of tech. Well, guess what? Just like that, the Apple versus Epic trial is already over. Closing arguments are today, in fact, but what we missed because of the weekend was the headliner, Tim Cook's testimony on Friday, quoting TechCrunch. Rather than a fiery condemnation of Epic's shenanigans and allegations, Cook offered a mild, carefully tended ignorance that left many of the lawsuit's key questions unanswered or unanswerable.

The facade of innocent ignorance began when he was asked about Apple's R&D numbers, $15 to $20 billion annually for the last three years. Specifically, he said that Apple couldn't estimate how much of that money was directed toward the App Store because, quote, we don't allocate like that, i.e., research budgets for individual products aren't broken out from the rest. Now, that doesn't sound right, does it? A company like Apple knows down to the penny how much it spends on its products and research, even if it can't be perfectly broken down, an advance in macOS code may play into a feature

in the app store. The company must know, to some extent, how its resources are being deployed and to what effect. The differences between a conservative and liberal estimation of the app stores R&D allocation might be large in the hundreds of millions, perhaps, but make no mistake, those estimates are almost certainly being made internally. To do otherwise would be folly. But because the numbers are not publicly declared and broken down, and because they are likely to be somewhat fuzzy, Cook can say truthfully that there's no single number like to invent an amount. App Store R&D was, say, $500 million in 2019. Cook then deployed a similar strategy of starving the competition with a preemptive shrug about

profits. He only addressed total net sales, which were about $275 billion at a 21% profit margin, saying Apple does not evaluate the App Store's income as a standalone business. Certainly, it is arguable that the App Store is very much a tightly integrated component of a larger business structure, but the idea that it cannot be addressed as a standalone business is ludicrous. It is, again, nearly certain that it, like all of Apple's divisions and product lines, is dissected and reported internally in excruciating detail. But, again, it is just plausible that for legal purposes, it is not straightforward enough to say,

The income and profits of the app store are such and such, thus denying Epic its datum. The mask slipped a tiny bit, however, when Epic's attorney asked Cook to break down the confidential income numbers that combined the Mac and iOS app stores, while Apple objected to this saying it was privileged information and could only be divulged in a closed court, Cook offered that the iOS numbers are, quote, a lot larger than the Mac numbers. What we see here is another piece of financial sleight of hand. By mixing the iOS and Mac income, Apple gets to muddy the water. of how much money is made and spent in and on them. Epic's attempt to unmix them was not successful,

but the judge is no fool. She sees the same things Epic does, but just as dimly. Apple is attempting to deny Epic a legal victory, even at the cost of looking rather shadowy and manipulative, end quote. Given all of that, in fact, it is interesting to hear how the judge parsed Cook's testimony, because maybe I've been right all along that it'll end up being gaming that will finally crack open the App Store, quoting protocol. Judge Yvonne Gonzalez-Rogers, who is presiding over the Epic v. Apple antitrust trial, saved her best for last. As Apple CEO Tim Cook prepared to leave the stand on Friday afternoon on the 15th and final day of courtroom testimony, Gonzalez-Rogers took nearly 10 minutes, the longest singular line of questioning she's put to a witness in the trial to grill

cook about both the business model of the app store and the very nature of its relationship with developers. The end result was the best hint yet how Gonzalez-Rogers is thinking about the Fortnite dispute, which one of Epic's many complaints she finds credible, and how she may decide to rule when the trial ends. In particular, the judge seems concerned about the rigidness of the 30% cut and Apple's rules against allowing developers to communicate ways to purchase digital goods off platform. At the beginning of her questioning, she got Cook to admit that gaming, and in particular in-app purchases for mobile games on the iPhone generate the majority of the app stores revenue.

From there, Gonzalez-Rogers forced Cook to answer a series of increasingly uncomfortable questions about whether Apple's conduct with regard to game developers is fair and not, in fact, anti-competitive. Quote, what is the problem with allowing users to have choice, especially in a gaming context to find a cheaper option for content? She asked Cook, who answered that users have, quote, a choice between many different Android models and an iPhone, end quote, if they're looking for choice. Quote, if they wanted to go and get a cheaper battle pass or V-bucks and they don't know they've got that option,

what is the problem with Apple giving them that option? The judge asked. Apple's App Store guidelines forbid developers from notifying users of alternative payment options, and both sides have argued in court about the limits of Apple's restrictions, though Cook has clarified that developers can ask for a user's email address and market to them over email. Cook, who seemed taken aback at the line of questioning, said Apple needs to get a return on its intellectual property investment, and that's essentially why the app store takes 30 percent and why the model has been so rigid over the years. Quoting the judge, the gaming industry seems to be generating a disproportionate amount of money

