← Tech Brew Ride Home
Tech Brew Ride Home
Nvidia Back In China… Maybe
00:20:20
Trump says Nvidia and others can ship chips to China, but the question is, will China take delivery. OpenAI is ending the code red in about a month, after getting a new model out the door. Meta wants a new Llama model, maybe in a month. And a new smart ring that is pretty intriguing…
Loading summary
Transcript10 lines
- [00:01]
A
The Uniswap Wallet makes it easier and safer to own and use crypto Created by pioneers of the crypto economy, the Uniswap protocol has powered over $3 trillion in trading volume, and it's trusted by tens of millions worldwide. With the Uniswap Wallet, you can discover, swap and manage your crypto all from your phone. Buy your first crypto assets in just a few taps and start exploring the freedom of decentralized finance with Uniswap. Tap the Banner to get started.
- [00:35]
B
Welcome to the Tech Brew Ride home for Tuesday, December 9, 2025. I'm Brian McCullough. Today Trump says Nvidia and others can ship chips to China, but the question is, will China take delivery? OpenAI is ending the Code Red in about a month after getting a new model out the door. Meta wants a new Llama model, maybe in a month as well, and a new smart ring that is pretty intriguing. Here's what you missed today in the world of tech.
- [01:05]
You may have noticed that your customers love webinar and video content, but if you've ever put together a webinar or video, then you know that it can eat up a lot of your time and budget. Now, however, thankfully, there's a single tool that can streamline your team's video and webinar workflows. Wistia. Wistia can scale your content output with AI powered tools that help you create, edit and repurpose videos and webinars fast. And speaking of webinars, you can host engagement, engaging, easy to set up webinars in Wistia too, complete with built in analytics. With Wistia you don't have to pay for multiple video tools, hop between platforms, or constantly re upload files, create, edit, collaborate and publish all in one place. Head to wistia.com brew to learn more. That's W-I-S-T-I-A.com brew. With Wistia you can expect less work and more plays.
- [01:57]
President Trump said the US will let Nvidia ship its H200 chips to approved customers in China and elsewhere on the condition that the U.S. gets a 25% cut. The president also said the Commerce Department is working on the same approach for amd, intel and other great American companies as Nvidia, which okay, good day for Jensen at All right. Except sources are also telling the FT that Beijing is set to limit access to Nvidia's H200 chips in its push for chip self sufficiency. Quoting CNBC, Chinese President Xi Jinping responded positively to the proposal, Trump wrote in a Truth Social post. The policy, quote, will support American jobs, strengthen U.S. manufacturing and benefit American taxpayers, trump wrote. The Department of Commerce is finalizing the details and the same approach will apply to amd, intel and other great American companies, he added in the post. Both Nvidia and chip rival amd, short for Advanced Micro Devices, agreed in August to share 15 of the revenue from China chip sales with the US government. But around that same time, China reportedly warned companies against using the H20AI chip that Nvidia designed especially for the country. The H200 is a higher grade chip than the H20, but not the company's top of the line product. Nvidia shares climbed earlier Monday on news that the Commerce Department was set to approve the China sales, but later paired those gains. The Stock rose about 2% after hours. We applaud President Trump's decision to allow America's chip industry to compete to support high paying jobs and manufacturing in America, a spokesman for Nvidia told CNBC in a statement. Offering H200 to approved commercial customers vetted by the Department of Commerce strikes a thoughtful balance that is great for America, the spokesman said. End quote. And quoting the FT Beijing is set to limit access to Nvidia's advanced H200 chips despite Donald Trump's decision to allow the export of the technology to China as it pushes to achieve self sufficiency in semiconductor production, according to two people with knowledge of the matter. Regulators in Beijing have been discussing ways to permit limited access to the H200, Nvidia's second best generation of artificial intelligence chips. Buyers would probably be required to go through an approval process, the people said, submitting requests to purchase the chips and explaining why domestic providers were unable to meet their needs. No final decision has been made yet, the people added. China has used various bans to push domestic chip makers to develop products to compete with Nvidia. Moves include stepping up customers, customs checks of chip imports and offering energy subsidies to data centers using domestic chips. The two regulators in charge of Beijing's years long semiconductor independence campaign, the National Development