Tech Brew Ride Home - Wed. 01/18 – The HomePod Is Back

Episode Date: January 18, 2023

The original HomePod is basically back, even though it’s not the original cause it’s been updated. Apple’s AR/VR plans continue to be maddeningly confusing, at least to me. Microsoft has announced huge layoffs. Does even the AI revolution require human moderators working in boiler-room like conditions? And I guess we have all the specs for those new Galaxy phones about to be announced. Links: Apple announces revamped full-size HomePod two years after discontinuing original (The Verge) Apple Delays AR Glasses, Plans Cheaper Mixed-Reality Headset (Bloomberg) Full memo: Microsoft to cut 10k jobs, about 5% of workforce, and take $1.2B restructuring charge (GeekWire) Exclusive: OpenAI Used Kenyan Workers on Less Than $2 Per Hour to Make ChatGPT Less Toxic (Time) Galaxy S23 spec leak leaves few surprises ahead of February launch (The Verge) Twitter Manager: Daily Revenue Has Dropped 40%, 500 Top Advertisers Have Left (The Information) 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 Wednesday, January 18th, 2023. I'm Brian McCullough today.

The original HomePod is basically back, even though it's not the original because it's been updated. Apple's AR VR-VR plans continue to be maddeningly confusing, at least for me. Microsoft has announced huge layoffs. Does even the AI Revolution require human moderators working in boiler room-like conditions? And I guess we have all the specs for those new Galaxy phones about to be announced. Here's what you miss today in the world of tech. Surprise. Apple this morning announced an updated HomePod featuring a similar design to the original device from 2018, which, somehow this had slipped my mind, had been discontinued a while ago. But the new one is basically the old one, available for order for $299 today, shipping February 3rd. So I guess it's worth noting that that is about $50 cheaper than the original, quoting the verge. The device has a similar form factor to the original home pod release in 2018.

New features available with this second-generation model include support for the smart home standard matter, allowing for the speaker to control compatible accessories, and an upgraded S-7 chip first seen in the Apple Watch series 7. It will be sold alongside the HomePod Mini, the smaller version of the speaker introduced in 2020. The second-generation HomePod has a single woofer paired with five tweeters, down from seven in the original model. Apple says the speaker can sense how it's positioned in a room, adjusting its sound, depending on whether it's placed up against a wall or out in the middle of a space. Like the original HomePod, two speakers can be arranged in a stereo pair or used as speakers for the Apple TV4K. The original HomePod had an unusually rocky lifespan for an Apple product, originally announced in 2017 with a December ship date.

The $349 smart speaker was subsequently delayed to February 2018. When it was eventually released, the HomePod's voice assistant Siri, fared poorly against Alexa and Google Assistant, and reviewers criticized how locked into Apple's ecosystem the smart speaker was. Despite getting a $50 price cut in 2019, the HomePod was eventually discontinued in 2021. As well as its cheaper price, the HomePod Mini also introduced support for thread, a smart home protocol that subsequently became an integral part of the Matter spec. Like the new HomePod, the Home Mini can serve as a Matter smart home controller and thread border router for management of matter devices across thread, Wi-Fi, and Ethernet networks, end quote.

A couple other notes, in addition to the five tweeters down from seven, not sure what that does to the sound quality. The microphone array is four, down from the original six. Not sure what that does for voice activation. And it requires iOS 16.3, so expect that to come out sometime before February 3rd. This is just downright confusing at this point. Mark German's sources are now saying that Apple has indefinitely postponed its lightweight AR glasses due to technical difficulties and instead planning a lower-cost mixed-reality headset for 2024 or 2025. That doesn't mean we shouldn't expect some sort of headset announcement this year. That's still going to happen. Let's see if we can unpack this quote. Apple is still planning to unveil

its first mixed-reality headset this year, but an even more important follow-up product, lightweight augmented reality glasses has been postponed due to technical challenges. The company had originally hoped to release the AR glasses after the debut of its mixed reality headset, which combines both AR and virtual reality, but that part of the plan is now on hold. Instead, Apple will follow up with a lower cost version of the mixed reality headset as soon as 2024 or early 2025, according to people familiar with the deliberations. Apple's initial dream of offering a lightweight pair of AR glasses that people could wear all day now appears many years away, if it happens at all. High costs could also make an

Apple headset more of a niche product. The initial mixed reality device due this year will cost around $3,000. The hefty price stems from its use of advanced and high-resolution displays, more than 10 cameras, sensors to determine where a user is looking, and both a Mac-grade M2 processor, and a dedicated chip for handling AR and VR visuals. Apple is aiming to lower the price of the follow-up mixed reality device by using chips on par with those in the iPhone rather than components found in higher-end Mac computers. The company will be competing with meta's mixed-reality headset, which costs $1,500. That's a price Apple may strive to get closer to with its lower-end model.

