← Tech Brew Ride Home
Tech Brew Ride Home
Pour Moi, C'est Le Déluge
00:21:11
Aaannddd…. Right on time here come the Chinese AI models. Elon Musk kicks off a major reorg of xAI. Google is warning of AI distillation attacks. New Waymo cars hit the road. And another interesting AI essay to read to you.
Loading summary
Transcript1 lines
- [00:04]
A
Welcome to the Tech Brew Ride home for Thursday, February 12, 2026. I'm Brad McCullough today and right on time, here come the Chinese AI models. Elon Musk kicks off a major reorg of Xai. Google is warning of AI distillation attacks, new Waymo cars hit the road and another interesting AI essay to read to you. Here's what you missed today in the world of tech. Well, the deluge has begun. The deluge of Chinese AI models. I mean Zai, also known as Zhipu, has launched GLM5, saying its flagship open weight model has quote best in class performance among all open source models in reasoning, coding and agentic tasks. Note again open weights as Chinese companies are want to do, but also they're going after the hotness right now, which is of course agents. Quoting Reuters the open source GLM5 model features enhanced coding capabilities and the ability to perform long running agent tasks approaching rival Anthropic's Claude Opus 4.5, encoding benchmark tests and surpassing Google's Gemini 3 Pro on some benchmarks, the company said in a press release. The latest model was developed using domestically manufactured chips for inference, including Huawei's flagship Ascend chip and products from leading industry players such as More Threads, Cambercon and Kunluxin, according to the statement. Zhipu is considered one of China's AI Tigers, a group of promising AI startups in the country fing with the United States to lead the development of this frontier technology. Zhipu went public on the Hong Kong Stock Exchange last month alongside rival Minimax, another AI Tiger. Both stocks have rallied strongly as investors bet on the companies benefiting from China's AI Boom. The GLM5 release follows a series of updates, including version 4.7 last month and version 4.6 in September. The latest model is also optimized for with AI agents such as OpenClaw, the company said. The company has positioned its models as having strong coding and agentic capabilities that can perform multi step tasks. Zhipu, which faces US sanctions, derives most of its revenue from the domestic Chinese market but has overseas ambitions. Chief Executive Zhang Peng told Reuters in a September interview that overseas revenue was beginning to gain traction, though the company has yet to directly compete with US Models in consumer subscriptions. End quote well, speaking of another Tiger mention there, Minimax today released M2.5, claiming that model delivers on the quote intelligence too cheap to meter promise that it made recently priced at $0.30 per 1 million input tokens and $1.20 per 1 million output tokens, which okay that's aggressive and ByteDance's new AI video generation model C dance 2.0 that I told you about recently has apparently gone viral in China, with one state backed newspaper saying it bigger than Deep Seeks Sputnik moment. Quoting Reuters Again, ByteDance, which officially unveiled Seed Dance 2.0 on Thursday, said in a statement that the system was designed for professional film, e commerce and advertising productions because it was capable of processing text, images, audio and video simultaneously, lowering the cost of creating content on Chinese social media. Seed Dance 2.0 drew comparisons to Deep Seek's meteoric rise to fame earlier last year. The release of deep seq R1 sparked heated debate in the US tech community over a Sputnik moment, chinese state backed newspaper Global Times wrote in an editorial on Wednesday. This year, the continued breakout success of Seedance 2.0 and similar innovations has gone even further, giving rise to a wave of admiration for China within Silicon Valley. The buzz generated by CDance 2.0 was underscored when the world's richest man, Elon Musk, replied to a post praising the model on his social media platform X by commenting, it's happening fast. Users on China's Weibo microblogging platform shared videos generated by the AI model that showcase the complexity and image quality of its output, no matter how bizarre. The prompt 12 minute video, which had been viewed around a million times on Weibo, depicted rapper and record producer Yi, formerly known as Kanye west, and reality TV star Kim Kardashian as characters in a palace drama set in Imperial China, speaking and singing in Mandarin. End quote. Elon Musk went ahead and announced a major reorganization of XAI that required parting ways with some people after two XAI co founders said they were leaving. This week, XAI will be reorganized into four core areas Grox, Chatbot and Voice Products Coding Imagine and Micro Hard, which will build digital agents to run companies. Again. Microhard is a typical Elon joke, but it's technically more of a play on the name Microsoft. Also, at an all hands meeting, Elon continued to walk back his Mars ambitions for SpaceX, telling employees that Xai and SpaceX need a factory on the moon to build AI satellites and a massive catapult to launch them into space, quoting cnbc. The organizational overhaul was done, quote, to improve speed of execution, musk said in a post on X. He did not say which employees may have been cut as a result of the restructuring or which employees may have resigned of their own volition. We are hiring aggressively, musk added. Earlier this week, XAI co founders Jimmy Ba and Tony Woo announced their exits after several other founding members, including Igor Babushkin, Kyle Kossik, Christian Segerty and Greg Yang, previously departed, end quote, quoting Bloomberg. What matters is velocity and acceleration, musk told employees. If you are moving faster, you will be the leader. He also thanked the people who have departed the company. Most of the AI compute is going to be understanding real time video generation, musk said. And we expect to be leaders in that. Nikita Beer, who is in charge of X's, said the social network and its adjacent apps, including