Logistics technology increases marketing ROI by integrating real-time inventory and fulfillment data directly into ad channels, preventing wasted ad spend on out-of-stock products while boosting conversion rates through precise delivery promises. Connecting warehouse management and shipping infrastructure to marketing systems lowers customer acquisition costs (CAC) and increases long-term customer lifetime value (LTV).
When e-commerce brands evaluate campaign performance, they often look exclusively at ad copy, bidding algorithms, or audience targeting. However, marketing ROI is heavily dictated by what happens after a shopper clicks an ad: whether the product is in stock, how quickly it will arrive, and whether the fulfillment experience encourages repeat business. Modern logistics technology turns supply chain operations into an active conversion driver rather than a passive back-office cost center. Quantifying that connection means looking past ad platform dashboards, and resources that break down the ROI of logistics technology investments can help teams build the financial case for making that shift.
Quick Take
Running advertising campaigns without real-time supply chain integration leads to wasted ad spend, higher bounce rates, and lower customer retention. Connecting inventory, order management, and shipping systems directly to marketing platforms dynamically pauses ads for unavailable inventory, surfaces precise delivery dates on product pages, and protects campaign profitability. Integrating logistics into your growth strategy converts operational accuracy into higher conversion rates and superior Return on Ad Spend (ROAS).
The Direct Connection: How Operations Dictate Marketing Efficiency
Digital advertising returns do not depend solely on campaign settings inside Google Ads or Meta Ads Manager. Profitability relies on operational synchronization between supply chains and marketing channels. When a brand advertises a product that is backordered or temporarily out of stock, every paid click results in immediate site abandonment, inflating acquisition costs without yielding revenue.
Advertisement
Conversely, when logistics systems share real-time data with storefronts and ad networks, marketing teams can optimize campaigns based on physical stock levels, fulfillment center proximity, and shipping transit times. Aligning operational execution with customer acquisition forms a foundational pillar of a comprehensive digital marketing strategy.
1. Stopping Ad Spend Waste with Real-Time Inventory Syncing
One of the most immediate leaks in e-commerce marketing budgets is paid traffic directed to out-of-stock products. When a popular item sells out in the warehouse, ad platforms often continue displaying Shopping ads or dynamic retargeting banners for hours or days if product feeds are updated manually or via slow batch processing.
Integrating modern Warehouse Management Systems (WMS) or Order Management Systems (OMS) with catalog feeds automates ad availability. For example, e-commerce platforms can feed inventory thresholds directly into advertising platforms. Rather than deleting listings, merchants can deploy the official Google Merchant Center pause attribute to temporarily halt ad delivery without losing historical listing status or performance history.
Automating this pipeline ensures ad spend is directed exclusively to items that are ready to pack and ship immediately. Teams evaluating their fulfillment stack alongside specialized inventory management software tools can eliminate wasted pay-per-click (PPC) budget while maintaining steady inventory turnover.
Advertisement
2. Converting Shoppers with Dynamic Delivery Promises
Shoppers consistently rate shipping speed and delivery clarity among their top purchasing criteria. Displaying vague shipping timelines like “3–7 business days” introduces purchasing friction at the exact moment a visitor considers buying.
Logistics platforms calculate real-time transit times by analyzing warehouse processing schedules, carrier transit limits, cut-off times, and destination zip codes. According to Kibo Commerce fulfillment research, 75% of online shoppers are more likely to complete a purchase when an exact Estimated Delivery Date (EDD) is displayed on the product page and checkout screen.
Implementing continuous delivery date calculations relies on shipping management engines like ShipperHQ delivery promise solutions, which evaluate multi-origin fulfillment constraints to surface accurate delivery dates. Pairing accurate delivery promises with an e-commerce conversion rate optimization framework directly increases site-wide conversion yield without requiring additional media spend.
3. Maximizing Customer Lifetime Value Through Post-Purchase Delivery
Acquiring a new customer is expensive, and first-order profit margins are often thin due to initial ad costs. Sustainable profitability relies on building high Customer Lifetime Value (LTV) through repeat orders. A single late delivery or lost package damages brand trust and eliminates the likelihood of future purchases.
