Tech Brew Ride Home - Is Gemini 4 Argon Just Benchmaxxing?

Episode Date: October 1, 2026

Google unveiled Gemini 4 Argon, which you and I can't use yet, and insiders say it's weaker in practice than on benchmarks. Amazon refreshed the Kindle, Samsung quietly hiked Galaxy S26 prices, and McDonald's AI-priced your Big Mac, location by location. Google rolls out Gemini 4 Argon to trusted cyber defenders through Fairwind and says it is participating in the US government's voluntary pre-release process (Axios) Google announces Gemini 4 Argon, claiming industry-leading coding, knowledge work, and cybersecurity performance, with 1 million token output and $2/$10 API pricing, but it's still limited to trusted testers (Ars Technica) Google grapples with employee skepticism about Gemini 4: insiders say it shines on benchmarks but struggles with coding and front-end design, after Google abandoned Gemini 3.5 Pro (Bloomberg) Amazon updates the Kindle, Kindle Paperwhite, and Colorsoft with new colors and up to 32GB of storage, and unveils the $35 Kindle Click page-turning remote (The Verge) Hands-on with the new Kindles: the $34.99 Kindle Click Bluetooth remote, plus a $79.99 Page-Turn Cover that gives the Signature Editions of the Paperwhite and Colorsoft physical buttons (Gizmodo) Samsung quietly raises US prices for most of its Galaxy S26 lineup by $100 and the 1TB Galaxy S26 Ultra by $200; the Galaxy Z Fold 8 and Z Flip 8 are unchanged (Phone Arena) Investigation: McDonald's is utilizing an AI pricing engine across nearly 14,000 restaurants to generate location-specific “optimal prices” for menu items (Reuters) Subscribe to the ad-free feed.

Welcome to the TechBrew right home for Thursday, October 1st, 2026. I'm Brian McCullough today. Google unveiled Gemini 4 Argon, which you and I can't use yet. And some insiders say is weaker in practice than on benchmarks. Amazon refresh the Kindle, Samsung quietly hiked Galaxy S26 prices. And McDonald's is AI pricing your Big Mac location by location. Here's what you miss today in the world of tech. Every day shareholders meet to discuss important matters about the companies you invest in. Now you can make your voice heard too. Vanguard Investor Choice makes it easy to set your proxy voting preference for your eligible Vanguard index funds. Whether you hold a Vanguard fund directly or through another brokerage firm, all it takes is a few clicks to select your proxy voting preference and be heard on important shareholder topics like executive pay and director elections. Visit vanguard.com slash investor choice to learn more. It's your shares. It's your voice. It's easy. Vanguard investors own shares of our index funds and those funds own shares of the companies they invest in. Guard Marketing Corporation distributor.

Google has rolled out Gemini for Argon, but you and I cannot use it, quoting Ars Technica. Google promised Gemini 3.5 Pro in June, but it spent the summer trotting out smaller flash models. Now Google is ready to take on the frontier again with Gemini for Argon. The company claims this new AI offers industry-leading performance in coding, knowledge work, and cybersecurity, but you aren't allowed to use it yet. Most of us will have to wait to test Gemini for. Google says engineers inside the company are already using the new model extensively. Argon reportedly used fleet-wide telemetry data to help Google save 300 tebabytes of memory across its data centers. Meanwhile, Argon agents have been working to migrate C, C++ codebases to rust across Google, including thousands of lines in the core RE2 and LibGav1 libraries and more than 800. thousand lines in the fuchsia OS Zircon kernel. Google has also come armed with a raft of benchmarks to back up its claims on the software engineering deep sui v1.1 benchmark. Gemini for Argonne hits 77.9

percent, which is higher than GPT6 Astra, Fable 5.1, and Opus 5.5. Google promises similar power across a range of long horizon tasks pointing to Argon's industry leading score in the economic analysis VALs index test. This model is still in limited testing, but Google has announced API pricing. For a limited time, Argon will offer rates of $2 per million input tokens and $10 per million output tokens, and cashed input tokens will be discounted 95%. The company has also confirmed that Gemini 4-Argon will support a much higher output limit of 1 million tokens. That's up from 64,000 tokens in previous Gemini models.