Grok, have reached about 1 billion users. January was the best month ever in terms of engagement for X, he said. He also noted that new users spend 55% more time a day in the app than they did six months ago. The app has been rebuilt to be better than ever and is now generating $1 billion in annual recurring revenue tied to subscriptions, he said, end quote. And quoting the Times, you have to go to the moon, Mr. Musk said during an all hands meeting, which was heard by the New York Times. The move would help XAI harness more power than other companies to build it's AI, he said. It's difficult to imagine what an intelligence of that scale would think about, but it's going to be incredibly exciting to see it happen, he added. Mr. Musk's fixation with the Moon is a recent one. Since founding SpaceX in 2002, he has said making humanity multi planetary first by establishing a colony on Mars was the company's raison d'. Etre. But in recent months he has posted frequently on X's social media platform about the company's new focus, which is the moon. Two former SpaceX executives told the Times on the condition of anonymity because they were not authorized to speak speak publicly about corporate plans that the moon had never been a main focus for the company. In his remarks on Tuesday, Mr. Musk described the moon as a stepping stone to Mars. First, he said the company would build a self sustaining city on the moon, then travel to Mars and finally explore star systems in search of aliens. If you're moving faster than anyone else in any given technology arena, you will be the leader and XAI is moving faster than any other company. No one's even close, Mr. Musk said. Because we've reached a certain scale, we're organizing the company to be more effective at this scale. And actually when this happens, there are some people who are better suited for the early stages of a company and less suited for the later stages, end quote. Google says Gemini has been inundated by commercially motivated actors who are trying to clone it, including one campaign that prompted gemini more than 100,000 times, quoting NBC News In a report published Thursday, Google said it has increasingly come under what are known as distillation attacks or repeated questions designed to get a chatbot to reveal its inner workings. Google described the activity as model extraction, in which would be copycats probe the system for the patterns and logic that make it work. The attackers appear to want to use the information to build or bolster their own AI, it said. The company believes the culprits are mostly private companies or researchers looking to gain a competitive advantage. A spokesperson told NBC News that Google believes the attacks have come from around the world world, but declined to share additional details about what was known about the suspects. The scope of the attacks on Gemini indicates that they are most likely or soon will be common against smaller companies custom AI tools as well, said John Holtquist, the chief analyst of Google's Threat Intelligence group. We are going to be the canary in the coal mine for far more incidents, holtquist said. He declined to name suspects. The company considers distillation to be intellectual property theft, it said, even though they have mechanisms to try to identify distillation attacks and block the people behind them. Major LLMs are inherently vulnerable to distillation because they are open to anyone on the Internet. OpenAI, the company behind ChatGPT, accused its Chinese rival Deepseek last year of conducting distillation attacks to improve its models. Many of the attacks were crafted to tease out the algorithms that help Gemini reason or decide how to process information, Google said. Hultquist said that as more companies design their own custom LLMs trained on potentially sensitive data, they become vulnerable to similar attacks. Tax let's say your LLM has been trained on 100 years of secret thinking of the way you trade. Theoretically, you could distill some of that, he said. End quote. You've been chosen to lead AI onboarding for your team that includes everyone from seasoned technical pros to AI newcomers. No pressure, right? That's where ARIA comes in. Areas no Code, Low Code and Pro Code platform makes it easy for your organization to embrace AI, regardless of technical experience. With Aeria's unified Security layer, advanced threat detection and robust compliance measures, your AI ecosystem can stay safe while your team innovates. Worried about staying on budget, ARIA's cost optimization tools can help you manage and forecast your AI spend so there are no surprises. Get started@aeria.com morningbrew that's that's area.com morningbrew if you've got Waymos in your town, you're about to see some different looking Waymos. Quoting cnbc, Waymo on Thursday said it has begun using its sixth generation driverless system to provide robotaxi rides to employees on Ojai vehicles, which use a base model made by Chinese automaker Geely. By upgrading their driverless tech and adding more vehicles to its fleet, Waymo aims to extend its U.S. lead and lock in loyal riders. The Alphabet owned company said its sixth generation Waymo driver uses more cost effective parts and should be able to navigate through harsher weather conditions than previous generations. The new system will serve as the primary engine for our next era of expansion, waymo Vice President of Engineering Satish Jayachandran said in a statement. Waymo is offering service on its Ojai vehicles, two employees and their guests in the San Francisco Bay Area and Los Angeles first before gradually expanding to new cities with a goal to open to public writers later this year. Waymo's decision to use Chinese electric vehicles in its US Fleet has raised GOP lawmakers concerns. We're locked in a race with China, but it seems like you're getting in bed with China, senator Bernie Moreno of Ohio said to Waymo's safety chief at a committee hearing last week. Spokesperson Sandy Karp told CNBC that Waymo will not provide, quote, any access to its closely held autonomous driving technology sensor data nor any writer information to