ZAI Launches Agentic Coding Environment

This episode is brought to you by Accenture. When your advertising operations fall out of sync, everything else follows. Spotify and Accenture are working together to reinvent the rhythm of ad sales using automation, analytics, and smarter workflows to simplify campaign delivery and access better data across the business. The result Less time spent on operations, more time connecting brands with the moments and fandoms that matter most. Learn more at Accenture d. com slash Spotify Welcome to the teech. We Witeome for Thursday, july second, twenty six. I'm Ban McCullll. Today Open eye floated giving the U.S. government a five percent stake to court the Trump administration's favor Nvidia promised to backstop cloud providers for a revenue cut. SpaceX showed investors an AI device prototype, Apple ramped up photable iPhone orders and ZAI launched Z code. Here's what you missed today in the world of tech The older I get, the more I really need a full eight hours of sleep just to function. att least that's until I tried Blueprint's longevity mix created by Brian Johnson. Just one S scoop helps support energy, cognitive functions, mood, focus, sleep, cellular resilience, and healthy aging The longevity mix is packed full of ingredients like magnesium, toreine, glycine, and creatine, all to help you age better and survive on less sleep Science backed, precision dose, no BS For a limited time, only our listeners get twenty percent off plus free shipping at blueprint. brya andjohnson dot com by using code TbrH at checkout. That's code TBrh at blueprint. brianjohnson.ot com for twenty percent off And after you purchase, they will ask you where you heard about them. Please support me and tell them I sent you. Sources tell the FT that openpenAI has discussed giving a five percent stake in itself to the U.S. government seeking to clear political obstacles by securing buy in from the Trump administration But let me pause and underline something that I just said so you don't miss it. I said, giving the US government, not selling the US government a five percent stake. Sam Altman, Chief executive of the Chat GPT Maker has argued that giving the public a financial stake in the company is the best way to share the upside of AI and has suggested a stake of this size in early conversations with the administration, according to two people familiar with the talks. proposed arrangement would involve other USAI companies handing over a similar stake, although it is not clear if the other labs would be willing to do so Giving the government an ownership stake could help secure good relations with the administration and would mark an attempt to address political blowback by sharing the wealth generated by AI with the public. AI labs have faced an increasingly challenging environment in Washington as the American public and politicians Grow more concerned about vast data center construction and the implications of AI for jobs and cybersecurity opening eye, and its chief rial anthropic. Recently, both had the release of their cutting edge models held up by U. S. scrutiny, while some Republicans and advisers to President Donald Trump favor much tighter regulation of the sector The two rivals are also preparing for public listings, which would expand their ownership base and generate big gains for current investors, although opening Eyes float may not take place until next year Altman and other open AI executives have suggested that each of America's leading AI developers allot five percent of their equity to a vehicle like the Alaska Permanent fund a sovereign fund that invests the state's oil wealth into stocks and pays dividends to the state governments and residents The companies might include Anthropic as well as Google, Meta and others, although it is not clear if any of these groups would agree with Open eye's proposal. After publicly attacking Intel's chief President Trump has swung behind the U S. chipmaker after the government took a ten percent stake in that company Conceptual talks between the government and open AIe were in the early stages and any deal might require an act of Congress to implement, the people said But the discussions point to a potential mechanism to distribute the financial gains from the technology Alldtman has been in active talks with the administration about the issue of public ownership, including Trump, Commerce Secretary Howard Lutniick, and Treasury Secretary Scott Besscent, according to multiple people familiar with the matter The openpen AI chief has also spoken to Democratic Senator Bernie Sanders in recent weeks. Sanders has pushed for public ownership of closer to half of each US AI company through a sovereign wealth fund, end quote. oting Dar Abasanjo. Intel stock is up over five hundred percent since the Trump administration took a ten percent stake in August of last year It makes sense for openen eye to seek a similar outcome, especially given Trump's recent interest in regulating AI models and the benefits of sweetheart government deals, given a Trump administration equity stake and quote