Tech Brew Ride Home - Mon. 11/06 – Elon “Groks” AI

Episode Date: November 6, 2023

Elon releases his AI model. The Chinese AI startup that has hit unicorn status in just 8 months by going the open source route. The reviews of the new Mac with M3 chips are out, and they’re generall...y good. But if you’ve been pining for a new 27 inch iMac, I’ve got some very bad news for you. Sponsors: Traceroute Podcast Kolide.com/ride ipn.ibotta.com/ride Links: Elon Musk debuts ‘Grok’ AI bot to rival ChatGPT, others (CNBC) AI Pioneer Kai-Fu Lee Builds $1 Billion Startup in Eight Months (Bloomberg) Bumble CEO Whitney Wolfe Herd Steps Down (WSJ) Review: Apple’s 16-inch M3 Max MacBook Pro crams Ultra-level speed into a laptop (ArsTechnica) Apple has no plans to make a 27-inch iMac with Apple Silicon (The Verge) Learn more about your ad choices. Visit megaphone.fm/adchoices

Transcript
Discussion (0)
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 TechMame right home for Monday, November 6, 2023. I'm Brian McCullough today. Elon releases his AI model, the Chinese AI startup that has hit unicorn status in just eight months by going the open source route. The reviews of the new Macs with the M3 chips are out and they're
Starting point is 00:00:50 generally good, but if you've been pining for a new 27-inch iMac, I've got some very bad news for you. Here's what you missed today in the world of tech. You should be aware of the fact that today Open AI is holding its Developer Day, and that has become one of the biggest deals out there right now for obvious reasons. I debated holding the show today till later in order to report on the keynote, but while I will watch the keynote live myself, there was enough news today that I figured we'll wait and run down everything announced today tomorrow so that we don't miss out on important things to tell you about right now. Now, maybe because of the Developer Day happening, there is quite a bit of AI news today. Maybe they're trying to get ahead of OpenAI's news. For example, Elon Musk's OpenAI competitor,
Starting point is 00:01:41 XAI, has released GROC. A new AI model that the startup says surpasses rivals in its compute class, including GPT3.5 to a limited number of U.S. users. Quoting CNBC, GROC, the company said, is modeled on the hitchhiker's Guide to the Galaxy. It is supposed to have a bit of wit, a rebellious streak, and it should answer the... the spicy questions that other AI might dodge, according to a Saturday statement from XAI. Leading up to the release, Musk posted on X, formerly Twitter, an example of Grock responding to a request for a step-by-step cocaine recipe. Oh, sure, Grock responded, just a moment while I pull up the recipe for homemade cocaine. You know, because I'm totally going to help you with that,
Starting point is 00:02:27 end quote. In case you didn't get that, that was supposed to be sarcastic. Grock also has access to data from X, which XAI said will give it a leg up. Musk on Saturday, posted a side-by-side comparison of GROC answering a question versus other AI bots, which he said had less current information. Still, XAI hedged in its statement, as with any large language model or LLM, GROC, quote, can still generate false or contradictory information, end quote. The prototype is in its early beta phase, only two months in training, and is available to a select number of users to test out before the company releases it more widely. Users can sign up for a waitlist for a chance to use the bot.
Starting point is 00:03:08 Eventually, Musk said on X, GROC will be a feature of X Premium Plus, which costs $16 per month. The Tesla and SpaceX CEO appears to be positioning XAI as a challenger to companies like OpenAI, inflection, an anthropic. On an initial round of tests based on middle school math problems and Python coding tasks, the company said that GROC surpassed all other models in its compute class, including chat GPT3.5, and inflection 1. It was outperformed by bots with larger data troves. Quote, in some important respects, it is the best that currently exists, Musk said in an ex post on Friday, leading up to the GROC announcement. GROC is a term coined by Robert Heinlein in his
Starting point is 00:03:51 1961 science fiction novel Stranger in a Strangeland. In the book, GROC is a Martian term with no direct earthling translation. Critics have debated the word's exact definition, but have settled on some version of having very deep empathy or intuition with something. Merriam Webster defines it simply as a transitive verb that means to understand profoundly and intuitively. XAI launched in July with a team stacked with former employees of OpenAI, DeepMind, and more. It is still hiring for several roles, end quote.