zeekr. The Chinese automaker, a subsidiary of Geely, is responsible for providing base vehicles and Waymo installs its autonomous driving technology in the cars in the U.S. the company's sixth generation systems will also work with Robo taxis built on the Hyundai Iconic 5 line. Waymo's current Jaguar I Pace vehicles will continue to run on its fifth generation systems. We have consistently operated mixed fleets for years, including when we Transitioned from the 4th gen driver on the Pacifica to the 5th gen driver on the I Pace, karp said in an email. The Ojai is a boxier ride with a lower step and a higher ceiling than Waymo's existing robotaxis, but has about the same footprint as the Jaguar I Pace. End quote. Finally today, another day, another essay about AI to summarize for you, in his essay titled the AI Vampire, Steve Yeager offers a sobering critique of how he says artificial intelligence is reshaping the software engineering landscape. While Yegi has long been a vocal proponent of AI productivity, often touting 10x gains through vibe coding and advanced agents, this piece serves as a pivot toward the human and economic costs of that acceleration. He argues that AI has become a vampire that is systematically draining the life and value out of the developer community. The core of Yegi's argument centers on who actually benefits from AI driven productivity. He presents two extreme scenarios to illustrate what he calls a value capture problem. Scenario a. You use AI to work at 10x productivity for a full 8 hour day. Your employer captures 100% of this extra value. You receive the same salary, but you are now doing the work of 10 people. This leads to immediate burnout and sets an unsustainable new baseline for the entire industry. Scenario B. You use AI to complete your daily task in just one hour, spending the rest of your time resting or pursuing personal interests While you capture 100% of the value, which is your time. Your company becomes uncompetitive and eventually goes out of business because other firms are forcing their employees into scenario A. Yegi posits that we are currently stuck in scenario A. Companies functioning as capitalistic extraction machines, in his words, naturally lean into AI to squeeze every drop of productivity from their workforce. The result is a fracking of human intelligence, injecting AI into the workflow to extract value until the engineer is pumped out and exhausted. Yegi admits openly to his own so called addiction to these tools. Describing a cycle of intense high speed coding followed by nap attacks and massive unexplained fatigue. He suggests that the sheer volume of decisions an engineer must make when moving at 10x speed causes a level of cognitive load that the human brain isn't designed to handle. When you use an agent like Claude Opus 4.6, you aren't just typing, you are reviewing auditing and integrating code at a blistering pace. This is agentic burnout. It's a new phenomenon where the physical act of coding is replaced by the much more taxing act of continuous high stakes decision making. The essay reflects on how the world is accelerating against its will almost. Yegi notes that in the 1980s society had time to reflect on news and events. Today, the pace of AI development and the pressure to adopt it leave no room for reflection. This creates what he calls a multi whammy for developers. They are pressured by early adopters who control the narrative, by CEOs who are panicking to integrate AI, and by the fear of being left behind. Yehee concludes that the current path is unsustainable. If AI truly makes us 10x more productive, the logical and humane response should be not to work 10x more, but a radical reduction in working hours. He advocates for a move toward A four hour workday without a fundamental shift in how we value labor and time. Yegi warns that the AI vampire will continue to drain the industry. We are at a crossroads where we must either fight for a new normal that prioritizes human well being or face a future of perpetual AI accelerated burnout. According to Yegate, I've actually been thinking all night and all day about that essay. We led the show with yesterday the Is this a Covid Moment for AI Advances essay. Longtime listeners will remember that for various accidental reasons, I was early to the COVID concerns. We launched the COVID Daily podcast a week before lockdowns happened happened. So I was thinking, what if we made this show something of an early warning system for the ideas in that essay yesterday? We're all about the cutting edge of tech news around here, right? You all listening are the front lines of tech. So I want to make this request. If you see AI taking over your job, tell me about it and maybe we'll talk about it on the show. I'm not saying I'll bring you on or whatever, but send me, you know, anecdotes and things like that, and then I will mention them when appropriate. I'm not necessarily talking about layoffs because those will show up as headlines, though I guess I am talking about layoffs. If you're at a small startup and they suddenly do something crazy like replace the entire engineering team with AI agents. Tell me about that. But what I am talking about is the idea that was core to yesterday's essay. If you go into work tomorrow and see AI implemented in a way that makes you think, well, there's no more need for my job anymore. That's what I want to hear about. I want to know when that revelation comes to you, if it comes at all. And I want to know about it months, weeks before the layoffs hit again, this might never come to pass, but let's all band together and use the power of the mutant podcast army to collectively all get an early warning system for the tipping point when AI is truly making people redundant. If such a thing happens to you, get in touch@brianidehomefund.com I can promise you anonymity and all that good stuff if you want. Nothing again may come of all this. But like the COVID podcast, if the tidal wave of history does wash over all of us, let's harness the hive mind of this podcast audience to know about it before everyone else does. Thanks in advance. Talk to you tomorrow.