Google says this allows users to complete more daunting tasks in a single, Google's step. The main focus for Gemini 4 right now is cyber defense. Google says models of this scale call for a phased release, so it's starting with a small group of trusted testers. Partners in the company's Fairwind program can get access to leverage the model's cybersecurity defense capabilities. Whiz is apparently already using Argon and has used it to uncover a critical vulnerability that could expose personal information in a system used at hospitals around the world. Google claims other frontier models miss this flaw, but it didn't provide any specifics. There's a lot of hand-wringing about model misalignment after several

high-profile hacking incidents over the summer. Google claims it designed Argon with systems that monitor the model's chain of thought and can stop it in its tracks if the model steps out of bounds. This, Google says, is why reasoning transparency is important in frontier model development. Gemini 4-4-Argon will eventually be available to enterprise and consumer customers, but Google isn't making any specific timeline promises. We do. do know that following the cybersecurity tests, general availability will start with paid API users and Google AI ultra subscribers, end quote. Artificial analysis says Gemini 4-Argon high matches GPT6 Astra Max on its intelligence index

and has a 15% hallucination rate compared with 51% for Astra. But, quoting Bloomberg, some insiders say those metrics don't tell the whole story, While Gemini 4 has performed well on benchmarks widely used to gauge model efficacy, it does less well when employees actually put it to work, according to people with direct access to the effort. The model struggles to handle certain coding tasks, said the people who requested anonymity to discuss an internal matter. Google has labored recently to develop models that can compete with OpenAI and Anthropic.

The company had planned to release a different version in June, dubbed Gemini 3.5 Pro, but abandoned the effort. the people said. Google said it would be inaccurate to say that Gemini 4 is underperforming in areas such as coding, but apparently there is a spectrum of opinion inside Google. Some employees believe Anthropics Fable and Open AIs astro models are improving at a faster rate than Gemini. These people believe that Gemini 4, even at its best, will still lag behind those models in some areas. Other employees believe the coming version has caught up with the leading AI labs. A Google employee familiar with the model development said there is large consensus internally at the company that Gemini 4 is at the frontier.

This person said the company had conducted rigorous tests of the models and denied that they struggle with messy real-world coding tasks. Google badly needs Gemini 4 to succeed. Versions of the model underpin nearly every product the company sells from the AI answers atop search, Google's main profit engine to Maps, Gmail, and Chrome. Each of those products has more than a billion users, a distribution advantage some of the company's rivals lack. Last November, Google debuted Gemini 3, a well-received model that was widely seen as a turning point for the company's efforts to keep up with Open AI and Anthropic. Google announced a new iteration called Gemini 3.5 Pro at its I.O. Conference in May and pledged

to release the model the following month. But that deadline passed, and the company has since abandoned 3.5 Pro, according to people familiar with the matter. Besides hampering Google's AI ambitions, the decision likely costs the company dearly in time and money. Training runs to build such a model can require spending as much as $400 million, according to Bloomberg intelligence analyst Mandip Singh. Using highly paid AI researchers could boost the price tag. Now Google is encountering various challenges with Gemini 4. Its coding abilities are uneven, according to people familiar with the model's internal evaluations. Gemini isn't particularly adept at front-end design, which shapes how apps and websites look and feel, one person said. That's a potentially

serious setback because Google has struggled to compete in the red hot market for AI coding tools. Moreover, it's a very large model, according to a person familiar with its development. Typically, big models are expensive to run potentially putting pressure on Google's margins. Experts say Google could be suffering from an industry tendency to focus on benchmarks, a phenomenon known as bench maxing, when engineers concentrate more on achieving. a good score than creating a product that does a job well. AI labs tend to do this because customers often judge models by their benchmark scores. Gemini 4 appears to be affected by this process, said to people familiar with the model. Edwin Chen, the founder of AI startup surge AI,

said relying on benchmarks can prompt labs to focus on building models that write code in a particular language rather than creating apps that are easy to use or well designed. An analogy would be, oh yeah, my kid got a really good score on the SAT, but the SAT does that. doesn't translate into real-world performance, Chen said. It's an incredibly pernicious problem, end quote. Samsung has quietly raised U.S. prices for most of its Galaxy S-26 lineup by $100 and the one-terabyte Galaxy S-26 Ultra by $200. The Galaxy Z-Fold 8 and Z-Flip 8 are unchanged, quoting Phone Arena. How bad is this price increase? In short, pretty, bad and it's obviously made a lot worse by the fact that the Galaxy S-26 S-26 plus and even certain Galaxy