relative to the IP you have given them and everyone else, in a sense, it's almost as if they are subsidizing everybody else, Gonzalez-Rogers said. Even Cook's rather deft attempts to deflect didn't say. seem to work out well for Apple. He said that Apple allows free apps on the iPhone as well as apps selling physical goods without requiring they pay a commission because, quote, it increases the traffic on the store dramatically, end quote. Gonzalez-Rogers said that using that logic, the app store's design is not about a return on investment so much as it is about controlling access to the iPhone customer base. And with regard to games, quote, you're charging

the gamers to subsidize Wells Fargo, end quote. Quoting the judge again, I understand this notion that somehow Apple's bringing the customers to the users, but after that first time, after that first interaction, the developers are keeping the customer with the games. Apple's just profiting off that, it seems to me, Gonzalez-Rogers said, nearing the end of her questioning. At that point, Cook simply said, quote, I view it differently than you do. The judge added, quote, it doesn't seem to me you feel any pressure or competition to actually change the manner in which you act to address the concerns of developers, end quote. The general takeaway is that Gonzalez Rogers expressed deep skepticism of Apple's claims that it operates

the app store the way it does out of the goodness of its heart. Apple executives have reiterated throughout the trial that they built iOS and the app store this way out of a concern for security and privacy for an end-to-end experience. But Gonzalez-Rogers says there were also clear financial incentives to do so and that it appears Apple is incapable of responding to any concerns that may threaten the benefits it receives, end quote. Dun-d-dun-dun. Remember how Apple announced new Apple Music Losless audio, but very few of their hardware devices actually supported it?

Apple now says the HomePod and HomePod Mini will receive a software update in the future to natively support Apple Music Losless, which cool, but still doesn't explain why this is an after-the-fact matter, quoting 9-5 Mac. Although we now know all models of HomePod will receive support for lossless audio, it's still unclear whether you'll need a pair of home pods to enable this higher quality or if only one smart speaker is enough. For example, HomePod Mini will not

support Dolby Atmos. The original HomePod does, with videos, at least with a pair. About AirPods Max, Apple confirms it doesn't stream lossless and high-res lossless over Bluetooth, but with the lightning to three and a half millimeter audio cable, this headphone will deliver, and quote, exceptional audio quality. However, given the analog to digital conversion in the cable, the playback will not be completely lossless, end quote. Apple also clarifies that broadcast radio, live radio, on-demand content from Apple Music One and music videos won't support lossless audio either. iTunes purchases also cannot be downloaded again in lossless as it's only available from the Apple Music catalog. The company also reinforces that by listening to high-res lossless quality,

the user will need an external digital-to-analog converter. To listen to lossless 24-bit 48 kilohertz, users will need a wired connection to headphones, receivers, or powered speakers with a compatible Apple device. Apple Music Loslis launches in June, end quote. Leaked emails show that Citizen App is testing an on-demand private security service that will respond to, quote, incidents as requested by app users, which sounds like a terrible idea to me. What could possibly go wrong with this sort of an idea?

Quoting vice, the introduction of in-person private security forces drastically alters the service and potential impact that citizen may offer in the future and provides more context as to why a citizen-branded vehicle has been spotted driving around Los Angeles. The news comes after citizen offered a $30,000 bounty against a person it falsely accused of starting a wildfire, quote, the broad master plan was to create a privatized secondary emergency response network. One former citizen employee told Motherboard, Motherboard granted multiple sources anonymity to protect them from retaliation from the company. Quote, it's been something discussed for a while, but I personally never expected it to make

it this far, another citizen source told motherboard. In short, the product described as security response in internal emails would have citizens send a car with a private security force, to an app user according to the former employee. A private security company working with Citizen would provide the response staff. The former employee added a second Citizen source confirmed this description of the service. Citizen has been actively testing the program with what the company describes as quick response times and instant communication between citizens and security partners according to the emails, end quote. So, private security guards and good old vigilanteism as a service.

