Tech Brew Ride Home - Tue. 11/28 – NameDrop Is Safe, Despite What Police Say

Episode Date: November 28, 2023

AI foundation models are coming to more AWS products. Why are police departments scaremongering about that NameDrop feature on iPhones? Why is Google Drive losing peoples’ files? And Ikea’s new super cheap smart-home starter devices. Sponsors: Miro.com/podcast Traceroute Podcast Links: AWS enhances AI services with foundation model capabilities for improved performance (SiliconAngle) Amazon wants businesses to use its palm-scanning tech to let employees into the office (CNBC) US Thanksgiving weekend sales hit record on big discounts, online boost (Reuters) NameDrop is safe. The fearmongering about it is not. (Washington Post) Google investigating missing files on Drive, caused by desktop app (9to5Google) OpenAI Is Still an $86 Billion Nonprofit (Matt Levine/Bloomberg) Ikea debuts a trio of affordable smart home sensors (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 Tuesday, November 28th, 2023. I'm Brian McCullough today. AI Foundation models are coming to more AWS products. Why are police departments scaremongering about that name drop feature on iPhones? Why is Google Drive losing some people's files

and IKEA's new super cheap smart home starter devices? Here's what you miss today in the world of tech. Again, the AWS Reinvent Conference kicked off in Vegas yesterday, but the actual keynote doesn't happened for another few hours today, I guess. I don't know. I'm honestly confused about their timeline. So there may be some more headlines from that later, but what has already trickled out is that they have decided to add AI Foundation Model Enhancements to their automated speech recognition service known as Transcribe, now supporting more than 100 languages, quoting Silicon Angle. The enhanced capabilities include Amazon Transcribe now offering Foundation Model-powered language support and AI-enhanced call analytics. Amazon Personalize now using foundation models to generate more compiling content,

and Amazon Lex now using large language models to provide accurate and conversation responses. The new FM-enhanced Amazon Transcribe, Amazon's automatic speech recognition service, delivers what AWS says, is significant accuracy improvements between 20 and 50% across most languages. The new ASR system provides differentiating features across all supported languages, now more than 100 related to ease of use, customization, user safety, and privacy. Example features include automatic punctuation, custom vocabulary, automatic language identification, speaker diarization, the process of identifying and separating different speakers in an audio recording, word level, confidence scores, and custom vocabulary filters.

The support for a large number of languages and value-added feature sets is claimed to empower enterprises to unlock rich insights from their audio content and increase the accessibility and discoverability of their audio and video content across various domains. Amazon Personalize, Amazon's machine learning service designed to help developers build personalized recommendations for their customers, now offers hyper-personalization with foundation models through a feature called Content Generator. The new feature uses natural language to create simple and engaging text that describes the thematic connections between recommended items.

According to AWS, this enables companies to automatically generate engaging titles or email subject lines to invite customers to click on videos or purchase items. items. AWS also now offers personalized on the open source Langchain framework to allow customers to build their own foundation model-based applications. With the integration, users can invoke Amazon personalize, retrieve recommendations for a campaign or recommender, and feed them into their FM-powered applications with the Langchain ecosystem. Finally, Amazon Lex, Amazon's fully managed AI service for building conversational interfaces into any application using voice and text, is also getting foundation model-powered capabilities to build bots faster and improve containment, end quote.

Also, Amazon is launching Amazon One Enterprise into preview, allowing U.S. businesses to use its palm scanning technology as an authentication tool to let employees into the office, as an example. Quoting CNBC, companies already signed up for the biometric technology include IHG hotels and resorts, turnstile manufacturer Boone Adam, and Kone, an escalator and elevator provider. Amazon didn't disclose pricing for the service, which is available in preview in the U.S. starting Monday. Amazon is pitching the service as a cheaper and more secure solution for enterprises compared with traditional security and authentication tools like fobs, badges, or passwords. Amazon debuted its biometric payment system in 2020. The technology was originally conceived as a way for shoppers to pay for purchases by swiping their hand. It's been deployed in Whole Foods grocery stores, some Go convenience smarts, and

several fresh supermarkets. Sports and entertainment venues have also adopted the technology, as have some Panera bread restaurants, and a handful of Starbucks locations. Advocacy groups have criticized the service over privacy and security concerns, contending it leads to increased surveillance. Amazon says palm recognition is more private than other biometric systems, quote, because you can't determine a person's identity by looking at an image of their palm, end quote. The company also claims it doesn't capture purchase data from scans collected by non-Amazon stores. By reimagining Amazon 1 for enterprises, the company is following a playbook it's adopted elsewhere. Earlier this month, Amazon debuted Astro for Business, a version of its household robot,