Advertisement
Empirical research on e-commerce logistics published in RSIS International last-mile delivery studies demonstrates that delivery accuracy is the single strongest predictor of customer loyalty (beta = 0.412), outranking both overall delivery speed and pricing flexibility. When fulfillment technology provides accurate delivery tracking and proactive exception handling, customer satisfaction rises and repeat order rates increase.
Optimizing post-purchase tracking reduces “Where Is My Order?” (WISMO) support calls while keeping shoppers engaged with branded tracking portals. Higher retention rates allow marketing teams to invest in aggressive customer lifetime value strategies, knowing that operational efficiency will protect long-term margins.
4. Geo-Targeted Ad Bidding Based on Fulfillment Proximity
Shipping costs vary significantly based on geographic distance between fulfillment centers and customers. Standard national ad campaigns treat all regions equally, bidding the same amount for a click regardless of whether the order will ship locally via inexpensive ground transit or cross-country via expensive express air freight. Underlying transportation costs, including the fuel consumed by delivery fleets themselves, factor into which regions stay profitable to advertise into, which is one reason logistics operations look at fleet fuel cards as a way to keep per-mile fuel spend more predictable across their delivery vehicles.
Logistics technology provides spatial inventory awareness, permitting brands to align advertising bids with regional fulfillment capabilities within an omnichannel retail marketing strategy. Marketing platforms can ingest real-time regional stock levels to make strategic adjustments:
Advertisement
Proximity Bidding: Increase bid adjustments in zip codes surrounding regional warehouses where ground shipping offers 1-to-2 day delivery at minimal cost.
Regional Ad Pausing: Automatically scale down or pause heavy promotion of heavy SKUs in zones where freight surcharges erode profit margins.
Localized Inventory Promotion: Target specific product campaigns strictly to geographic regions where local fulfillment nodes hold surplus stock.
5. Protecting Campaign Margins via Reverse Logistics Automation
Product returns present a major challenge to digital marketing ROI. High return rates dilute reported Return on Ad Spend (ROAS) by reversing top-line revenue after ad costs have already been incurred. Traditional manual return processing is slow and expensive, often resulting in full cash refunds and unsellable inventory.
Automated reverse logistics platforms transform return flows by automating exchanges and inventory grading. As detailed in Redo reverse logistics operational benchmarks, deploying self-service return portals and automated exchange logic preserves revenue that would otherwise be lost to refunds, converting up to 30% of return requests into replacement sizes or alternative products.
Combining reverse logistics software with a dedicated post-purchase customer experience plan protects net revenues, keeps acquired customers inside your ecosystem, and prevents marketing acquisition budgets from dissolving through unrecovered product returns.
Common Misconceptions About Logistics and Marketing
Aligning operations with marketing requires avoiding several common operational pitfalls:
Assuming Speed Matters More Than Accuracy: Promoting ultra-fast delivery promises that the fulfillment network cannot reliably hit damages brand trust far more than setting a realistic, guaranteed delivery date.
Treating Logistics as an Isolated Cost Center: Viewing warehousing and shipping strictly as expenses rather than revenue drivers prevents organizations from leveraging operational performance in ad messaging.
Relying on Manual Product Feed Updates: Updating product availability feeds manually or on 24-hour batch cycles leaves massive windows where out-of-stock items continue consuming ad spend.
Ignoring Return Data in Campaign Bidding: Failing to feed return rate data back into marketing analytics can lead ad channels to over-invest in high-volume SKUs that carry unsustainably high return rates.
Where Logistics Technology Cannot Fix Marketing Returns
While supply chain integration improves marketing efficiency, technology cannot overcome fundamental product or strategic flaws. Logistics software will fail to boost marketing ROI under the following circumstances:
Poor Product-Market Fit: Fast shipping and real-time inventory tracking cannot compel customers to buy products that lack perceived value or competitive pricing.
Weak Ad Creative and Messaging: Operational efficiency ensures accurate delivery, but uninspiring ad copy and weak value propositions will fail to generate initial click-through traffic.
Insufficient Inventory Volume: Advanced geo-bidding and dynamic delivery engines require baseline inventory availability across multiple nodes to function effectively.