Yeah, that's clearly what the world needs right now, right? Maybe Amazon will buy Citizen 2 and merge it with Ring and we'll finally have those private enclaves as described in the classic cyberpunk novel Snow Crash. Of course, Snow Crash is also where the term Metaverse got popularized. So with every passing day, Snow Crash continues to be the predictive text for technology with a dystopian tinge to it. Given the recent controversies over at Google's AI efforts, it's interesting to me that over 500 AI researchers are apparently teaming up to work on a project to build an open source large language model that will be used to conduct research independent of any company, quoting MIT Technology Review. On May 18th, Google CEO Sundar Pichai announced an impressive new tool, an AI system called Lambda that can chat to users about any subject. To start, Google plans to integrate Lambda into its main search portal, its voice assistant

and workplace, its collection of cloud-based work software that includes Gmail, docs, and drive. But the eventual goal, said Pichai, is to create a conversational interface that allows people to retrieve any kind of information, texts, visual, audio, across all of Google's products just by asking. Lambda's rollout signals yet another way in which language technologies are becoming enmeshed in our day-to-day lives. But Google's flashy presentation belied the ethical debate that now surrounds such cutting-edge systems. Lambda is what's known as a large language model, or LLM, a deep learning algorithm trained on enormous amounts of text data. It's not just Google that is deploying this technology. The highest profile language models so

far have been OpenAI's GPT2 and GPT3, which spew remarkably convincing passages of text and can even be repurposed to finish off music compositions and computer code. Microsoft now exclusively licensed GPT3 to incorporate into yet unannounced products. Facebook has developed its own LLMs for translation and content moderation, and startups are creating dozens of products and services based on the tech giants models. Soon enough, all of our digital interactions when we email, search, or post on social media will be filtered through LLMs. More than 500 researchers around the world are now racing to learn more about the capabilities and limitations of these models, working together under the Big Science Project led by Hugging Face, a startup that takes an open science approach to understanding natural language processing.

They seek to build an open source LLM that will serve as a shared resource for the scientific community. The goal is to generate as much scholarship as possible within a single focused year. Their central question, how and when should LLMs be developed and deployed to reap the benefits without their harmful consequences? We can't really stop this craziness around large language models where everybody wants to train them, says Thomas Wolfe, the chief science officer at Hugging Face, who is co-leading the initiative, but what we can do is try to nudge them in a direction that is in the end more beneficial, end quote. It was controversial papers by Timit Jebru and Margaret Mitchell that caused that whole brouhaha on Google's AI team that we've spoken about. So I'm noting this in reference to all of that.

Quoting one more time, because when Google announced Lambda last week, I for one didn't realize how tied up that was in all of these controversies. Google already uses an LLM to optimize some of its search results. With its latest announcement of Lambda and a recent proposal it published in a pre-print paper, the company has made clear it will only increase its reliance on the technology. Noble worries this could make the problems she uncovered even worse. quote, the fact that Google's ethical AI team was fired for raising very important questions about the racist and sexist patterns of discrimination embedded in large language models should have been a wake-up call, end quote. So I guess this is that. Click through on the link story for

an in-depth explainer on all of this. Finally today, here's a novel new way to thwart hackers. What if you created hardware that was so secure it could prevent any sort of bad actions from taking place on the software side, quoting New Atlas. Engineers have designed a computer processor that thwarts hackers by randomly changing its microarchitecture every few milliseconds. Known as Morpheus, the puzzling processor, has now aced its first major tests, repelling hundreds of professional hackers in a DARPA security challenge. In 2017, DARPA backed the University of Michigan's Morpheus project with $3.6 million in funding, and now the novel processor has been put to the test. Over four months in 2020, DARPA ran a bug bounty program called Finding Exploits to Thwart Tampering,

pitting 525 professional security researchers against Morpheus and a range of other processors. The goal of the program was to test new hardware-based security systems, which could protect data no matter how vulnerable the underlying software was. Morpheus was mocked up to resemble a medical database, complete with software vulnerabilities, and yet not a single attack made it through its defenses. Computer scientists are increasingly realizing that hardware can play an important role in security. To design a piece of malware, hackers need to understand the micro-architecture of a processor so they can figure out where to inject their malicious code.

Locking down the system at the hardware level could potentially end the arms race once and for all. That was the design philosophy behind Morpheus. Essentially, the processor starts by encrypting key information, such as the location, format, and content of data. But that's not enough on its own. A dedicated hacker could still crack that code within a few hours. and that's where Morpheus gets clever. The system shuffles that encryption randomly every few hundred milliseconds. That way, even if a hacker somehow manages to get a picture of the entire processor, it'll completely change before they have a chance to act on it.

Quote, imagine trying to solve a Rubik's cube that rearranges itself every time you blink, says Todd Austin, lead researcher on the Morpheus project. That's what hackers are up against with Morpheus. It makes the computer an unsolvable puzzle, end quote. That's all for today. Talk to you tomorrow.