Tech Brew Ride Home - Wed. 11/15 – Bridging The Great Blue Bubble Divide

Starting point is 00:00:00 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, November 15th, 2023. I'm Brian McCullough. Today, nothing has come up with the first large-scale way to send eye messages even if you're on Android. Those new OpenAI products are so popular they're actually pausing your ability to use them.

Starting point is 00:00:50 DeepMind has a model that can predict the weather more accurately than humans, and the Cadillac of Web Cameras has a new model. Here's what you miss today in the world of tech. Has the whole blue-bubble, green-bubble device? been bridged, nothing, and I should probably clarify for your ears, the company nothing, is rolling out an early version of nothing chats, which can send iMessages to contacts with iPhones to nothing phone to users in partnership with a startup called Sunbird, quoting the Washington Post. Here's a glimmer of hope for anyone who has ever felt the social stigma that

Starting point is 00:01:31 comes with forcing a text exchange into green bubbles rather than blue, starting on Friday, the London-based device maker Nothing, will roll out an early version of a new messaging app called Nothing Chats. In addition to sending standard text, Nothing Chatters, can send iMessages to friends with iPhones as long as they have one of the company's phone-to devices. For the most part, it really does work. Engage someone in an iMessage conversation from Nothing chats, and they'll see their messages inside Blue Bubbles, same as if they were chatting with another iPhone owner. If they've added your Apple ID to their contact card, those iMessages will appear in a thread alongside your other previously exchanged texts. Sending images in full quality

Starting point is 00:02:12 is generally no problem, and neither is firing off voice memos. When the app goes live this week, Nothing Phone 2 owners will be able to download it directly from Google's Play Store. One of the first times an iMessage on Android tool like this has been made available to many at once. This past weekend, I used Nothing chats to text my partner about a matter of great importance, whose holiday movies are better, lifetimes or hallmarks? And could a big-city iPhone user visiting her small hometown fall in love with a local Christmas tree purveyor who uses an Android? Our comments flew around in blue bubbles for a while until we decided that Hallmark

Starting point is 00:02:47 was the clear winner. Nothing about the conversation felt out of the ordinary, except that I was sending eye messages from an Android phone. So what's the catch? Apart from needing a specific phone, there are key limitations that Sunbird, the New York startup that developed the messaging system hasn't worked out. For starters, you can't edit messages you've previously sent like you can in iOS. Tapbacks, the micro responses you can use to react to iMessages don't fully work yet. Group chats work only if everyone in the conversation is on

Starting point is 00:03:17 iMessage, and sometimes messages just don't go through unless you send them a few times. The experience is still very much a work in progress, but still blue bubbles from an Android phone. That said, nothing's limited release means it's not likely you will be able to try this out for yourself. The company says that it is sold about six figures worth of phone two devices in North America, Britain and Europe, the regions where the messaging app will be available. That means there are at least 100,000 Android phone owners who will soon be able to bust through the green bubble barrier, judging by the popularity of other projects trying to make iMessages available on Android. The interest in breaking down those walls is real.

Starting point is 00:03:56 Sunbird, which maintains its own version of the iMessage-combatable messaging app for other Android smartphones, has around 150,000 people waiting to get a taste of Blue Bubble Life. Unfortunately, their weight seems to have just gotten longer. In an email, Sunbird, CEO Danny Mizrahi said, For the next few months, the only way to get Sunbird is to have a nothing phone too. And as of early November, Beeper, an open-source project that lets Android users access iMessage and other messaging services from one app has more than 130,000 people on its own. wait list, end quote. Intel has patched a bug affecting virtually all modern Intel CPUs that lets

Starting point is 00:04:39 code running inside a virtual machine crash hypervisors, which would be a big risk to cloud providers in particular, quoting Ars Technica. The flaw affecting virtually all modern Intel CPUs causes them to, quote, enter a glitch state where the normal rules don't apply. Tavis Ormandy, one of several security researchers inside Google who discovered the bug, reported. Once triggered, the glitch state results in unexpected and potentially serious behavior, most notably system crashes that occur when untrusted code is executed within a guest account of a virtual machine, which, under most cloud security models, is assumed to be safe from such fault. Escalation of privileges is also a possibility.