and Reform Commission and the Ministry of Industry and Information Technology could apply other measures to ensure the competitiveness of domestic chips, the people said, including banning China's public sector from buying the H200. The return of Nvidia's advanced chips would be welcomed by tech giants such as Alibaba, ByteDance and Tencent, which have been using more Chinese chips for some basic AI functions but still prefer Nvidia's products because of their higher performance and easier maintenance. Many of them are training their AI models abroad to access Nvidia chips banned at home. While Trump has signaled his approval for the export of Nvidia's advanced chips to China, he faces opposition in Congress. A group of US Senators has introduced legislation that would prevent the administration from approving exports of chips, including the H200, to China for 30 months. Washington might also adopt an approval process that allows sales of H200 chips only to companies it considers safe, said the people familiar with the matter. Nvidia has already been approved to export the H20, a watered down version of the H200 made specifically for China, after the company in August agreed to pay the US government 15% of its revenues from chip sales in China. Beijing, however, has restricted tech companies access to the H20, arguing that chip's performance is not significantly better than Chinese alternatives. End quote.
- [06:04]
Sources say that Sam Altman plans to end OpenAI's so called code Red after releasing a model next month in January with improved image generation, speed and personality. Quoting the journal with the code read, Altman instructed employees to boost ChatGPT in a specific way through better use of user signals, he wrote in his memo. With that directive, Altman was calling for turning up the crank on a controversial source of training data, including signals based on one click feedback from users rather than evaluations from professionals of the chatbot's responses. An internal shift to rely on that user feedback had helped make ChatGPT's 4.0 model so sycophantic earlier this year that it has been accused of exacerbating severe mental health issues for some users. Now, Altman thinks the company has mitigated the worst aspects of that approach, but is poised to capture the upside. It significantly boosted engagement as measured by performance on internal dashboards tracking daily active users. It was not a small, statistically significant bump, but like a wow bump, said one person who worked on the model. Behind Altman's Code Red declaration, however, are tensions between camps inside the company that have been festering for years, according to people familiar with the matter. A group including Fiji Simo, a former Meta Platforms executive who leads OpenAI's product efforts, and Chief Financial Officer Sarah Fryer have pushed the company to pour more resources into ChatGPT. Simo has also told staff that OpenAI needs to do a better job of making sure its users discover the value of ChatGPT's existing features before the company goes on to build new ones, and also wants to improve the chatbots speed and reliability. Researchers, meanwhile, have prioritized state of the art technology that could lead to artificial general intelligence, or AGI, but don't do as much to improve the basic chatbot experience. OpenAI is set to release a new model called 5.2 this week that executives hope will give it new momentum, particularly among coding and business customers. They overruled some employees who asked to push back the model's release so the company could have more time to make it better, according to people familiar with the matter. The company also plans to release another model in January with better images, improve speed and a better personality, and to end the Code Red after that, Altman said. An OpenAI spokeswoman says there's no conflict between the two philosophies and that broad adoption of AI tools is how the company plans to distribute AGI's benefit. End Quote.
- [08:34]
We all remember the choices that shaped the course of our lives in business. World renowned venture capital firm Sequoia Capital calls them Crucible Moments. Their podcast brings you inside the pivotal decisions that defined some of today's most influential companies. Hosted by Sequoia's Rulof Botha, Crucible Moments Season 3 pulls back the curtain on the untold stories behind companies like Stripe, Zipline, Palo Alto Networks, Klarna, supercell, and more. I loved the recent episode with the founder of Zipline and how even late to the game, they are leapfrogging the bigger players to bring true autonomous drone delivery not just to hospitals, but now to customers of the likes of Walmart and Chipotle. I might win that burrito delivery bet someday soon. Tune in to Sequoia's new season of Crucible Moments to discover how some of the most transformational companies of the modern era were built. Crucible Moments is available everywhere you get your podcasts and@CrucibleMoments.com.