Starting point is 00:04:27 And let me introduce you to 01 AI, which launched just this past March. It was founded by computer scientist Kai Fu Lee. But it has already reached a more than $1 billion valuation on the release of its AI model Ye 34B in Chinese and English, quoting Bloomberg. A Chinese startup founded by computer scientist Kai Fu Lee has become a unicorn in less than eight months on the strength of a new open-source artificial intelligence model that outstrips Silicon Valley's best, at least on certain metrics. The company, 1.a.I, or 0.1.A.I, has reached a valuation of more than a billion dollars after a funding round that included Alibaba Group's cloud unit, Lee said in an interview.
Starting point is 00:05:12 The chief executive officer of venture firm, Synovation Ventures, will also be the CEO of the new startup. He began assembling the team for 0.01.AI in March and started operations in June. The Beijing startup's open-sourced foundational large language model, Yi 34B, is now available to developers around the world in Chinese and English. Large language models or LLMs are computer algorithms trained on large quantities of data to read, understand, and produce human-like text, images, and code. On key metrics, Yi-34B outperforms leading open-source models already on the market, including Meta's well-regarded Lama 2.
Starting point is 00:05:48 Hugging Face, which runs leaderboards for the best-performing LLMs in various categories, posted evaluations over the weekend, ranking the Chinese model first for what's known as pre-trained base LLMs. Lama 2 has been the gold standard and a big contribution to the open source community, Lee 61 said in an interview over Zoom, We want to provide a superior alternative not just for China, but for the global market, end quote. Referring to the geopolitical tensions between the U.S. and China, including export controls of high-end AI chips, Lee called the situation, quote, regrettable, but said, Zero1.a.I stockpiled the chips it needs for the foreseeable future.
Starting point is 00:06:22 The startup began amassing the semiconductors earlier this year, going as far as borrowing money from Synovation Ventures for the purchases. We basically bet the farm and overspent our original bank account, he said. We felt we had to do this. Lee, who worked at Google, Microsoft, and Apple before moving into venture capital, has built a team of more than 100 people at 0.01.AI, drawing former colleagues from the U.S. companies and Chinese nationals who have been working overseas. The group includes not just AI specialists, he said, but experience business people who can help with everything from mergers and acquisitions to an initial public offering.
Starting point is 00:06:57 Zero1.a.I is already plotting its business strategy beyond the open source model just introduced. The startup will work with customers on proprietary alternatives tailored for a particular industry or competitive situation. For example, Yi-34B gets its name from the 34 billion parameters used in training, but the startup is already working on a 100 billion-plus parameter model. Our proprietary model will be benchmarked with GPT4, said Lee, referring to OpenAI's LLM. Offering a system in English and Chinese will be an advantage for global companies
Starting point is 00:07:26 like banks, insurers, and trading companies, he said. The startup plans to add more languages in the future. Lee's work on AI dates back decades. In his 1982 application to graduate school at Carnegie Mellon University, he wrote that he wanted to devote his life to AI research because the technology would help humans understand themselves better. He went on to write two best-selling books, AI superpowers, China, Silicon Valley, and the New World Order, and AI 2041, 10 visions for the future. It's the biggest breakthrough for humanity, Lee said. It's also the final step to understanding ourselves, end quote. He said he
Starting point is 00:08:00 frequently wondered whether the technology's promise would be realized during his lifetime until he saw the power of the latest generation of LLMs. He decided to depart from his usual role as a venture capitalists and become a startup CEO because of the promise of the technology and the opportunity to connect the dots from his roles as computer scientist, author, and executive, end quote. Bumble founder Whitney Wolf Hurd plans to step down as CEO of Bumble nearly a decade after founding the dating app to be replaced by Slack CEO. Lidian Jones on January 2, 2024. Notable, as heard, has been maybe the most prominent female founder still at the company she founded, quoting the Wall Street Journal. The leadership change comes
Starting point is 00:08:48 at a challenging time for the dating app industry. Last week, Match Group, the dating app behemoth that owns platforms like Match.com, OKCupid, Hinge and Tinder, which Wolf Hurd, co-founded before Bumble, saw its stock fall after it reported its earnings for the third quarter. The company He projected lower than expected revenue for the final quarter of the year and reported losing paying customers at Tinder the crown jewel in its portfolio. Bumble, which went public in February 2021, saw its stock trading at more than $70 per share on opening day. It has since fallen below $14. Bumble reports its third quarter earnings on Tuesday. I want to be the person who is able to look around the corner and innovate for the future of Bumble, Inc.