Trademark filings made by Apple hint at the dual-device strategy. The documents have included the names Reality Pro and Reality One. The pro name is likely for the initial model, while the one suffix could be under consideration for the cheaper version. There's also a trademark filing for reality processor hinting at the name for the dedicated chip. The company is focusing on the lower cost headset instead of the AR glasses, which were once planned to be released about a year after the initial headset, said the people who asked not to be identified because the project is still under wraps. At one point, Apple aimed to release the glasses in 2023 before delaying the launch until around 2025. Now Apple has

postpone the rollout indefinitely and paired back its work on the AR device. Apple's first high-end headset will last about two hours on one charge and will require an external battery pack. The headset's battery also will be cumbersome with the company choosing not to include it in the device itself to reduce the weight and heating risks of the headworn product. The battery will connect to the headset over a cable and rest in a user's back pocket. Creating a lightweight device with advanced augmented reality and built-in battery that could last all day isn't feasible with today's technology, end quote. Microsoft plans to lay off 10,000 employees through March 31st, reducing their overall headcount

by less than 5% globally, but will take a $1.2 billion charge to facilitate the move, quoting Geekwire. Citing a need to, quote, align our cost structure with our revenue and where we see customer demand, Microsoft CEO Sachinadella told employees, day that the company will cut 10,000 jobs between now and the end of March. In addition, the company will take a $1.2 billion charge in its second fiscal quarter when it releases its full financial results next week. Nadella told employees that the change is, quote, related to severance costs, changes to our hardware portfolio, and the cost of lease consolidation as we create higher density across our workspaces, end quote. Microsoft is, quote, allocating both our capital and talent to areas of

secular growth and long-term competitiveness for the company while divesting in other areas, Nadella wrote in the memo to employees. These are the kinds of hard choices we have made throughout our 47-year history to remain a consequential company in this industry that is unforgiving to anyone who doesn't adapt to platform shifts, he added. It's the second largest workforce reduction in Microsoft's history. The company cut 18,000 employees in 2014 several months after Nadella became CEO. The majority of those cuts involved Nokia's devices and services business, which Microsoft had acquired under Nadella's predecessor as CEO, Steve Balmer, end quote. Well, for all of the newness of the current and recent AI revolution,

seems like some things are staying the same, disappointingly. An investigation by Time magazine claims that starting in November 2021, OpenAI contracted Kenya-based SAMA to label violent and toxic content, paying workers $2 an hour to improve ChatGPT. Quote, the work was vital for OpenAI. ChatGPT's predecessor, GPT3, had already shown an impressive ability to string sentences together, but it was a difficult sell as the app was also prone to blurting out violent, sexist, and racist remarks. This is because the AI had been trained on hundreds of billions of words scraped from the internet, a vast repository of human language. That huge training

dataset was the reason for GPT3's impressive linguistic capabilities, but was also perhaps its biggest curse. Since parts of the internet are replete with toxicity and bias, there was no easy way of purging those sections of the training data. Even a team of hundreds of humans would have taken decades to trawl through the enormous data set manually. It was only by building an additional AI-powered safety mechanism that OpenAI would be able to rein in that harm, producing a chatbot suitable for everyday use. To build that safety system, OpenAI took a leaf out of the playbook of social media companies like Facebook, who had already shown it was possible to build AIs that could detect toxic language like hate speech to help remove it from their platforms. The premise was simple.

Feed an AI with labeled examples of violence, hate speech, and sexual abuse, and that tool could learn to detect those forms of toxicity in the wild. That detector would be built into chat GPT to check whether it was echoing the toxicity of its training data and filter it out before it ever reached the user. It could also help scrub toxic text from the training datasets of future AI models. To get those labels, OpenAI sent tens of thousands of snippets of tech. to an outsourcing firm in Kenya beginning in November 2021. Much of that text appeared to have been pulled from the darkest recesses of the internet. Some of it described situations in graphic detail like child sexual abuse, bestiality, murder, suicide, torture, self-harm, and incest. OpenAI's

outsourcing partner in Kenya was SAMA, San Francisco-based firm that employs workers in Kenya, Uganda, and India to label data for Silicon Valley clients like Google, Meta, and Microsoft. Sama markets itself as an ethical AI company and claims to have helped lift more than 50,000 people out of poverty. The data labelers employed by SAMA.A.A.A. were paid a take-home wage between around $1.32 and $2 per hour, depending on seniority and performance. For the story, time reviewed hundreds of pages of internal SAMA.I. documents, including workers' pay slips and interviewed for SAMA. employees who worked on the project. All the employees spoke on condition of anonymity out of concern for their livelihoods, end quote. The full specs for the upcoming Samsung Galaxy S23, S23 Plus, and S23 Ultra have allegedly already leaked, showing 6.1, 6.6.6 and 6.8-inch screens,