- [09:28]
Keeping pace with data growth in the age of AI is like trying to find enough shelf space after a trip to a big box store. AI and data growth have outpaced the old storage model. Manual management of traditional storage can't keep up, so it's time for a new, unified approach from pure Storage. They help organizations simplify and automate how data is stored and managed, eliminating silos and putting intelligence at the center of operations. When you don't know where data lives or how it's used, governance slips, visibility and compliance can become constant challenges. The pure storage platform unifies storage into a single, intelligent layer that can turn data into a governed, virtualized cloud of data with guaranteed outcomes. Learn more@PureStorage.com Morning Brew that's PureStorage.com MorningBrew.
- [10:21]
I know it's hard to keep track of the various players in the AI horse race, but that's kind of what I get paid to do. So here's me telling you, don't forget about Meta. Remember, Zuck spent an absolute metric ton of money to hire people to reorganize Meta's AI efforts and at at some point they are going to come back out swinging, probably in a big way. Sources have told CNBC that Meta is pursuing a new Llama successor and Frontier AI model codenamed Avocado, set for release in Q1 of 2026, and it could be a proprietary model. As 2025 comes to a close, Meta's strategy remains scattershot, according to insiders and industry experts, feeding the perception that the company has fallen further behind its top AI rivals, whose models are rapidly gaining adoption in the consumer and enterprise. Meta is pursuing a new Llama successor and Frontier AI model, codenamed Avocado, CNBC has learned. People with knowledge of the matter said many within the company were expecting the model to be released before the end of this year, but that the plan now is for that to happen in the first quarter of 2026. The model is wrestling with various training related performance testing intended to ensure the system is well received when it eventually debuts, said the people, who asked not to be named because they weren't authorized to speak on the matter. Our model training efforts are going according to plan and have had no meaningful timing changes, am Meta spokesperson said in a statement. With its stock underperforming the broader tech sector this year and badly trailing Google parent Alphabet, Wall street is looking for a sense of direction and a path to a return on investment after Meta spent $14.3 billion in June to hire Scale AI founder Alexander Wang and a handful of his top engineers and researchers. Four months after that announcement, which included Meta purchasing a big stake in Scale, the social media company raised its 2025 guidance for capital expenditures to between 70 and $72 billion from a prior range of 60 to $72 billion. In many ways, Meta has been the opposite of Alphabet, where it entered the year as an AI winner and now faces more questions around investment levels and ROI, analysts at KeyBanc Capital Markets wrote in a November note to clients. The firm recommends buying both stocks. Avocado, when it's eventually made available, could be a proprietary model, according to people familiar with the matter. That means outside developers wouldn't be able to freely download its so called weights and related software components. Some at Meta were upset that the R1 model released by Chinese AI lab Deep Seq earlier this year incorporated pieces of Llama's architecture, the people said, further underscoring the risks of open source and hammering home the idea that the company should overhaul its strategy. The company's high priced AI hires and leaders of the recently launched Meta Superintelligence Lab, or msl, have questioned the open source strategy and favored creating a more powerful proprietary AI model, CNBC reported in July. Wang is now under pressure to deliver a top tier AI model that helps the company regain momentum against rivals like OpenAI, Anthropic and Google, the people said. That pressure has only increased as competitors stepped up their game. Google's Gemini 3, unveiled last month, has drawn solid reviews from users and analysts. OpenAI recently announced new updates to its GPT5AI model, while Anthropic debuted its Claude Opus 4.5 model in November shortly after releasing two other major models. Wang isn't the only Meta exec feeling the heat. Nat Friedman has also been tasked with producing a breakout AI product, the people said. He was responsible for Meta's September launch of Vibes, a feed of AI generated short videos, which is widely viewed as inferior to OpenAI's Sora 2, they said. Former employees and creators told CNBC that the product was rushed to market and lacked key features like the ability to generate realistic lip synced audio. Friedman has called for MSL to use newer tools that have been calibrated to incorporate multiple AI models and various kinds of coding automation software often called AI Agent. The people said they have this mantra now, saying demo, don't memo, lovable CEO Anton Osica said in October at the Masters of Scale Summit in San Francisco about Meta's new development process. End quote.