Starting point is 00:05:20 The bug tracked under the common name Reptar and the designation CVE 2023-23-3-58 is related to how affected CPUs manage prefixes, which changed the behavior of instructions sent by running software. Intel X64 decoding generally allows redundant prefixes, meaning those that don't make sense in a given context to be ignored without consequence. During testing in August, Ormendi noticed that the Rex prefix was generating unexpected results when running on Intel CPUs that support a newer feature known as fast, fast, short, repeat move, which was introduced in the Ice Lake architecture to fix microcoding bottlenecks. Jerry Bryant, Intel's Senior Director of Incident Response and Security Communications, said on Tuesday that company engineers were already aware of a functional bug in older CPU platforms that could result

Starting point is 00:06:11 in a temporary denial of service and had scheduled a fix for next March. The severity rating had tentatively been set at five out of a possible 10. Those plans were disrupted following discoveries within Intel and later inside Google, end quote. Sam Altman says that OpenAI is pausing new chat GPT plus signups for a bit, as, quote, the surge in usage post-dev day has exceeded our capacity. Quoting search engine journal, to ensure a high-quality user experience for existing users, chat GPT plus subscriptions offering features like the new GPT4 Turbo and custom GPs will be temporarily halted. Interested users can get notifications via a wait list. Following the introduction of

Starting point is 00:07:02 GPTs, developers and companies have built GPTs for a wide variety of purposes, such as graphic design firm Canva. The decision to pause new chat GPD signups follows a week where open AI services, including chat GPT and the API, experienced a series of outages related to high demand and DDoS attacks. Ideally, this means that developers working on building GPs and using the API should encounter less issues like being unable to save GPT drafts, but it could also mean a temporary decrease in new users of GPs since they are only available to plus subscribers, end quote. You know, didn't somebody once famously say that the definition of product market fit is when customers are literally ripping the product out of your hands?

Starting point is 00:07:47 Wasn't that person Sam Altman? I guess a corollary to that would be when your product is so popular that you have to rate limit it to. Apple is extending free usage of its emergency SOS service on iPhone 14 for two more years to September 2025. The company has not yet announced how much emergency SOS will cost when it does start costing, however, quoting 9 to 5 Mac. Previously, Apple gave iPhone 14 customers two free years after device activation, which would have begun expiring this time next year. But now all current iPhone 14 users will be able to use the service for free for another two years. Apple has not revealed how much it will charge for emergency SOS via satellite when the free period is up,

Starting point is 00:08:40 and today's announcement means that the company can defer any decision-making on that further into the future. The free period for iPhone 15 customers has not been extended, however. That means iPhone 15 users still currently are working on a two-free year model, which will begin to elapse in September 2024. This means iPhone 14 and iPhone 15 customers will see their free periods expire at roughly the same time beginning September 2025. It gives Apple more time to decide on pricing plans for these features. Emergency SOS via satellite allows compatible iPhone users to send short text messages to nearby emergency services even when outside of Wi-Fi or cellular signal range. It allows people to get help if they get into trouble in places where

Starting point is 00:09:23 usually they would have no way to contact anyone else. Users can also update their location in the Find My app using the satellite signal with iPhone 15. Apple also extended the emergency SOS feature to include contacting roadside assistance when your car breaks down, end quote. Google Deep Mind has detailed a weather forecasting AI model known as Graphcast, which they claim is more accurate than the best conventional systems for three to 10 day weather predictions. Quoting the Financial Times. Artificial Intelligence has for the first time convincingly outperformed conventional forecasting methods at predicting weather around the world up to 10 days into the future.