snapdegon 8 Gen 2 chips, a 200 megapixel sensor, at least on the Ultra, and a lot more. Quoting the verge. Starting with cameras, it appears as though the widespread reports of the Galaxy S23 Ultra coming with a 200 megapixel sensor are correct. Alongside it, the phone appears to have three additional rear cameras, a 12-mapsal ultra-wide, a 10-mapixel telephoto with 3x optical zoom, and a second 10-mapsal telephoto with a 10-X optical zoom. Meanwhile, both the S-23 and S-23 Plus reportedly have three cameras, a 50-mapixel main, 12-mgapixel ultra-wide, and a single 10-magixel telephoto with a 3-X optical zoom. In terms of screens, the Galaxy S-23 Ultra is apparently the largest of the three phones,

with a 6.8-inch 1440p OLED display with a 120-hertz refresh rate. The Galaxy S-23 and S-23 Plus reportedly have 6.1 and 6.6.6-inch OLED displays, both 1080p, both 120 hertz. All three phones are said to have a whole-punch notch with a 12-magpixel selfie camera and an IP68 rating for dust and water resistance. Internally, the leaks suggest all three phones will be powered by Qualcomm's latest flagship processor the Snapdragon 8 Gen 2, at least in the French and German markets where these spec sheets appear to come from. Samsung will almost certainly use the same processor in the U.S. based on previous models, but in the past, it's used its own Exenos processors in some international markets.

This is joined by a choice of 8 gigabytes or 12 gigabytes of RAM in the Ultra, and 256 gigabytes, 512 gigabytes or 1 terabyte of built-in storage. The S-23 and S-23 Plus will reportedly only come with 8 gigabytes of RAM, and between 128 gigabytes and 512 gigabytes of storage. Finally, there's the batteries, which sit at 3,900 mill-amp hours for the S-23, 4,700 milamph hours for the S-23 plus, and 5,000 for the Ultra. The plus and ultra will reportedly support fast charging up to 45 watts versus 25 watts on the base S-23, end quote. Remember, we are expecting these phones to be announced at the upcoming Samsung event on February 1st.

Finally today, from the information, a source is telling them that Twitter manager Sidarth Rowe told Twitter employees that daily revenue at Twitter was 40% lower than a year earlier, and more than 500 top advertisers have paused spending. Quote, while Elon Musk plans to diversify Twitter's revenue away from ads in the long run, continued deterioration of that business will make it harder for the company to break even on a free cash flow basis in 2023, as Musk predict. it would three weeks ago. That may rest on the company's ability to generate three billion dollars in revenue this year, which Musk said the company projected when he took over and

pay one and a half billion dollars a year in interest related to the debt he raised for the $44 billion acquisition. Twitter generated $5 billion in revenue in 2021. The last full year Twitter published financial statements. In the first quarter of 2022, Twitter generated $1.2 billion in revenue, which averages to $13.3 million a day. If Twitter's daily revenue, is running 40% below that figure, Twitter would be currently generating about $8 million a day, translating to $720 million in the first quarter of this year. The ad problems could also prompt Musk to cut more costs after previously axing 75% of the company's 7,500 employees and ordering the closure of one of its data centers. Some large advertisers have balked at Musk's

approach to content moderation, including reinstating the accounts of individuals whom Twitter had previously banned. Just as important, Musk has also fired most of Twitter's salespeople, including many of those who handled the company's major advertisers, as well as around 50 engineers and data scientists working on improving Twitter's ad product. Rao recently replaced Luke Simon, an engineering director who was fired and had been leading Twitter's revenue organization. Clients of Group M, the world's largest ad buying firm, have cut their spending by nearly half since Musk took over, and ad-holding companies Omnicom and Interpublic Group have advised their clients. to pause all advertising on Twitter while waiting to see what Musk and Twitter do next,

the information reported last week, end quote. So I chose that one to be the last story of the day on purpose, so I could tell you this. I announced this on Twitter last week, but in case you missed it, my ride home fund has invested in T2, that new upstart Twitter clone that you heard about on our bonus episode last month when the founder came on the show to talk about it. I, of course, wanted you to be aware of this investment because you should have that in your mind going forward when we talk about Twitter on this show. My intention is to cover Twitter like I always have with a mixture of skepticism and deep, deep love.

But on some level, of course, by investing in a Twitter competitor, some part of me is rooting for them to be supplanted. Again, my intention is for my coverage of and even slant on Twitter to remain unchanged, same as it's been for five years. But in the interest of full disclosure, you deserve to know that this investment has been made so you can judge my biases for yourself. As I've said many times, I've never claimed to be a journalist, so I don't have that sort of neutrality that journalists are supposed to have. I try to tell you what is happening in Silicon Valley every day as close to the metal, as close to the truth as I can

get it, but it is always going to be the truth as I see it, and I know you understand that. What I never want to do is obfuscate anything from you for personal reasons if I can help it. So again, full disclosure. And one last interesting thing to keep in mind, this investment in T2 is being made in this quarter. So if you were interested in investing in T2 yourself, if that's intriguing to you, as long as you invest in the Ride Home Fund before the end of February, your money will go into this T2 investment. So, again, if that sounds interesting to you, go to Ridehomefund.com and put your money in the podcast. Talk to you tomorrow.