Key Takeaways
Sync Feeds in Real Time: Connect WMS/OMS platforms directly to ad channels to automatically pause ads for out-of-stock items using Google Merchant Center attributes.
Display Precise Delivery Dates: Surface dynamic Estimated Delivery Dates (EDDs) across product pages and checkout to reduce cart abandonment and increase conversion rates.
Focus on Delivery Accuracy: Recognize that on-time delivery accuracy is the primary operational driver of customer retention and repeat purchase LTV.
Leverage Geo-Targeted Bidding: Adjust ad bids based on regional warehouse stock and ground shipping proximity to maximize order profitability.
Automate Reverse Logistics: Use self-service return platforms to convert return requests into product exchanges, preserving net revenue and campaign ROAS.
Frequently Asked Questions
How does connecting an inventory feed to ad platforms lower Cost Per Acquisition (CPA)?
Syncing inventory feeds directly to ad platforms lowers CPA by eliminating ad spend on out-of-stock items. When an item sells out, real-time feed updates automatically pause campaigns or update ad attributes. This ensures that every paid click lands on an actionable, in-stock product page capable of converting, reducing wasted clicks and driving down total customer acquisition costs.
Will displaying an Estimated Delivery Date (EDD) slow down website page speed?
Not if implemented using asynchronous API calls and server-side logic. Modern delivery promise platforms calculate transit times using optimized cloud microservices that query warehouse schedules and carrier data without delaying initial page rendering. Asynchronous execution preserves site loading speed and Core Web Vitals performance while displaying accurate delivery estimates.
Advertisement
How does reverse logistics automation directly protect marketing Return on Ad Spend (ROAS)?
Reverse logistics platforms protect ROAS by automating immediate product exchanges during the return initiation process. By offering dynamic size swaps, color variations, or instant store credit within a self-service portal, the platform retains customer revenue from the initial ad conversion rather than issuing a full cash refund that negates the ad spend efficiency.
Samsung’s upcoming “budget” flagship has been making the rounds in the rumor mill over the past couple of weeks. From disappointing camera leaks to build refinements, nearly everything about the Galaxy S26 FE has been revealed. Now, another fresh leak has even given us a proper look at the phone along with its color options.
A fresh set of leaked renders from AndroidHeadlines gives us a clear look at the Galaxy S26 FE in all three of its expected color options. The images line up with an earlier color leak pointing to Graphite, Aqua Green, and a third finish sitting somewhere between blue and purple.
At least one color has some personality
Graphite is exactly what you would expect from a modern Samsung phone. It is the dark, understated, and safe look. Aqua Green brings more life to the lineup, while the blue-purple option is easily the most distinctive of the three in the leaked renders. The latter also looks close to the new color that debuted along with the Galaxy S26.
AndroidHeadlines
There could still be additional Samsung.com-exclusive finishes that have yet to leak, but the current reporting points specifically to these three colors for the main lineup. Samsung has not officially confirmed any of them.
It really wants to look like a Galaxy S26
The renders also reinforce the biggest visual change we had already seen in earlier leaks. The Galaxy S25 FE’s three individually protruding rear cameras are now being replaced by the new design language as seen in the Galaxy S26 lineup.
Advertisement
AndroidHeadlines
Previous leaks have pointed toward an Exynos 2500, 8GB of RAM, and Android 17. But this is still just unconfirmed rumors so far. We covered the alleged Exynos 2500 benchmark recently, where the S26 FE trailed the newer Exynos 2600 inside the standard Galaxy S26 by a sizeable margin.
Samsung is also expected to launch the phone around September, which would line up closely with last year’s Galaxy S25 FE release. This is still just a leak, so we’ll have to wait for an official Samsung announcement to know if these colors are actually confirmed.
Christian Ivan worked for months within Blender, carefully creating a flawless 3D model of Manhattan Island, from a desolate landscape in 1600 to 2026. The final output was an 8-minute piece of work that scarcely claimed to be historically accurate. Ivan admitted to using ancient maps and atlas photographs from the New York Public Library to achieve the desired effect.