Well, all of this is funny capitalism lately, if you ask me, back to the so called circular investing Nvidia is apparently promising to financially backstop young cloud providers like Fermmus rent out its AI chips in exchange for a revenue share through a new program quoting the information The backstop would come in the form of promising to rent back unused GPUs if the companies can't find AI developers to rent them. according to GPU cloud providers Fermis and Sharon AI, which are participants in the new program and three executives at other firms that do business with NVIidia The deals are part of a program some people at NVidia have dubbed the AI Compute Partnership, one of the people said NvidDia has offered to pay a guaranteed rate for the cloud providers' unsold GPU capacity. An NVvidia spokesperson also confirmed the program, which shows how the world's most valuable firm is increasingly leaning on its financial position to become a kind of central bank to the hundreds of companies that buy its chips in bulk GPU's are typically the most expensive part of an AI data center so guaranteeing to rent unsold GPU capacity makes it much easier for chip buyers with subpar credit ratings to get the loans they need Nvidia kills two birds with one stone in such deals, said a dataenter executive. If NvIDIia only backstops the leases tied to facilities, quote, then you're stuck with how do you finance the GPU's? But if NvIidia guarantees it will pay for unused computing capacity in the facilities, the GPU's get financed and the data center gets financed as well, the person said, end quote The journal says that SpaceX showed investors a handset like device prototype with AI tech from XAI inside it, as well as a proprietary OS, a snapdragonip and a design Simmer than an iPhone, quote SpaceX told investors that the project was at an early stage. The design could change, and it is unclear whether such a device will be made Representatives for SpaceX andQualcom didn't immediately respond to requests for comment The project is a sign of Elon Musk's sprawling ambitions as he builds a leading global satellite connectivity network, grows his rocket company, and creates new AI tools. AI companies are placing a variety of bets on the future form and functionality of AI powered devices The Wall Street Journal has reported that Elon Musk has in the past weighed building a smartphone due to frustration over how Apple controls distribution of third party apps such as X It isn't easy, however, for a newcomer to break into the hardware business The idea of making a phone makes me want to die, Musk said last October, but if we have to make a phone, we will. In February, Musk denied that his company was developing a phone pushing back against a Reuter' report that SpaceX was building one that would connect directly to its Starlink satellite network. We are not developing a phone, he posted on X But some investors in SpaceX and Tesla were told that Musk has long envisioned a consumer device to serve as a platform for his ventures technologies. A device could also help Musk become less reliant on other companies. Musk's XAI chatbot is typically used via Apple or Android devices. Musk's Starlink satellite service already sells a dish to connect to the internet and it offers cell phone service in dead zones through partnerships with companies including TMobile The device prototype SpaceX recently showed to investors draws on the philosophy behind Musk's everything app. A concept he championed when acquiring social media platform Twitter, now X in twenty twenty two Some of the people familiar with the matter said. Often referred to as super apps, these software programs are popular in Asia and incorporate services that Americans usually access by downloading separate individual apps. Users on China's WeCat and Alipay, owned by local technology giants Tenscent and Ant respectively, tap many apps and features available on the platforms to transfer money orrder food, book, travel and play games. This model has challenged the traditional app store business Chinese companies are turning to AI to reinvent this super app model. They are integrating AI agents into existing super apppps or trying to turn popular AI chatbots into the nucleus of new everything apps, end quote Sources say that Apple has told suppliers to produce around ten million foldable iPhones in twenty twenty six. That's up from their original estimate of seven to eight million and eighty million iPhones in total across all new models this year. Coting nine to five Mac The additional eighty million units of new iPhone models being introduced in the second half of twenty twenty six would bring the total units ordered for the year to two hundred twenty million per the report IDC recently forecasted that Apple would ship close to two hundred forty million iPhones in twenty twenty six. The report later says that Apple has told some suppliers to expect as many as eighty five new iPhone orders in the second half of the year As for the first foldable iPhone, Apple is expected to tag it with a premium price. IDC recently