Starting point is 00:09:26 and to take us 10 years ahead, Wolf Hurd, 34 said in an interview in 2014 at 24 years old, Wolf Hurd launched the dating app Bumble on the idea that having women initiate conversations with dating prospects would help weed out unwanted messages and put them in control of their romantic lives, end quote. The reviews for the Macs with those new M3 chips inside them are out, and I'm not really going to get into the IMAQ reviews, though I do have news on the IMAC, as I said in a second. Basically, the consensus seems to be that the IMAX with the M3s offer impressive gains over the M1. It's good for gaming.
Starting point is 00:10:09 has an excellent webcam, but the relatively few ports and poor peripheral options, like the keyboard without backlighting, are problematic. And actually, let's align over the MacBook Pro entry level with M3 reviews as well, though the consensus there seems to be great performance, long battery life, a worthy entry-level pro model, but 8 gigabytes of RAM is silly, and they actually took away a Thunderbolt port, which is not useful at all. Let's focus on the high end, though, and I'm going to turn to Ars Technica to do that this time. They said that the M3 MacBook Pro surpasses the performance of the M1 Ultra. It's got the usual nice design and good Pro-motion display, but the base 8 gigabytes of RAM in the 1599 config and memory is pricey.
Starting point is 00:11:02 Quote, like the early 2023 MacBook Pros, these late 2023 models are iterative improvements to the 2021 redesigns. They keep the things that made those laptops such a big improvement over the late model Intel MacBook Pros, while adding just a little more performance and one or two other minor improvements to entice people who still haven't made the Apple Silicon Switch. We can only paint a partial picture of these new notebooks performance since we were only able to get a fully loaded M3 Max version of the 16-inch MacBook Pro for testing. But the short version is that two years of updates and the brand-new, more efficient manufacturing process that the M3 uses should make these an appealing upgrade to anyone who couldn't quite justify
Starting point is 00:11:42 paying for an upgrade before now. Just get ready to shell out because top-end MacBook Pro configurations are more expensive than ever, thanks to the new 128-gibight RAM option. Switching to its own chips continues to pay dividends for Apple. In the three years since Apple moved away from Intel, Intel has put out three successive chip generations with the exact same integrated GPU in them, processors that have improved CPU performance by a lot. but often at the cost of battery life. And there's no Windows notebook I've tested, whether we're talking about a thin and light ultra book
Starting point is 00:12:13 or a ridiculous gaming laptop that doesn't sound like a jet engine when it's running a game or some heavy CPU intensive task for more than a few seconds. The end of Apple's Intel era is coming. It could be here as early as next year, but all signs point to the last Intel Macs being dropped by 2025's MacOS release. All Apple Silicon Macs have been some kind of an upgrade
Starting point is 00:12:34 compared to the Intel models they've replaced, But the M3 generation will be perfectly poised to catch a lot of people who own those last two or three Intel Mac generations, ones made between 2018 and 2020. For them, it will be an immense upgrade, everything that was good about the M1 and M2 releases, but with a bit of extra speed and a handful of minor hardware refinements. The worst thing I can say about the new MacBook Pro is that it doesn't really fix any of the problems that might keep people away from a pro-level Mac in the first place. entry-level models like the 1599 MacBook Pro still leave something to be desired for a lot of pro-level
Starting point is 00:13:08 users, like it's paltry 8 gigabytes of RAM or lacking external display support. And upgrade prices for additional RAM and storage remain absurd, easily inflating Apple's base prices by hundreds or thousands of dollars. Storage prices in particular are positively usurious compared to what high-end PC storage costs, upgrading from 1 terabyte to 8 terabytes of storage costs $2,200. This is more than a lot of... laptops even cost. It's also worth saying that I don't think people who already own any kind of M1 or M2 series Apple Silicon Mac need to consider upgrading to any of these M3 models yet.
