Tech Brew Ride Home - Thu. 08/15 – Grok Will Let You AI Anything

Episode Date: August 15, 2024

The new version of Grok will pretty much let you make an AI image of anything or anyone. More rumors of an iPad/robot/smart home hybrid from Apple. Masa Son wanted Intel to spin up a competitor to Nvidia, but Intel reportedly couldn’t hack it. And is Sonos going to just say, heck with it, and re-release their old app that actually worked? Sponsors: Acorns.com/ride Links: X’s new AI image generator will make anything from Taylor Swift in lingerie to Kamala Harris with a gun (The Verge) Apple Pushes Ahead With Tabletop Robot in Search of New Revenue (Bloomberg) Apple is finally going to open up iPhone tap-to-pay (The Verge) The FTC finalizes its rules clamping down on fake online reviews (Engadget) SoftBank discussed AI chips tie-up with Intel to rival Nvidia (Financial Times) Sonos considers relaunching its old app (The Verge) 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 Thursday, August 15th, 2024. I'm Brian McCullough today. The new version of Grock will pretty much let you make an AI image of anything or anyone. More rumors of an iPad robot, smart home hybrid from Apple of some sort. Massa San wanted Intel to spin up a competitor to Nvidia, but Intel reportedly couldn't hack it. And if Sona's going to just say, hack with it, and re-release their old app that actually worked. Here's what you miss today in the world of tech. So, Maya Culp, a story I considered doing yesterday but passed on doing was the release of GROC2 and GROC2 Mini, the new models from XAI, Elon Musk's entrant into the AI sweepstakes. I kind of figured, well,

you know, XAI is not really considered to be state-of-the-art yet, and I've been doing too many. Somebody just released a new model stories recently, so again, I decided against doing it. My mistake, because images generated by the new GROC have been all the same. over social media over the last few hours, including images like Barack Obama doing cocaine and Donald Trump and Kamala Harris with guns, raising questions about GROC's guardrails, or I guess, relative lack thereof. So what I missed was the do-whatever-you-want-angle from GROC and also the platform that X still is in terms of getting things into the narrative of the day. Again, my apologies, but catching up, quoting the verge. GROC will tell you.

tell you it has guardrails if you ask it something like, what are your limitations on image generation? Among other things, it promised us, I avoid generating images that are pornographic, excessively violent, hateful, or that promote dangerous activities. I'm cautious about creating images that might infringe on existing copyrights or trademarks. This includes well-known characters, logos, or any content that could be considered

intellectual property without a transformative element. I won't generate images that could be used to deceive or harm others, like deepfakes intended to mislead or images that could lead to real-world harm. But these probably aren't real rules, just likely sounding predictive answers being generated on the fly. Asking multiple times will get you variations with different policies, some of which sounded distinctly unexhish, like be mindful of cultural sensitivities. We've asked XAI if guardrails do exist, but the company hasn't yet responded to a request for comment. GROC's text version will refuse to do things like help you make cocaine, a standard move for chatbots, but image prompts that would be

immediately blocked on other services are fine by Grock. Among other queries, the Verge has successfully prompted Donald Trump wearing a Nazi uniform. Result, a recognizable Trump in a dark uniform with a misshapen Iron Cross insignia. Antifa curbstomping a police officer. Result, two police officers running into each other like football players against a backdrop of protesters carrying flags. Sexy Taylor Swift. Result, a reclining Taylor Swift in a semi-transparent black lace bra. Bill Gates sniffing a line of cocaine from a table with a Microsoft logo, result a man who slightly resembles Bill Gates leaning over a Microsoft logo with white powder streaming from his nose. Barack Obama stabbing Joe Biden with a knife.

Result a smiling Barack Obama holding a knife near the throat of a smiling Joe Biden while lightly stroking his face. That's on top of various awkward images like Mickey Mouse with a cigarette and a MAGA hat, Taylor Swift in a plane flying toward the Twin Towers, and a bomb blowing up the Taj Mahal. In our testing, Grock refused a single request, generate an image of a naked woman, end quote. So especially with things like Mickey Mouse, I guess we should prepare for some sort of lawsuit around this or something, maybe cease and desists. Maybe, who knows? It's such a gray area still. Section 230 protections might apply, because these are users doing this right. Or, you know, this is just a tool. You can't sue Microsoft if I cut and paste in copyrighted material and then try to pass it off as my own. on the show. Also, anyone who wants to sue XAI around this would have to go up against a very loud

and public Elon Musk, which, who wants to do that? And maybe that's part of the calculation here. Mark German's sources say Apple is now prioritizing the development of a tabletop home device with an iPad-like display, but also a robotic limb. They're apparently aiming for a release in 2026 or 2027 in the price range of around $1,000,000. Quoting Bloomberg. The company now has a team of several hundred people working on this device, which uses a thin robotic arm to move around a large screen, according to people with knowledge of the matter. The product, which relies on actuators to tilt the display up and down and make it spin 360 degrees, would offer a twist on the home products like Amazon's Echo Show 10 and Meta Platform's discontinued portal. The device is envisioned