predicted that the iPhone Ultra will carry an average selling price of two thousand five hundred dollars with storage options priced as much as three thousand dollars. Apple is expected to unveil the iPhone eighteen Pro, iPhone eighteen Pro Max and fooldable iPhone Ultra in September The foldable iPhone could launch after the iPhone eighteen proro models, although it's still expected to arrive this year Meanwhile, Apple isn't expected to replace the standard iPhone seventeen with the iPhone eighteen until spring This would extend the life of the iPhone seventeen on the market from the typical twelve month run to around eighteen months Apple is similarly expected to upgrade the iPhone air introduced in september twenty twenty five with a new model in the spring of twenty twenty six, end quote You know, you could be experiencing productivity through a whole new lens literally Even G two are productivity smart glasses with teleprompting, conversation support, real time translation and AI. help you stay on top of work and daily life. They're made to look and feel like premium eyeewear with no camera and a lightweight thirty six gram design you can wear all day. The more context you give them, the smarter they get, adapting to how you work and what you need To learn more about E G two, go to evenrealities d. comot Use promo code tech Bw to get ten percent off E ring one and or even clip When you add them to your even G two order, H ever spent hours of your workday tracking down information only to find that it wasn't documented at all? It's an age old corporate conundrum Critical knowledge isn't available or accessible to the folks who need it That's the very problem our sponsor, Scribe was built to fix. Scribe is a workflow AI platform that automatically turns your workflows into clear cut documentation All you do is turn on the Scribe browser extension or desktop app and go through your processes as you would normally. Scribe builds a guide as you go, capturing every click, step, and screenshot automatically. So what would have taken hours gets done in under a minute. to see what sccribe could look like for your org, head to sccribe dot how slash ride home and mention ride home for your first month of sccribe capture free on select plans. That's SCRiB E d. how slash ride home Our roll battery is the same It's like castking ev full socer players are the same. Take Messi, the most decorated player ever Is there any other player who has achieved that? No, just him. Now take Duracl. Is there any other battery with power boost ingredients inside? No, just Duracl. Remember, goats only trust goats because they're built different. and Messi only trusts Duracl. And you might not think this is potentially the biggest story of the day, but It might turn out to be quoting Venture Beat. ZAI, the Beijing based artificial intelligence lab formerly known as Jipu AI. On Wednesday, officially launched ZCode, a free desktop application it describes as an aentic development environment purpose built for its flagship GLM five point two large language model The move marks the company's most aggressive push yet into the fast growing AI powered coding tool market where it now competes directly with Cursor, Clad code, GitHub C pilot, and Google's anti Gravity Red one way, Z code is simply another entant into a crowded market. Red another, It is a single product that crystallizes three of the most consequential trends in enterprise software today, the race to the bottom pricing of frontier AI models, the geopolitical balkanization of the AI stack, and the rapid maturation of agentic coding agents into wet Gartner estimates is a roughly ten billion dollars market Unlike traditional IDEs that bolt an AI on through a chat sidebar or autocomte extension, ZCode is best understood as an agent first development environment. Its core design is built around long horizon tasks. The user describes an outcome, the agent plans the work, edits files, runs checks, reviews progress, and continues across multiple iterations until the goal is met Z code organizes the development experience around the Z code agent Deeply tuned for GLM five point two with emphasis on deep integration, the model, tools, and execution workflow tuned together so that the agent fits continuous multi step real world development tasks The environment supports continuous followup across devices, desktop, mobile remote, and Faceo slash WeCat bot which can all keep the same workspace task moving Sensitive commands, file changes, and high permission actions go through confirmation before execution That remote control feature, the ability to steer a running coding agent from WeCat, FSu, or telegram on a phone is a differentiator that speaks directly to the Chinese developer market where those messaging platforms dominate professional communication You can keep checking progress and adding instructions while long running work continues from any device with these messaging apps The tool is free to download. Revenue flows through ZAI's GLM coding plan subscription tiers, which start at sixteen dollars twenty cents per month for