that it's framing as a roving security guard for businesses, end quote. According to Adobe, U.S. Cyber Monday sales rose 9.6% year over year to a record $12.4 billion, driven by deep discounts that peaked at 31% in the electronics category and 23% for apparel. Quoting Reuters. Choppers are leaning on flexible features like Buy Now Pay Later or BNPL services, which help consumers avoid the extra fees and interests that come with credit card payments to stretch their Christmas budgets. A record $940 million worth of purchases were made through BNPL on Cyber Monday,

surging 42 and a half percent from last year and trouncing Adobe's earlier estimate for an 18.8 percent jump. Online consumer spending jumped 9.6 percent on Monday from the $11.3 billion seen last year, outstripping Adobe's initial expectations for a 6.1% rise to $12 billion, end quote. Public service announcement for you here, according to Shira Ovida in the Washington Post, a bunch of local U.S. police departments have been issuing warnings about the privacy risk of the iPhone's new name drop feature used to wirelessly share contacts. The only problem is, this is fake news, basically.

So if you were at Thanksgiving and your uncle warned you about this, said he saw something on Facebook about it, you can tell him to relax. Quote, Chester Wineski, a digital security specialist at Sophos, called the warnings about NameDrop hysteria and nonsense. So you shouldn't worry about name drop, but you should worry that police and news organizations are failing you by sounding false alarms about technology. AirDrop works with Apple devices that are within roughly 30 feet or so of one another, and if you have both Wi-Fi and Bluetooth turned on for your phone. Apple says that NameDrop only works if your Apple device is within a few centimeters of another one. For those of us clueless about the metric system, that's roughly an inch or two. Each of the devices needs to be

unlocked for NameDrop to work, and you have to pick which pieces of contact information you want to share. There is always a risk that a creep or crook gets within a couple inches from your phone, and you mistakenly agree to share your contact information with them, but it's unlikely that you'll share contact information without intending to. Why then have police departments in Pennsylvania, Ohio, Oklahoma, and other places posted similar Facebook messages, warning about the privacy risk of name drop. I don't know, end quote. However, this, unfortunately, does seem to be a real problem. Google says it is investigating an issue caused by Google Drive's desktop app after users reported in recent days that files have been disappearing, quoting 9 to 5 Google. To summarize the ongoing

situation, Google Drive users have noticed that suddenly months or even years of files have gone missing. The issue simply eliminates the data from the account, almost reverting the account back to before that data was made. In some cases, spreadsheet data would be missing from recent weeks, months or years, but present from a former date. Obviously, it's a concerning problem, especially for business users. In a brief post on its community forums, Google formally acknowledged the missing file issue and confirmed that the problem stems from the drive for desktop app. While there's no explanation for the bug, Google says that a, quote, limited subset of drive for desktop users, are seeing the problem and that it's related to versions 84.0.0.0.0 through 84.0.4.4.0. which recently rolled out. Google says it is still investigating the problem, so there's no word of a fix just yet.

Google further tells users to avoid disconnecting their Google account from the drive for desktop app and that they should not delete or move data from the DriveFS folder on their machines located under app data, local Google, DriveFS on Windows, and library application support Google DriveFS on MacOS. Google does recommend making a copy of this folder, though, if there is enough space on the system, end quote. Remember the days when we regularly turn to Matt Levine's newsletter to help us make sense of things like Elon buying Twitter and the whole FDX blow up? Well, the good old days are back, I guess, because here's Matt's take on how things concluded with the whole Sam Altman saga. Quote, University of Kentucky law professor Alan James Clugel emailed me to suggest a theory

that maps the open AI conflict between its nonprofit board that worried about AI safety and its employees who love Altman and wanted him back onto a straightforward valuation dispute. Quote, the board is made up of AI evangelists. The reason they openly worry about AI getting too powerful is out of a belief in the potential for a godlike AI, or at least out of concern that this soon-to-be ubiquitous technology should be in its best possible shape before being distributed to the world. The employees, however, are familiar with all of the AI's limitations and problems and costs, and being Silicon Valley veterans, are also familiar with the hype cycle at play here.

In other words, this is a story about the employees wanting to secure the bag, while the unrealized potential of their product has captured everyone's attention and imagination, and Sam Altman's fundraising, and the Thrive Capital Tender offer in particular, was going to be their golden ticket until the starry-eyed board killed their payday in a flurry of techno-optimistic excitement, end quote. This is absolutely not at all what anyone was saying, and I suspect that it is not what anyone was thinking, but I like it as an objective explanation of what they were doing. It's not unheard of for a startup to get a pretty high valuation and for its employees to think, hey, let's cash out while the money is there, while its board members are venture capitalists