Starting point is 00:10:10 The Graphcast AI model marks a turning point in weather forecasting, its developers at Google DeepMind said in a peer-reviewed paper published in the journal Science on Tuesday. An extensive evaluation showed that Graphcast was more accurate than the world's leading conventional system for predictions three to 10 days ahead, which is run by the European Center for Medium Range weather forecasts. It outperformed the ECMWF product in 90% of the 1,30 metrics used, which included temperature, pressure, wind speed, and direction, and humidity at different levels of the atmosphere. Matthew Chantry, machine learning co-coordinator at ECMWF, said AI systems in meteorology had progressed far sooner and more impressively than we expected

Starting point is 00:10:56 even two years ago. EECMWF, an intergovernmental body based in Reading in the UK, has been running live forecasts by AI models from Huawei and Nvidia, as well as DeepMind alongside its own integrated forecasting system. Jantry endorsed DeepMind's claim that its system is the most accurate. Quote, we find GraphCast to be consistently more skillful than the other machine learning models, Pangu Weather from Huawei and ForecastNet from Nvidia, and on lots of scores, it is more accurate than our own forecasting system, he told the Financial Times. Graphcast uses a machine learning architecture called Graph Neural Network, which learnt from more than 40 years of past ECMWF data

Starting point is 00:11:37 about how weather systems develop and move around the globe. The inputs for its forecasts are the states of the atmosphere worldwide at the current time and six hours earlier, assembled by ECMWF from global weather observations. Graphcast produces a 10-day forecast within a minute on a single Google TPUV4 cloud computer. In contrast to this data-derived blackbox approach, the conventional method used by ECMWF and the world's national meteorological offices, known as numerical weather prediction, uses supercomputers to crunch equations based on scientific knowledge of atmospheric physics in energy-intensive process that takes several hours.

Starting point is 00:12:18 Once trained, Graphcast is tremendously cheap to operate, said Chantry. We might be talking about a thousand times cheaper in terms of energy consumption. That's a miraculous improvement. As an example of a successful forecast, deep-mine scientists mentioned Hurricane Lee in the North Atlantic in September. Graphcast was able to predict correctly that Lee would make landfall in Nova Scotia nine days before it happened, in comparison with only six days for traditional approaches, said Remy Lamb, lead author of the science paper. that gave people three more days to prepare for its arrival, end quote. However, AI performed no better than conventional physical models in predicting the sudden explosive intensification of Hurricane Otis

Starting point is 00:12:59 off Mexico's Pacific coast, which devastated Alcapulco with little warning on October 25th. The next step for ECMWF would be to build its own AI model and look at combining that with its numerical weather prediction system, Chantry said, there is room to inject our understanding of physics into these machine learning systems, which can seem like black boxes, end quote. Finally today, remember when we briefly did those interesting gadgets episodes for Ride Home Plus subscribers? Well, this segment would firmly be in that category. Are you familiar with Opel? They make high-end webcams for people for whom looking good on a webcam is important. They have an enthusiastic following among a certain class of extremely online folks, especially Alexis O'Hanian, whom I believe is an investor in the company. Well,

Starting point is 00:13:50 two years after their first product, they've launched a second. Say hello to the Tadpole. Quoting the Verge. Opal has a real hardware team and supply chain now. It was able to go back to suppliers and develop custom parts and tools instead of buying whatever parts it could find in China. It could also actually go to China, unlike in the heat of the pandemic. As a result, the $175 tadpole is not exactly a successor to Opel's first camera, the C-1. It's something different. A webcam designed specifically to be clipped to your laptop not stuck on top of a desktop monitor. But it's also everything the C1 is not. Tiny, compatible with everything, including Windows devices finally, and really, really

Starting point is 00:14:28 easy to use. I spent some time using the tadpole during a recent cross-country trip, and I'm really impressed. The device itself is a tiny square, 1.25 inches in both dimensions and about as thick as a pack of gum. Opel says it weighs about as much as a AA battery. I don't have a scale on me, but holding both in my hands, it's about right. And the point is, it is really small. I get a real iPod shuffle vibe from this thing, and I mean that in a good way. It comes in black and white and has a clip on the back that attaches to your laptop lid and an integrated USBC cable that plugs into your computer. Opel built this device, the company says, because they kept hearing from users who wanted to take their cameras on the road. As people began to go back to the office and on business travel, they still wanted an upgrade on their webcam.