S-26 ultra-variants were already costlier at launch than their Galaxy S-25 series equivalents from last year. Basically, every single Galaxy S-26 series model apart from one configuration is $100 more expensive than before all of a sudden. And that particular S-26 Ultra version is unfortunately receiving a $200 hike over an already extravagant $1,800 list price. That means the 6.3-inch S-26 now starts at a whopping $999.99 on its manufacturer's official U.S. website, while the most affordable 6.7-inch S-26 plus and 6.9-inch S-26 alter configurations are available for no less than $1,199 and $1,39, respectively. The good news, at least the recently released Galaxy S-26FE is not getting costlier just yet.

And the same, fortunately, goes for the Galaxy Z Fold 8, Z Fold 8, Ultra, Z Flip 8, as well as all Galaxy A series, Midrangers, and Samsung tablets in circulation right now. Don't forget you have options, though. For starters, you have the option of boycotting Samsung and even the mobile industry as a whole and hang on to your existing smartphone a little longer. If you can't do that for whatever reason, my advice is to wait until October 6th, yes, that's just a few days away, when Amazon's second epic Prime Day, sale of the year is scheduled to kick off. Amazon, by the way, was the first major U.S.

retailer to signal this Galaxy S-26 series price increase. But while the Vanilla S-26 and the S-26 are listed at their new regular prices at the e-commerce giant, the S-26 Ultra, is weirdly enough, still available for $1,09 and up, end quote. Introducing the Pod 6 from 8th Sleep. What makes it different from? the pod five. Think of this as an engine upgrade. Pod six is 50% smaller than the previous generation, packs nine times more sensors and delivers more powerful active heating and cooling. The pod can now read more of what your body is actually doing overnight and feeds a better signal into autopilot. New sizes include twin and twin XL and members see up to 34% more deep sleep. Use code ride home

at 8Sleep.com slash TechBrew Ride Home for a special offer. The pod six starts at $1,99, and you can have 30 nights to try it at home. Return it if you don't love it. Many people use nicotine as a pre-workout thinking it will make them better, but the reality is, in fact, the complete opposite. Nicotine is actually detrimental to your workouts for several reasons. It's a vaso-constrictor, so it narrows your blood vessels and restricts oxygen delivery to muscles. it raises cortisol, which can lead to muscle breakdown and longer recovery, and it can create

unnecessary cardiovascular stress during workouts. That's why so many people are using ultra pouches instead. They're made with clinically proven neutropics and adaptogens to deliver focus and energy, and they're 100% nicotine-free. Give ultra-pouches a try before your next workout and see how different you feel. Use code TBRH for 15% off your order at Take Ultra.com. That's TakeA Ultra.com with code TBRH for 15% off. Amazon has updated the Kindle, Kindle, Kindle, Paperwhite, and ColorSoft with new colors and up to 32

gigabytes of storage, and unveiled a $35 Kindle click page turning remote. Quoting the verge. Kindles don't usually change much with each new version, but Amazon is launching next generation models today that look and feel a whole lot nicer. The base Kindle, Kindle, Kindle Paperwhite, and Kindle ColorSoft are a millimeter thinner and 30% faster at opening books, and their power buttons have moved from the bottom to the right edge near the top. They ship in vibrant new colors with a soft touch plastic case, and for the first time, there's option for a couple shades of aluminum, giving them a higher end look and feel, not to mention double the storage of the plastic version at 32 gigabytes versus 16. Amazon VP of Kindle Kevin Keith told the verge that they all featured. user replaceable batteries to be in compliance with EU regulations too. The updated base Kindle is available today starting at $149.99 in a new purple color called