- [14:30]
Finally today, some very interesting new hardware, or at least interesting to me, pebble has unveiled the Pebble Index 01, a $99 smart ring with an on device LLM for processing voice notes, shipping in March of 2026 initially for just $75. Quoting Wired Unlike every other pebble product, this isn't a watch but a smart ring. And unlike most other smart rings, the Pebble Index doesn't measure your heart rate or track your sleep. Its purpose revolves around memory. It's a tool to remember things. There's a physical button on the ring. Wear it on your index finger, then use your thumb to press and hold the button as you speak to it. The index can remember an important tidbit, save a reminder, start a timer or create a calendar event. Pre orders start today with the index, priced at $75, and it's expected to ship in March. Another company very recently announced a similar gadget, Whisper, into Sandbar's Stream Ring to have it record your stream of consciousness musings. But where Sandbar's product leans heavily on AI processing and requires a subscription, the Pebble Index veers in the opposite direction. I'm not as interest interested in the AI Persona, pebble founder Eric Migikovsky tells Wired. The hardware is simple because Mijikovsky wanted to ensure it works 100% of the time. There are no haptics or vibration motors. It's just a physical button and a microphone, one that Michikovsky assures will be able to pick up your voice even in loud environments. The stainless steel ring comes in three colors matte black, polished silver and polished gold, and eight sizes plus it's water resistant. It has some internal memory to save your audio clips even offline, but when your phone is in range, it'll sync the files to the pebble app. The index works with both Android and iOS. There's no way to recharge the ring. Mijikovsky said he didn't want yet another gadget to charge every day, so instead the Pebble Index has non rechargeable silver oxide hearing aid batteries designed to last two years with average use. Once the device's battery is nearly dead, users will receive a notification in the app and the idea is you'll buy a new Pebble Index, an idea that's easier to get behind knowing the ring costs just $75, though the price will jump to $99 after the first batch. You'll also be able to send your old index to the company for recycling. When your audio is sent to your phone, an open source speech to text AI model processes it locally to convert your voice notes to text. Then an on device large language model will categorize the audio, deciding whether it's a reminder, a timer or a general note. A feed shows all your memory logs and you scroll through it to find and listen to each clip. None of this data is ever sent to the cloud. It all stays on your phone. These are your innermost thoughts, migicovsky says. You don't want to send them anywhere. By default, all your musings with the Ring are handled by the pebble app, so if you had it set a reminder, you'll get one from the pebble app. However, you can customize the destination if you prefer to use your own service. If you use the Notion app for notes and tasks for example, you can set it up so that your reminders and thoughts will be sent there. The open source nature of the pebble app means there's no limit to customize, you press and hold the button to log a note, but you can have a single press trigger an action. Migicovsky says he set his to play or pause music, and a double press switches tracks. But you can set it to take a photo remotely or activate a smart home routine. There will be an Actions category in the Pebble App Store where folks can publish their custom actions with a double press and hold. You can go a step further with a secondary suite of actions. These actions are sent to a chatgpt like bot to help you answer basic everyday questions like what's the weather? Or something you type into Google Search. These more open ended results naturally will require an Internet connection, Mijicovsky says. They won't work all the time and this capability is unavailable by default. People will be able to write and publish their own model context protocol actions, expanding the use of the ring. As for the output, if you have earbuds, you'll be able to hear the results of your queries. If you have a Pebble smartwatch, you may be able to see those weather results populate on the watch face. End quote.
- [18:51]
Nothing more for you today. Talk to you tomorrow.