The time-lapse begins showing the island appearing desolate, with the occasional tree and some sand beneath a perfectly clear blue sky. There’s a hint to the Lenape people at the outset, before a handful of Dutch ships arrive and a little cluster of structures forms on the southern side. And that’s when New Amsterdam emerges, complete with twisting roads and fortresses to keep things under control. As the town grows, more figures emerge on the screen, serving as a cruel reminder that each new house or warehouse was erected on the backs of a small group of people who live on this tiny island.
THE CITY THAT NEVER SLEEPS – Construct the ultimate souvenir with the LEGO Architecture New York City – The Big Apple (21066) building set for…
CREATE 6 ICONIC LANDMARKS – Builders can create six NYC architectural icons in minimalist white, including the Brooklyn Bridge, Empire State…
DISCOVER HIDDEN DETAILS – As they construct their city skyline, builders can discover details that are uniquely New York, such as yellow taxi cabs…
When the British take over, the settlement begins to expand out little. More docks are erected along the waterfront, and the street begins to grow north. It’s moving at a fairly steady pace, and the shoreline is shifting as rivers are progressively swallowed up by piers and filled in. As we enter the early American years, we begin to see a few more dwellings and dock work being built, as well as a continual stream of ships arriving faster than ever before. Warehouses begin to proliferate, and what was once a few colonies begins to resemble a proper operating port metropolis.
Then comes the Commissioner’s Plan of 1811, and in a single stroke, the middle of the island becomes gridlocked, with the old straight avenues and numbered streets marching north right where all that previously empty or sparsely populated land used to be, and that was a single decision made, and we still have it that way to this day. Growth accelerates from there. During the nineteenth century, the old empty blocks began to fill up with large brick buildings, factories, tenements, and industrial businesses, all piled on top of one other. As a result, the skyline begins to rise; at first, it is just trudging along, but you can tell it is starting to lift, and before long, you can see the skyline reaching into the sky.
Ivan demonstrates his technique by walking through the entire process three times in a row. First, you get a good bird’s-eye view of the entire island, giving you an idea of its overall form and how development progressed from south to north. The camera then zooms in on Lower Manhattan, where the oldest streets still exist and the first skyscrapers are beginning to rise. Following that, the camera moves on to Midtown, where the grid has become more organized and the towers have begun to cluster. Watching the same old history unfold from three distinct perspectives makes it much simpler to grasp the scope of the changes that occurred. Lower Manhattan is the first to be settled, and its unique early layout remains until steel frames arrive and skyscrapers begin to take off. Midtown is filled in a little later and more methodically, with avenues converted into office skyscraper corridors.
The twentieth century begins in a rather obvious manner. The Empire State Building appears on the scene, and additional Art Deco buildings spring up all around it. The twin towers of the World Trade Center emerge briefly before disappearing, leaving you with a tranquil visual reminder that huge changes may not always arrive as expected. The newer skyscrapers keep appearing in the last act, and the model is changed correspondingly. The island’s population continues to rise until it reaches its current level. [Source]
Apple has released macOS Tahoe 26.6.1 to fix a Screen Sharing vulnerability that could let an attacker on the same network authenticate without valid credentials.
The timing makes this update more interesting than its small version number suggests. Apple’s security notes for macOS 26.6, released July 27, already listed three separate vulnerabilities affecting Screen Sharing Server.
Apple is now following that release with another Screen Sharing fix less than two weeks later. The company has also issued corresponding updates for macOS Sequoia and Sonoma.
Rachit Agarwal / Digital Trends
What Apple fixed in macOS 26.6
Apple’s macOS 26.6 security notes identify three distinct Screen Sharing Server vulnerabilities, each with its own CVE.
CVE-2026-43779 could allow an app to intercept network connections intended for another process. CVE-2026-43777 covered a remote denial-of-service risk, while CVE-2026-43760 could allow an app to access sensitive user data.
Advertisement
Those aren’t three attempts to squash the same bug. They’re separate vulnerabilities with different potential consequences, which is worth keeping in mind when looking at the latest update.
Why the new flaw is different
The issue fixed in macOS 26.6.1 is more directly tied to authentication. Macworld reports that an attacker on the same network could potentially authenticate to Screen Sharing without valid credentials.