with diversified portfolios and liquidation preferences who are more willing to wait and gamble. Venture capitalist board members are supposed to be able to take the long view and bet on changing the world, while employees are often more risk-averse and need cash to pay the mortgage. OpenAI's board members are not venture capitalists, don't own equity at all, are not motivated by hopes of a trillion dollar valuation and were in fact adverse to its venture capitalist investors. And yet, I think the model applies. They took a very long and grandiose view of the importance of their product and its ability to change the world, while the employees would like to see some cash now. Also, I really do have to quote last week's incredible Wall Street Journal report

about the board's non-explanation of what Altman did to get himself fired. Quote, on the call, the leadership team pressed the board over the course of about 40 minutes for specific examples of Altman's lack of candor, the people said, the board refused, citing legal reasons, the people said. The board agreed to discuss the matter with their counsel. After a few hours, they returned, still unwilling to provide specifics. They said that Altman wasn't candid and often got his way. The board said that Altman had been so deft, they couldn't even give a specific example, according to the people familiar with the executives, end quote. Without realizing it, we were gradually overmatched by a superior intelligence until he ended up controlling us in ways

that are too subtle for us to even explain, thought the AI-nervious board of OpenAI? I love them. Their fears about rogue AI are such obvious metaphors for their mundane real-life problems. Finally, it is a long-running schick of this column that whenever I take a day off, Elon Musk does something crazy. I took much of last week off for Thanksgiving. I guess by Monday Elon Musk is going to own Open AI and Binance I threaded, but I wasn't really worried until OpenAI's new board was announced. The chairman is Brett Taylor, who was also the chairman of the board of Twitter when Musk bought it. I don't really think that Musk is going to buy OpenAI, but I am going to take some time off for the holidays in December, so who knows, end quote. Finally today, some

interesting gadgets. What if I told you you could set up some very basic smart home things on the cheap? And by cheap, I mean very cheap, like $10 cheap. Well, you can soon, thanks to IKEA. Quoting the verge. IKEA's push to democratize the smart home continues with the introduction of three new Zigby sensors that won't break the bank. There's Parasol, the door and window sensor, Valhorn, the motion sensor, and bad ring, the water leakage sensor. They'll all be priced at less than $10 when they go on sale globally in the first half of next year.

Parasol is a typical window and door sensor that can be discreetly mounted to trigger an automation when an open-slash-close event is detected. It can also be paired directly with an IKEA light bulb right out of the box without needing to buy and configure an IKEA home smart hub. It's priced at 9 euros 99 cents in Europe, but exact US pricing is yet to be confirmed for it or any of IKEA's three new sensors. The Valhorn motion sensor can be used both indoors and outdoors with IP44 splash protection against rain to activate lights or other automations when movement is detected. It's powered by three AAA batteries and can be paired to directly control up to 10 IKEA smart bulbs. right out of the box. It costs just 7 euros 99 and can sense more of the room than IKEA's existing $14.99-motion sensor that's smaller, but only usable indoors and needs its own

coin cell battery replaced more frequently. The battering water sensor includes a built-in siren that can alert you when it senses a leak. It can also trigger a mobile notification in the IKEA Home Smart app for Homes with an IKEA DiraJera Hub, $69.99 installed. Sensors like these can save homeowners a ton of money before a water leak has the opportunity to create real damage. It'll cost $9.99. Of the three, the Parasol and Badring sensors are not compatible with older Tradfri Home Smart Gateway's, all support the newer Dearygera Hub, of course, which fully integrates the sensors into IKEA's burgeoning lineup of smart home products and home smart apps.

The hub also allows IKEA's devices to interoperate with smart home ecosystems from Google, Amazon, and Apple when at home or away. Despite Matter support being just, quote, a couple months away, over a year ago, the company still doesn't support it. IKEA plans to eventually enable matter by turning on the thread radio inside the Derrigerra hub to bridge IKEA's existing Zigby-based devices to the next-generation smart home standard in a statement sent to the verge. IKEA Digital Product Area Manager Jonas Soderquist said the company has decided to delay this

functionality and will provide an update when it's time. The ongoing delay is understandable given that IKEA's products are already integrated well with other platforms, and the company is focused on keeping things as simple as possible for anyone who delves into the smart home on a whim while shopping for a new bookcase. And when you consider the hurdles required to get everything running on matter networks during this period of transition, IKEA's delay is more than justified, end quote. Hey, does anybody know of an AI avatar service that allows you to use your own voice? Like, I know things like Synthesia, you can pick any number of avatars, but you can only input text to make those avatars talk, and you have to use the voices that they provide. But what if I want to

use an avatar for video, but I want to upload my own voice. Like, I literally want to upload an mp3 file that I recorded using my voice instead of inputting text, but have it come out of the avatar's mouth. Does anybody know of somebody doing something like that yet? If so, can you let me know who's doing it so I can test it out? Thanks in advance. Talk to you tomorrow.