Starting point is 00:15:13 One of the most popular requests Opal received was for a travel case for the C-1. And then we started thinking, okay, why do people want a case? One of the founders says. They discovered that people still wanted to look good, but they needed to be able to do so from an airport lounge, a co-working space, or a conference room. The camera itself is a half-inch 48-pixel Sony IMX-582 sensor, which you might have seen in the OnePlus 7 Pro or some Samsung A-series phones, circa 2019 and 2020, and the ability to capture video goes up to 4K. It's a bigger, better sensor than even the C1,

Starting point is 00:15:49 and it's a vast improvement over your average built-in laptop webcam and roughly on par with what you'd get from a more expensive device like the Insta 360 link. The tadpoles quality is enough of an improvement that I've had people comment on how sharp I looked on calls and is particularly useful in bad lighting. You know that setup where you're sitting right in front of a window and your webcam basically turns you into a silhouette in front of a blown-out background. Stefan Solstrom, Opel's other co-founder says, that's the hardest setup for any webcam to solve, and while the tadpole doesn't totally fix it,

Starting point is 00:16:20 it's noticeably better than my MacBook errors built-in webcam. The tadpole also has a built-in mic that Opo calls the Vizamike. The idea behind the Vizamike was that the TadPol's microphone should only be able to hear what the mic can see, which is a pretty good way to guess what's me participating in the meeting and what's stranger at the next table on a loud phone call. It's a clever idea, but it hasn't really held up in my testing. The tadpole's mic is good, and it does attempt to turn down some out-of-frame sound, but it's not noticeably better to my ears than any other noise cancellation system.

Starting point is 00:16:54 You can mute the mic by touching a capacitive pad on the USB plug, which is neat, but it doesn't integrate with any chat apps. It's just a hardware switch, so you run the risk of the occasional, where am I muted, confusion. Overall, I like the tadpole a lot and can see it being useful for people who travel a lot and need an easy way to upgrade their meetings. And that last bit is the real story here. The tadpole is remarkably easy to use, which couldn't be more unlike the C1, end quote. Okay, something to make you aware of. The Ride Home AI Fund is still raising money if people want in, but that has a minimum check size of $100,000. And the Ride Home Fund is always open to new investors, the Rolling Fund,

Starting point is 00:17:44 but the minimum investment there is $20k or $5,000 a quarter for a minimum of four quarters. But there's a third type of thing that we can do, a syndicate. With a syndicate, if we get some extra allocation in a deal, one of our funds does, we can create what's known as an SPV, essentially a sidecar investment vehicle just for that investment in that one company. It's a case-by-case type of an investment, not a fund. But the advantage of doing an SPV is that the minimum investment in one of those is $1,000. Now, we've never done a sidecar investment vehicle yet, and we might not ever do one,

Starting point is 00:18:28 but there are some potential investments coming down the pike where we might want to consider doing one or two sidecars. There's not an investment deal on offer yet, but if you want to know when one does become available, the two bottom links in today's show notes are to two separate syndicates. One is for the Right Home AI Fund, and the other is for the Right Home Fund, the Rolling Fund. Again, there are no investments on offer from these syndicates yet, and there might never be. But in case we do do one of these sidecars soon, if you want to hear about it so that you can consider participating, I just wanted to offer you all the chance to join the syndicates now so that you can be aware of the deals if and when they come. It doesn't cost anything.

Starting point is 00:19:10 you're not required to invest in deals when and if they come, you can decide on a case-by-case basis whether to participate or not, but you do need to sign up to be notified of any deals. So if interested, hit up those links at the bottom of today's show notes. Talk to you tomorrow.

Tech Brew Ride Home - Wed. 11/15 – Bridging The Great Blue Bubble Divide

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.