Tech Brew Ride Home - Mon 01/30 – Every Platform Is Hustling To Unveil An AI Play

Episode Date: January 30, 2023

Après ChatGPT, prepare yourself for the generative AI deluge. Amazon says the free lunch for grocery delivery is over. More speculation about foldable iPads. San Francisco wants Waymo and Cruise to pump the breaks on their self-driving car experiments because the actual cars have been literally pumping the breaks too much. And my wager on whether or not 2023 will actually go down as the year that AI breaks through. Sponsors: The Business of Tech (Listener ad!) RelationshipHero.com/techmeme (Listener ad!) Links: Chinese Search Giant Baidu to Launch ChatGPT-Style Bot (Bloomberg) Amazon adding new fees for grocery delivery (GeekWire) Apple Supplier in India Begins Making Components for AirPods (Bloomberg) Foldable iPad could arrive as early as next year, claims noted Apple analyst (The Verge) Two companies race to deploy robotaxis in San Francisco. The city wants them to hit the brakes. (NBCNews) Scientists at Salesforce develop proteins with AI that can eat trash (NeoWin) AI has designed bacteria-killing proteins from scratch – and they work (New Scientist) 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 ride home for Monday, January 30th, 2020. I'm Brian McCalla today. Apprae chat GPT. Prepare yourself for the generative AI deluge. Amazon says the free lunch for grocery delivery is over. More speculation about foldable iPads. San Francisco wants Waymo and

Cruz to pump the brakes on their self-driving car experiments because the actual cars have literally been pumping the brakes too much. And my wager on whether or not 2023 will actually go down as the year that AI broke through. Here's what you missed today in the world of tech. So there are a bunch of parallel storylines running right now with this sudden generative AI mania. There's the whole VC investment angle. I could give you about a half a dozen behind the scenes rumors about big funding rounds that are coming all at the same time right now for various players in the space. And of course, there's the whole narrative of people feeling around in the dark right now trying to find out, as we discussed on the most recent bonus episode,

if you can actually build a viable business on top of this stuff, and if so, what that might look like. But another very interesting angle here is this. The big existing incumbent platform angle. Everyone in tech has an AI team. Everyone has been doing AI research for the better part of a decade now, sometimes as a loss leader for their cloud business experiments, sometimes as a, just in case this stuff becomes real insurance policy, Well, it's sort of become real, or at least real enough, all the sudden. So that sound that you are hearing is a ton of major tech platforms being like, crap, open AI has gone ahead and broken the seal on this thing. They've got a real product out there on the market. So we could have a product on the market, and we could have had it all this time, but we didn't think it was ready. So what do we do now? We've talked about this in regards to Google, but here's another case and point. According to Bloomberg, Chinese search giant Baidu plans to launch an AI chatbot similar to OpenAI's ChatGPT in March, initially embedded in its main search services. Quote, the tool whose name hasn't been decided will allow users to get conversation-style search results much like OpenAI's popular platform. The company's shares rose as much as 5.8% after

Bloomberg's report, the largest intraday gain in almost four weeks. Bidu has spent billions of dollars researching AI in years-long efforts to transition from online marketing to deeper technology. Its Ernie system, a large-scale machine learning model that's been trained on data over several years, will be the foundation of its upcoming chat GPT-like tool, the person said, a Baidu representative declined to comment. Baidu, Alibaba Group, Tencent Holdings, and Bight Dance Limited control much of China's internet. But the search company has been trying to revive growth in the mobile era after increasingly lagging its larger rivals in arenas such as mobile advertising, video and social media. Apart from research and AI, the search giant is now also developing

autonomous driving technology. Bidu chief executive officer Robin Lee raised Chat GPT as an example of where the tech giant can take the lead during an internal talk in December, according to a transcript viewed by Bloomberg News. I'm so glad that the technology we are pondering every day can attract so many people's attention. That's not easy, he said. He warned that the commercialization of generative AI by making it a, quote, product that everyone needs could be challenging, though. ChatGPT also piqued the interest of Chinese internet users who, like people elsewhere, shared screenshots of surprising conversations with the AI bot on local social media. That's despite a heavily censored domestic internet, largely walled off from the rest of the world,

a model that's helped companies like Baidu thrive as local equivalents to Google, Amazon, and Facebook. Apart from Baidu, several Chinese startups are also exploring generative AI and have attracted investors such as Sequoia and Synovation Ventures, end quote. Mark my words here. The next six months are going to be surprise announcement after surprise announcement where major platforms take the wrappings off their new generative AI products, whether they're ready for the prime time or not. I would hazard a guess that the Google and meta-developer conferences this year are going to be wild.