Apple
That puts it in a different category from the three flaws patched in macOS 26.6. Still, seeing several unrelated Screen Sharing vulnerabilities addressed across consecutive macOS releases makes the follow-up patch harder to dismiss as routine maintenance.
Which Macs should get the update
The fix extends beyond macOS Tahoe. Apple also released macOS Sequoia 15.7.9 and macOS Sonoma 14.8.9 for the same Screen Sharing authentication issue.
If you use Screen Sharing, or keep it enabled on a Mac that regularly connects to shared networks, installing the update promptly is the sensible move. The latest flaw involves authentication, and Apple has now pushed fixes across three macOS generations rather than limiting the patch to Tahoe.
There are a few other affordable gaming laptops I haven’t tested yet that are worth considering. Asus and Dell both have RTX 5050 options, but neither is under $1,000. The Asus ROG A14 is out there, but it is currently only available with an RTX 5060.
Other affordable RTX 5050 laptops include the Omen 16, which is currently selling for $1,100. I haven’t tested that device in a while, but it’s a solid deal. The cheapest deal I’ve seen is the HP Victus 15 with the RTX 5050, which is only $999. If you want to upgrade to the next level of GPU performance, such as an RTX 5070, you’ll see a significant jump in price. One of the cheapest RTX 5070 gaming laptops is the Asus ROG Strix G16, which sells for $1,760. That should give you an idea of how performance scales up from what is considered “budget.” Read our Best Gaming Laptops guide for more high-end picks.
Advertisement
Here are two other affordable gaming laptops I’ve tested that are worth mentioning:
Photograph: Luke Larsen
Lenovo LOQ 15 Essential Gen 11 for $1,850: If the version of the LOQ 15 I recommended above isn’t available, I would consider this one more seriously. If (and when) that older model goes out of stock, my opinion here will change. But as it stands, it’s hard to justify this “newer” laptop, which pairs a three-year-old GPU with a three-year-old CPU (the Core i7-13650HX). The good news is the display is on par with the other LOQ 15, which is a step up in color from the Alienware 15. But the deal-killer is the included 135-watt power adapter. Like the Acer Nitro V 16 below, it loses power when playing games in Performance mode despite being plugged in. That means the extra 10 to 15 percent of higher frame rates you get in Performance mode is on a time limit. Also, at its full price of $1,850, it’s way overpriced, so wait for a discount.
Photograph: Luke Larsen
Acer Nitro V 16 for $1,040: Much can be forgiven if the price is low enough, but unfortunately, the Acer Nitro V 16 isn’t as cheap as it once was. The display is bright enough and has a fast 180-Hz refresh rate, but the colors are an issue. As I saw in my tests, the color accuracy and gamut are wildly off, resulting in some funky, desaturated coloring. The other big issue with the Acer Nitro V 16 is the power supply. Like the LOQ 15 Essential, it also only comes with a 135-watt charger rather than a larger 180-watt brick that most RTX 5050 laptops have. While the smaller power brick is nice, it means you can’t run intensive games in Turbo mode without draining the battery—even while plugged in. I was told that the updated 2026 model would fix this problem, but it still lists a 135-watt charger. I was fine to overlook that issue when the laptop was selling for less than the competition, sometimes as low as $750. At its current price, though, you’re better off with the MSI Cyborg A15.
Advertisement
What You Should Look For in a Cheap Gaming Laptop
Most cheap gaming laptops share a lot in common. They’re all usually between 0.8 and 1 inch thick and tend to have bare-bones gaming-capable hardware. Here are some of the key specs to look for:
Display: 15-inch or 16-inch display. Depending on the aspect ratio, they’ll have basic displays with a standard 1920 x 1080 or 1920 x 1200 resolution. You won’t find higher-resolution panels on gaming laptops under $1,000. Also, take a look at the refresh rate. 144 Hz is the standard, but the higher the better for less motion blur and smoother animation. While OLED and mini-LED are more common in higher-end gaming laptops, all budget-oriented options use LED IPS.