a light plan and scaled one hundred forty four dollars per month for a max prices that undercut anthropics clawed code and cursors compatible tiers by significant margins Through july thirty first, ZCode is offering a promotional one point five X effective quota bonus for coding plan subscribers with off peak token consumption charged at six seven X as a coefficient. The platform also supports multiple AI models and agents, including clogged code, Codex, Gemini, and openCode, a pragmatic concession to the reality that no single model wins every task ZCode's value proposition is inseparable from GLM five point two, the model it was designed to showcase. ZAI released GLM five point two on june sixteenth first to its coding plan subscribers and subsequently as open source weights under the MIT license on Hugging Face, a sequencing decision that prioritized distribution over the traditional benchmark led launch The model's specifications are formidable. GLM five point two is a seven hundred forty four billion parameter mixture of experts architecture with forty billion active parameters, a genuine one million token context window, five times the two hundred thousand limit on its predecessor, and training on twenty eight point five trillion tokens It ranked second globally on Code Arena as of mid June, trailing only Anthropics Claud Fable five, making it one of the highest performing publicly available models for coding tasks On benchmarks, GLM five point two performs within striking distance of the best proprietary systems. It trails anthropics Clawed Ous four point eight by just one percentage point on frontier SWE, a benchmark measuring multi hour autonomous engineering projects while edging out OpenAI's GPT five point five It's API pricing, a dollar fortyents per million input tokens and four dollars and forty cents per million output are a cost reduction of up to eighty two percent compared to Anthropics Claud Opus four point eight at five and twenty five dollars respectively. Because Z code is a first party tool from the same company that makes the model, it requires no manual endpoint configuration. The model is just wired in ZCode's arrival cannot be separated from the geopolitical drama that has roiled the AI industry over the past three weeks regarding the Claud Mhos and Fable bands. ZAI's timing was surgical. On the same day the Trump administration ordered Anthropics's most advanced models blocked for foreign nationals, Jipu announced the open source release of GLM five point two with no usage restrictions The South China Morning Post reported that GLM five point two would be available to all users of JipPU's new GLM coding plan subscription priced at just a tenth of Anthropics premium Claud code and CladMax tiers The market responded accordingly, Jipu's AI market capitalization crossed one hundred twenty eight billion dollars on june twenty second, driven by a forty two percent intray share surge JP Morgan raised its twenty twenty six through twenty thirty revenue forecast for JPU by between seven and sixteen percent Following the launch, projecting an over five hundred and thirty four percent revenue surge for twenty twenty six and expecting the AI firm to turn a profit by twenty twenty eight The Fable five episode did more than embarrass anthropic. It introduced a new risk category into enterprise AI procurement, Sovereign access risk. When a government can disable a commercially deployed AI model overnight The traditional evaluation criteria of developer experience, benchmark scores and pricing becomes secondary to a more fundamental question Will this tool still work tomorrow againgst the entrenched coding players, Z Code's pitch rests on three pillars Deep first party integration with GLM five point two that no third party editor can replicate, aggressive pricing that starts at a fraction of Western competitors, and MIT licensed open weights that allow enterprises to self host eliminating the regulatory kill switch risk that the Fable ban made viscerally real, end quote Okay, so with july fourth coming up The observed holiday, as well as the actual holiday. I will be taking tomorrow off and also Monday off. So the next regular news episode will be on Tuesday, july seventh, but I do have a bonus episode for you tomorrow as well as one on Monday, two completely different and new how I AI episodes. one on how someone transformed their job using AI and one on how someone is managing their backyard with AI? Enjoy that. Talk to you on Tuesday Hey it's Ryan Renold here from MitMobile. Now, I was looking for fun ways to tell you that Mint's offer of unlimited premium wireless for fifteen dollars a month is back. So I thought it would be fun if we made fifteen dollars bills, but it turns out that's very illegal. So there goes my big idea for the commercial. Give it a try at mintmobile dot com slash switch Upront pay forty five dollars for months ninety dollars for six months one hundred ninety dollars forel month, reired fifteen dollars for a month equal to taxas bes extra. initial term only greater than fifty gigabyes be slow netk is busy t