Ube and a planar graphite color. It comes with 16 gigabytes of storage and lock screen ads and it looks nicer than the last gen version thanks to a seamless transition from frame to screen, moving on from the old design that had a recessed display. New for this revision is the choice of aluminum which bumps the price up to 18999, doubles the storage, and ditches ads. The Kindle and aluminum is a very nice compact thing to hold, especially in its new folio case that it can magnetically attach to. Though at 23999, altogether 18999 for the e-reader and 4999 for the case, you're firmly in the territory of the paper white with its more pixel-dense and auto-dimming screen for 199-99-99. The display specs for the new Kindle Paperwhite and Color Soft models coming October 28th,

available to pre-order now haven't changed, though the color options have. The new plastic and aluminum colors available for each cover the entirety of the e-readers, not just their backsides. Amazon's Keith credits a new technique of applying a black layer between the edges of the panel to prevent the case's color from visually bleeding onto the screen. The aluminum options for the paperwhite and color soft costing $249 and $319, respectively include more than the base Kindle's high-end version. These are the new signature edition versions with Pogo pins on their rear that are compatible with a $79.99-paget

magnetic case that adds physical page-turning buttons a la the Kindo Oasis. Wireless charging isn't offered in the new signature additions. Amazon will sell a $59.99 charging dock that can charge this edition upright and attaches via Pogo Pins, end quote. And quoting Gizmodo. In addition to the refreshed Kindle lineup, which also includes kids' versions, Amazon is also selling a Bluetooth page-turning remote called the Kindle Click. The 3499 clicker has two big buttons for turning back and forward on e-books, and two buttons on the right side for adjusting screen brightness. It's basically Amazon's answer to the page-turning remotes like the Bookes Tappy.

The more interesting page-turning accessory is the page-turn cover accessory, in my opinion. The 79-99 case gives the signature additions of the... new Kindle Paperwhite and Kindle ColorSoft physical buttons. It only works with the signature additions of those two e-readers because it uses the Pogo Pins to connect. It's a good, simple solution to giving consumers who want page turn buttons without forcing it onto the device design, end quote. I've said this before, but this is a particular hobby horse of mine. If you make an e-reader without physical page turn buttons, you've done it wrong. This single issue is why I abandoned the Kindle ecosystem years ago for the books, and I don't regret it.

According to a Reuters investigation, McDonald's is utilizing an AI pricing engine across nearly 14,000 restaurants to generate location-specific what they're calling optimal prices for menu items. Quote, McDonald's is increasingly using artificial intelligence to guide menu prices across the U.S. and some global markets, a plan that aims to boost headquarters profit, but risks alienating customers and attracting antitrust scrutiny. One pricing factor supercharged by AI, an estimate of how much each store's patrons are willing to pay. The details of how that pricing system works, its extensive use of AI, the company's regulatory concerns, and the tensions with its franchisees have it previously been reported. For this story, Reuters reviewed screenshots of

the company pricing engine taken in August and interviewed nine sources with firsthand knowledge of the burger chain's strategy. McDonald's pricing engine uses machine learning algorithms to continually analyze data from millions of daily transactions across McDonald's nearly 14,000 restaurants, and generate what the company calls the optimal price at each location for each menu item, from Big Macs to discounted coffee for seniors. The engine has widened existing price differences for the same product between restaurants, including from neighborhood to neighborhood within the same area,

three franchisees told Reuters. Screenshots of the interface franchisees use, reviewed by Reuters, show messages including your restaurant is showing medium sensitivity to price based in part on customer willingness to pay in your area. The platform also contains public price information pulled from online menus of nearby restaurants of competitors, including Wendy's and Burger King.

Those chains said they do not use AI in pricing decisions. A Reuters check of prices in September on the McDonald's mobile app showed price differences. For example, a company-run store in Fresno, California sells a Big Mac for $569, but another company-run restaurant, two miles away, sells the same sandwich for $6.89, a 21% premium. Reuters could not confirm if the price difference is the result of the price engine's recommendations or other factors. McDonald's says its franchisees are free to set their own prices, but five store owners told Reuters the company pressured them to use the AI pricing tools.

A franchisee document from June reviewed by Reuters shows McDonald's records franchisees deviations from the recommendations in detail. In January, McDonald's started requiring franchisees to be constructively engaged with McDonald's approved pricing consultant and tools as part of its new business standards according to an internal communication sent to franchisees and reviewed by Reuters. McDonald's said in a statement, costs and other factors vary across its stores and that restaurants, even a few miles apart, can belong to distinct markets. Its pricing portal is a tool, not a mandate, designed to provide restaurant-specific recommendations to help franchisees deliver value for customers and make informed business decisions, McDonald's

said. The company also described Reuters reporting as speculative and uninformed claims that attempt to, quote, recast a standard business practice as something controversial, end quote. Nothing more for you today. Talk to you tomorrow. You know, Thank you.