In another sign that the pandemic economy is over, no matter to what degree the pandemic itself is or is not over, Amazon will charge U.S. Prime members variable fees for Amazon fresh grocery delivery orders under $150 starting in February, dropping the practice where you could get free deliveries on any order over $35 or more, quoting Geekwire. In an internal email to customers, Amazon said it will start adding delivery charges of $3.95 for orders between $100 and $150, $695 for orders between $50 and $9.95 for orders for orders under $50. Prime members in select areas previously received free delivery for any order over $35. We're introducing a service fee on some Amazon Fresh delivery orders to help

keep prices low in our online and physical grocery stores as we better cover grocery delivery costs and continue to enable offering a consistent fast and high-quality delivery experience. An Amazon spokesperson said in an emailed statement, the new fees, which are on top of the $140 annual prime membership will go into effect starting February 28th. Amazon operates dozens of fresh grocery stores in the U.S. and got into grocery delivery in 2007. The tech giant acquired Whole Foods in 2017. As it dramatically expanded its push into the grocery space, in 2021, Amazon added

a $9.95 fee to Whole Foods delivery orders, which were previously free to prime members who spent more than $35. Amazon said in its email to customers that it will continue to offer two-hour delivery windows for all orders, and customers in some areas will be able to select a longer delivery window for a reduced fee, end quote. Yep, but this is another news item that calls into question Amazon's entire grocery strategy. Something tells me that within the next year or so, Amazon is going to have to make a tough decision to stick or twist when it comes to groceries. Either they need to double down to make it work economically, or they might abandon it as a business strategy altogether. And I've been sitting on several incremental stories about Apple quietly and sometimes not so quietly

moving more and more of its supply chain out of China. So I'm using this as a sort of placeholder to make note of the fact that this is getting very real. Sources say that key Apple supplier, Jabil has begun making AirPods enclosures or the plastic bodies of AirPods in India before shipping them to China and Vietnam for assembly. So note that China is still part of this whole mix, and you could say that plastic enclosures, who cares, but something, something about the difficulty of turning an aircraft carrier around. It takes time, and it's a series of incremental steps that can add up to a cumulative hole. Quoting Bloomberg, Apple is built building out production in India to reduce its reliance on China, where U.S. trade restrictions and

COVID-related disruptions have made manufacturing more risky. Its India output has thus far been limited to the iPhone, making AirPods the second Apple product now partially manufactured in the country. Indian Prime Minister Narendra Modi has made it a national priority to grow India's manufacturing sector, providing financial incentives and government support for companies' expansion projects. Apple has played a central role in that effort with partners such as Hon. High Precision Industry company, producing more iPhones in the country for the latest generation than ever before. U.S. manufacturing services provider Jabil operates a 858,000 square foot or 80,000 square meter facility, employing more than 2,500 workers in Pune, Western India, according to its website.

While India is still some time away from getting Apple to fully manufacture AirPods locally, New Delhi has given initial clearances to more than a dozen of its Chinese suppliers to ramp up via joint ventures with Indian partners. Luxhare Precision Industry Company, one of Apple's Chinese suppliers which makes AirPods, is among the companies gaining that approval. Luxshare in 2020 agreed to take over a shuttered Motorola plant in India's southern Tamil Nadu state, but is yet to begin manufacturing Apple products locally. Making enclosures is typically the first step for full-fledged production of AirPods,

said Neil Shah, Vice President of Research at Counterpoint. Now that Apple has won initial approval for some suppliers, including Luxembourg, share, they are obviously building a supply chain for the end product, end quote. Also, Apple, and from the momentarily quite full Apple rumor file, our friend Ming Chi Quo says that after a quiet 2023, he expects for the iPad lineup where the headlines will probably largely be about spec bumps and tweaks, Apple plans to release a foldable iPad next year in 2024, featuring a carbon fiber kickstand as well as an upgraded iPad mini. Quoting the Verge.