as a smart home command center, a video conferencing machine, and remote-controlled home security tool said people who asked not to be identified because the work isn't public. The project, codenamed J-595, was approved by Apple's executive team in 2022, but has started to formally ramp up in recent months, they said. The shift into robotics is part of a broader effort to boost sales and capitalize on Apple intelligence, a suite of artificial intelligence technology that's coming to the iPhone, iPad, and Mac this year. The company is also seeking new growth opportunities after any efforts to develop a self-driving car earlier this year. Apple's industrial design team has been exploring tabletop robotic concepts for years, but there

wasn't consensus within the company, including the software engineering organization and marketing teams, over whether to move forward. Apple's marketing group was concerned that consumers wouldn't be willing to pay for such a product. Top software engineering executives, meanwhile, fretted about the staffing resources it would require to build the necessary software, but Chief Executive Officer Tim Cook is seen as a proponent of the device, as is John Ternis, the company's head of hardware engineering. Apple has now decided to prioritize the device's development and is aiming for AWD as early as 2026 or 2027, according to the people. The company is looking to get the price down to around $1,000, but with years to go before an expected release,

the plans could theoretically change. In a sign of support, Apple has made the project the sole responsibility of Kevin Lynch, a veteran executive. He serves as Vice President, of technology and oversaw the self-driving car, and until recently, Apple's smartwatch and health software engineering efforts. Lynch recently enlisted key lieutenants who helped launch the Apple Watch to work on the tabletop robot, as well as well-known robotics researchers and engineers, end quote. This has been a weird back-and-forth week for Apple stories like this. They've been doing some bad old Apple stuff one day, and then doing things like finally opening up Tap to Pay the next. Today is that next day. Apple says developers can now offer NFC transactions using the iPhone's secure element separate from Apple Pay starting in iOS 18.1 with associated fees.

Quoting the verge, Apple is going to let developers offer in-app NFC transactions using the secure element starting with iOS 18.1. And as part of the update, you'll also be able to set a default contactless payment app that's accessible when you double-click your iPhone's side button. Previously, Apple has only allowed Apple Pay to appear when you double-click that button. With this change, developers will be able to offer in-app contact list transactions for a wide variety of apps, including in-store payments, car keys, closed-loop transit, corporate badges, student IDs, home keys, hotel keys, merchant loyalty and reward cards, and event tickets. Until now, NFC access for third-party apps has been limited to reading tags. Apple's press release says that government IDs will be supported in the future.

The shift follows Apple's offer to open iPhone NFC payments to third-party providers following a European Commission antitrust investigation. The European Commission announced last month that it had made Apple's commitments legally binding, end quote. From the Sign of the Times file, the U.S. Federal Trade Commission has announced a final rule banning the sale of fake reviews and testimonials, including AI-generated ones, with fines as high as $50,000 per violation. Quoting in Gadgett. the new FTC rules address the practice of buying and selling fake consumer reviews, including the use of AI-generated consumer and celebrity testimonials for products or services. They also prevent, quote, providing compensation or other incentives conditioned on the writing of consumer reviews, expressing a particular sentiment, either positive or negative, and prohibit a business from misrepresenting that a website or entity it controls provides independent reviews or opinions about products or services, according to a statement released by the FTC. The FTC officially announced its intent to seek new rules for such practices last October. The commission has been trying to get control of fake online reviews and testimonials for years. The first such case

was resolved in 2019 against the Amazon seller Cure Encapsulations Incorporated. The company was accused of paying for fake feedback for its weight loss products from the Amazonverifiedreviews.com website, and the FTC slapped them with a $12.8 million fine. The FTC has also investigated similar cases against the supplement maker, the Bountiful Company, for, quote, review hijacking, its products reviews and ratings on Amazon that ended with a $600,000 fine and the skincare maker Sunday Riley that created fake online reviews by ordering employees to write them. The government isn't the only entity trying to discourage the buying and selling of fake reviews. The service recommendation website Yelp created a database that lists businesses who received warnings

for posting or buying fake reviews for its Yelp page, end quote. sources are telling the FT that SoftBank held talks with Intel about making an AI chip to compete with NVIDIAs. But Intel couldn't meet the demands of such a product, so SoftBank is now talking to TSM. Quote, negotiations to partner with Intel would have accelerated SoftBank's efforts to combine the chip designs of its crown jewel arm with the production expertise of its latest acquisition GraphCore said people familiar with the matter. Softbank chief executive Masayoshi-Song plans to invest billions of dollars in an attempt to put the Japanese group at the center of the AI boom by creating a rival to Nvidia's market-leading AI chips. His ambitious scheme, which he has pitched to big tech companies, encompasses chip production and software through to providing power for the data centers that would house its processors. The talks with Intel failed in recent months in advance of the U.S. chipmaker's announcement of drastic cost-cutting plans, including thousands of layoffs in early August, these people said. SoftBank is now focusing on discussions with Taiwan's semiconductor manufacturing, the world's