Starting point is 00:13:45 Performance is better, but it was already good. Upgrade if you're also stepping up in the lineup from an M1 Air to an M3 something pro, but most people can afford to wait for another even better chip generation or a more substantial hardware redesign before they replace a Mac that's still only a couple of years old, end quote. Finally, for the bad news on the iMac, as I mentioned. Apparently, Apple does not have plans to release an Apple Silicon version of the 27-inch IMac. Instead, they seem to be encouraging buyers to consider a Mac Studio or Mac Mini and a studio display, quoting the Verge. Apple PR representative Starlane Mesa confirmed the company's plans to the verge. The company encourages those who have been holding out hope for a larger iMac to consider the studio display and Mac Studio or Mac Mini,
Starting point is 00:14:39 which pair a 27-inch 5K screen with a separate computer compared to the all-on-one design of the iMac. For many years, the 27-inch iMac was Apple's flagship consumer desktop computer, and its reach spread far beyond the typical living room or home office into the studios and edit bays of many media companies. But since Apple never released a version with its in-house Apple Silicon processors, many video editors, developers, and other creatives have moved on to other machines such as the Mac Studio or MacBook Pro or more modular desktop PCs. Apple says the 24-inch IMac, which has a 4.5K display, neatly splits the difference between the old Intel models, which came in 21-inch 4K versions and the 27-inch 5K model. But the 24-inch IMac does not come with Apple's most capable processors, instead sticking with its base model chips.
Starting point is 00:15:26 Even today's base Apple Silicon processors are faster than the old. or Intel models, but in years past, you could equip an Intel IMac with very high-end chips and discrete GPUs that made it a powerful machine at the time. Those who want more performance than the base M3 chip now, particularly for GPU-focused workloads, have no choice but to look at the other Macs or outside of Apple's lineup. The larger IMAQ also had more power user features like a wide array of ports and an SD card slot. The current 24-inch IMA comes with a maximum of four USBC ports, and that's it. I've been testing the new IMAQ for a few days, and we'll be publishing my review later this week.
Starting point is 00:16:04 It's an excellent all-on-one desktop computer for most consumers, but as someone who uses a 27-inch monitor all day for work, I immediately feel more claustrophobic on the 24-inch screen and wish there were an option for a bigger model. But it's clear that Apple does not intend for the IMA to be anything more than a family computer or to look pretty at the receptionist desk of trendy retail stores and spas. The move away from power users also puts some clarity around the reality of the all-on-one desktop PC market, has been relegated to a niche status compared to the dominance of laptops or more personal devices like phones and tablets. The company's message today is specifically concerning a 27-inch iMac, which does leave the door open for an even bigger iMac to come in the future, as has been rumored on and off for some time. I personally would not hold out much hope for that, however, end quote. So some of you might have heard an ad at the start of the show today
Starting point is 00:17:03 and at the start of the show on Friday. I just want you to know, I'm not going to make a a habit of that. What happened was I screwed up an ad and had to do a make good for the advertiser and doing it as a pre-roll was the only way to fix the problem without having to do three ads one day in the middle. But then I screwed up the ad insertion on Friday, so I had to do a make-good to the make-good today. Just want you to know, I'm not getting greedy. I don't plan to do pre-roll ads going forward or in the future. I just, I had to fix something. Sorry. Because, as I always say, I'm a fan of podcasts myself. I'm not only a podcaster.
Starting point is 00:17:41 I'm also a client, if you're old enough to remember that reference. Nope, I'm a fan boy too, which is why on the most recent episode of the Flop House podcast at about the one hour, 33 minute mark, I almost shouted out loud when they read my letter.
Starting point is 00:17:57 Yes, the letter from Mr. McCullough, first name withheld, who wrote into the Flop House, that was me in case any of you heard that. So much fun. Thanks, boys. Talk to you tomorrow.

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