I'm positive about the foldable iPad in 2024 and expect this new model will boost shipments and improve the product mix. He tweeted early Monday. Quo expects it to be joined by a revamped iPad Mini due to enter mass production in early 2024. Quote in and offer many new details on the rumored iPad foldable but said that it will feature a carbon fiber kickstand produced by Chinese component manufacturer NG technology. A 2024 release date is significantly earlier than the last significant foldable iPad prediction. which came from display supply chain consultant analyst Ross Young last February. He reported that Apple is developing a foldable iPad slash MacBook hybrid with a roughly 20-inch folding screen, but anticipated that it wouldn't be ready for release until 2026. Bloomberg's Mark German later reported that Apple has been exploring a dual-screen foldable

and added that the bottom half of the display would serve as a virtual keyboard when the device is used as a MacBook-style clamshell. German didn't offer an exact release date prediction, but noted in October that Apple's foldable iPad could release, quote, later in the decade. This isn't the first time Quo has put a date on an Apple foldable prediction after he said a foldable iPhone might be released in 2023 two years ago. But it sounds like this might be a much smaller device with a screen size in the region of seven and a half to eight inches. German has even reported that Apple has discussed releasing a foldable device with a similar screen size to the 6.7-inch iPhone 12 Pro Mac. Given the lack of rumors about a foldable iPhone more recently, it's hard to imagine it releasing in 2023, as, quote, previously predicted, end quote. Before this past weekend, I hadn't been in San Francisco for about four years, since definitely

before the pandemic, so I hadn't had the pleasure of seeing self-driving cars on actual roads. They don't test them very often here in New York City, but damned if I didn't see dozens and dozens of them this weekend. When Chris and I were walking back from the comedy show on Market Street at about 11 p.m. on Saturday night, we saw one after another, after another, after another, which we thought was cool until we wondered if maybe the cars were doing that thing where they swarm or spawn to the same place for indeterminate reasons. And so it was interesting to me to see this. The city of San Francisco has asked California regulators to halt or scale back the expansion of Cruz and Waymo on San Francisco streets after repeated incidents with stopped and idle robotaxies.

Quoting NBC News. Cruise, which is majority owned by General Motors, won permission last year to use 30 vehicles as robotaxies in parts of San Francisco between 10 p.m. and 6 a.m. The vehicles do not have backup human drivers during that time. The company has since received permission to test driverless cars any time of day, but it needs a sign-off from the California Public Utilities. Commission to expand the hours of its commercial service. Neither vehicles from cruise or Waymo have killed anyone on the streets of San Francisco, but the companies need to overcome their sometimes comical errors, including one episode late last year in which a cruise car with nobody in it slowly tried to flee from a police officer. In one recent instance documented on social media and noted by city officials,

five disabled cruise vehicles in San Francisco's Mission District blocked a street so completely that a city bus with 45 riders couldn't get through and was delayed for at least 13 minutes. Cruises autonomous cars have also interfered with active firefighting, and firefighters once shattered a car's window to prevent it from driving over their fire hoses, the city said. San Francisco officials said they want to continue the experiment and even allow Cruz and Waymo to expand, but only if they do so slowly and with conditions. A series of limited deployments with incremental expansions rather than unlimited authorizations offers the best path forward towards public confidence in driving automation and industry success in

San Francisco and beyond. Three city officials wrote Thursday in a letter to the Utilities Commission, the state agency that decides if a company gets a robotaxy license. A second letter expressed concerns about Waymo. San Francisco doesn't want robotaxies operating in the city's downtown core, for example, or during morning and evening peak commuting times. And it wants more data on how the vehicles perform, end quote. And finally, one more. AI story, although I guess the previous segment was an AI story also. At the comedy show this weekend, I was asked by someone in the audience if I thought this really is AI's breakthrough moment or not. And I gave it 50-50 odds, at least at this moment. 50% odds that 10 years from now we would

look back at 2023 as the year that AI broke through. Also, though, 50% odds that this will be another false dawn that has a ton of hype, and AI will go back to being what it's been for a while now, something perpetually 10 years away. See, self-driving car technology. What could affect my assessment and change the odds would be if more tangible things start to break through beyond, you know, arguably the sort of gimmicky stuff we've already seen. For years, one of the promises of AI has been what it could do for science, for medicine. You know, AI could solve global warming, could cure cancer. So if more things like this pan out, I could adjust the odds. Researchers have detailed how Progen, a large language model trained on