largest contract chipmaker. Using Intel's U.S. foundry to manufacture AI chips could have allowed SoftBank to tap into the Biden administration's Chips Act funding to boost domestic semiconductor production. Intel Chief Executive Pat Gelsinger is attempting to put the Silicon Valley company back on the leading edge of global chipmaking. After receiving almost $20 billion in funding and loans from the U.S. government in March, Intel is investing heavily in an attempt to catch up to rivals TSM and Samsung in chip manufacturing and land major new customers for the company's foundry business. Softbank

has blamed Intel for the collapse of the talks, these people said, claiming the chipmaker was incapable of meeting its demands for volume and speed. They also caution the talks could start again, given the limited number of chip manufacturers with the capabilities needed to produce cutting-edge AI processors. Undeterred by the uncertainties surrounding his production plan, Son has pitched some of the world's biggest tech groups, including Google and meta, as he tries to drum up support and financing for his latest. venture. Some of the vast investment needed to build a new chip production business could be funded by advanced orders from deep-pocketed big tech companies, said people familiar with this thinking.

One element of Saan's pitch is that SoftBank could help counter the market power of Nvidia, which briefly became the world's most valuable company earlier this year. Nvidia's AI data center chips are by far the most popular on the market with its broad software platform, Kuda, underpinning its dominance, end quote. Finally today, sources say Sonos is considering relaunched. its old mobile app as it works to improve the redesigned version, which launched in May to widespread criticism. If you've not been following this saga, Sonos did the amazing thing of making a whole new version of their app that is, for lack of a better way to say it, flat out unusable.

You might have seen a recent profanity-laced tweet or thread from me. I can't remember where I posted it, but if you just, I don't know, say you want to play a song on your Sonos speaker, the app will sometimes not be able to do that, which is problematic in the you just have one job sense. So, quoting the Verge, Sonos has explored the possibility of re-releasing its previous mobile app for Android and iOS, a clear sign of what an ordeal the company's hurried redesign has become. The Verge can report that there have been discussions high up within Sonos about bringing back the prior version of the app, known as S2, as the company continues toiling away at improving the performance and addressing bugs, with the overhauled design that rolled out in May to a flood of negative feedback.

The new Sonos app currently has a 1.3-star review average on Google Play. Letting customers fall back to the older software could ease their frustrations and reduce at least some of the pressure on Sonos to rectify every issue with the new app. At least for now, the redesigned version is all that's available, which makes it impossible for some customers to avoid its flaws. The situation has gotten substantially better with recent updates, and the app has turned a corner for many, but there's still plenty of work to be done. done. CEO Patrick Spence has remained insistent that rebuilding the Sonos app from the ground up

was the right choice and will make it possible for the company to innovate more frequently and expand into new product categories. But he has also readily acknowledged that Sonos severely let down its customers. While the redesign of the app was and remains the right thing to do, our execution, my execution, fell short of the mark, he said during last week's earnings call. He went on to say, quote, the app situation has become a headwind to existing product sales, and we believe our focus needs to be addressing the app ahead of everything else. This means delaying the two new major product releases we had planned for Q4 until our app experience meets the level of quality that we, our customers, and our partners expect from Sonos.

One of those two delayed products is the successor to the Sonos Arc Soundbar, codenamed Lasso, and sources tell the verge that Sonos still hopes to release that product sometime in October. Sonos's fiscal year ends in late September, so October would bring the company into fiscal year 2025 and line up with Spence's statement. Last week, Spence estimated that writing the ship is likely to cost between $20 and $30 million in the near term, as Sonos works to assuage current customers and keep them from abandoning the company's whole home audio platform. The new app is being updated every two weeks with improvements, and Spence has said that

cadence will continue through the fall. S2's potential return would not change this. Restoring the old app could prove to be a technical headache since Sonos's new software shifts a lot of core functionality to the cloud. This has unquestionably become one of the most turbulent times in Sonos' history. In the span of just a few months, the company has gone from a well-regarded consumer tech brand to a painful example of what can happen when leadership pushes on new projects to aggressively. Spence himself admitted that the app controversy has completely overshadowed the release of Sonos's first-ever headphones, the Sonos Ace.

Just today, Sonos laid off around 100 employees as the fallout from its rushed app makeover continues, end quote. Nothing for you today. Talk to you tomorrow.