280 million proteins, designed new proteins with antimicrobial properties that were tested in real life and shown to work. Quoting Neowen, what do you get when the world's largest CRM breaks into the research industry and leverages AI to build their products? You get Progen, a new AI system that can make artificial enzymes from scratch that can work just as well as real ones found in nature. ProGen was made by Salesforce research. Yes, that Salesforce and uses language processing to learn about biology. In short, Progen takes amino acid sequences and turns them into proteins. The artificial designs are better than ones made by the normal process, says James Frazier, a scientist involved in the project. We can now make specific types of enzymes, like ones that

work well in hot temperatures or acid, end quote. To make ProGen, the scientists at Salesforce fed the system amino acid sequences from 280 million different proteins. The AI system quickly made a staggering 1 million protein sequences, of which 100 were picked to test, end quote. And now quoting from new scientists. The AI called ProGen works in a similar way to AIs that can generate text. ProGen learned how to generate new proteins by learning the grammar of how amino acids combined to form 280 million existing proteins. Instead of the researchers choosing a topic for the AI to write about, they could specify a group of similar proteins for it to focus on. In this case, they chose a group of proteins with antimicrobial activity. The researchers program checks into the

AI's process so it wouldn't produce amino acid gibberish, but they also tested a sample of the AI proposed molecules in real cells. Of the 100 molecules they physically created, 66 participated in chemical reactions similar to those of natural proteins that destroy bacteria in egg whites and saliva. This suggested that these new proteins could also kill bacteria. The researchers selected the five proteins with the most intense reactions and added them to a sample of E. coli bacteria. Two of the proteins destroyed the bacteria. The researchers then imaged them with x-rays, and even though their amino acid sequences were up to 30% different from any existing proteins, their shapes almost matched naturally occurring proteins.

James Fraser at the University of California, San Francisco, who was part of the team, says it was not clear from the outset that the AI could work out how to change the amino acid sequence so much and still produce the correct shape. He was surprised to have found a well-functioning protein in the first relatively small fraction of all the progen-generated proteins that they tested, end quote. Shout out to Salesforce for making our news roundup again for the third time in as many weeks. But what did I say about every major player in tech having had an AI research arm of one form or another for years now. Thank you to everyone who came out to the meetup on Saturday.

In the end, I think we were approaching 50 people. We kind of packed the bar. I tried to post pictures when people sent them to me. So send me more if you haven't, and I'll tweet more pictures. I tried my best to talk to everyone that I could, but I know I probably missed people or didn't get to talk to some folks as much as I'd like. For example,

there was one woman who I only briefly got to talk to about using AI for voice stuff. Hopefully you know who you are. Chris was telling me about you afterwards. Very interesting stuff.

Anyway, everyone was great and kind, and I hope you all had fun. And the comedy show was a lot of fun for me. Whenever Mosha posted, I'll let you know so you can listen if you want. Moshe Casher was a gentleman

and a scholar for setting the whole thing up. I hope I delivered the deeper and more philosophical angle that he was going for. The other comedians on the panel were very gracious to me, helping me hopefully not embarrass myself. Shout out, especially to Baron Vaughn. And also to Janine Grofalo, who talked to me later because she was worried that she had been too antagonistic

by expressing her tech ludite tendencies, but it was fine and great and actually very funny, and it sure helped to know that, like, one-third of the audience were members of the Mutant Podcast Army, so I knew I had support from the crowd out there, And thanks to Moshe again for introducing me, however briefly, to some of my comedy idols. I met Seth Morris, y'all. I met Padgett Brewster and Andy Richter and half the cast of the movie A Mighty Wind.

I didn't talk to him directly, but I did have a drink at the bar next to George Went. And when you have a drink next to Norm himself from Cheers, that's pretty cool. Anyway, I know I'm forgetting a ton of details to share with you from the weekend, but right now I'm in the state of jet lag where I'm completely disassociating. Three and a half hours of sleep for two days in a row is a hell of a drug. Talk to you tomorrow.