CPU: The latest processors from Intel and AMD will all be here, and in the budget tier, the differences aren’t as significant as in higher-end options. For AMD, that’s usually either the Ryzen 5 220 or Ryzen 7 250. Intel’s latest gaming chips in this price range are the Intel Core Ultra 5 225H or Core Ultra 7 240H. While Intel has announced its next-gen Core Ultra Series 3 chips, these haven’t come out yet.
Advertisement
GPU: We’re currently in Nvidia’s RTX 50-series graphics cards, which came out at the beginning of 2025. In gaming laptops under $1,000, you’ll be stuck with either the RTX 5050 or 5060. These likely won’t be replaced until at least 2027, so it’s safe to buy these for now.
Memory: You want at least 16 GB of RAM, and that’s typically what you’ll be stuck with in budget gaming laptops. Many gaming laptops let you upgrade RAM yourself later, though with the price of stand-alone memory these days, it might not be a bad idea to configure it with 32 GB upfront.
Storage: Gaming laptops start at 512 GB, and that will be enough for most. Upgrading to 1 terabyte isn’t a bad idea, though, whether that’s configured up-front or done yourself later. You can always store games on an external hard drive, but with the size of games these days, the more storage you have, the better.
What About Older Gaming Laptops?
Advertisement
The laptops above are the cheapest gaming laptops you should buy, but as you’ll discover with some research, there are some even lower-priced options out there—some as low as $500 or $600. Most come with older graphics cards and processors, meaning you’re getting a dip in performance. This is especially true if you buy something like an RTX 3050 or 2050. The RTX 50-series is the latest version, which came out at the beginning of 2025. The RTX 40-series was announced in 2023, meaning anything older than that is well over three years old. I wouldn’t recommend anything that old, especially not if it was already budget-oriented. I haven’t seen anything cheap enough that would make me think you shouldn’t go with one of the RTX 50 series options above instead.
How We Test Budget Gaming Laptops
Here at WIRED, we test budget gaming laptops the same way we test high-priced ones. Gaming comes first, and we run a series of in-game benchmarks to establish a baseline that can be compared apples-to-apples, including titles from different genres like Cyberpunk 2077, Marvel Rivals, and Monster Hunter: Wilds. This spits out a score, yes, but while it’s running, we pay attention to factors like fan noise, internal temperatures, and surface temperatures. We also test the game in more realistic scenarios to get a feel for how all the elements of the system come together, including the screen. We also use other CPU and GPU benchmarks like 3DMark Steel Nomad and Cinebench to evaluate performance.
Speaking of the screen, we use a colorimeter to test the breadth and accuracy of the colors, the peak brightness, and the contrast. These are all important and contribute to the overall experience of using the laptop both in and out of gaming. We also test battery life in local video playback. This is not so much for gaming, but more to see how well it works for other work or for school. We also test out the speakers and webcam to see how they hold up.
Advertisement
Lastly, we handle these laptops to get a sense for durability, build quality, and usability. That means typing on the keyboard, swiping on the touchpad, opening and closing the hinge, carrying it from place to place, and applying pressure to the lid and palm rests. Add in the included specs, price, and configuration options, and it comes together in something that I’d either personally recommend or not. For budget-oriented devices, price is given more weight in this evaluation. If one gaming laptop costs more than another, it needs to add something that offers a qualitative benefit.
When you think about it, time is all we’ve really got. Where to go from there is ultimately up to you. Maybe you use an app to track every task, or just go with the onboard timer. But that can be a lot of steps to begin with, and then the phone screen goes dark again. For some people, the whole out of sight, out of mind thing will kick in. At worst, you get distracted, start doing something else, and then feel guilty and frustrated when the timer starts going off.
But there’s hope for us visual simpletons, and the purveyor of that hope is [Edris] of Ponderly Robotics. You see, [Edris] created an extremely easy-to-use timer that looks like something you’d find on Captain Picard’s desk as a token of defeating the Borg. But the affably-named FlipBuddy is far more useful than that description implies.
[Edris] uses the open-source FlipBuddy every day, and swears by its simplicity. The point is accessibility, and respect for privacy. That said, there’s a companion app to provide insight.
Basically, you assign a task to each cube face. Choose one, and place the cube with that side facing up. FlipBuddy wakes up, connects to WiFi, and then pushes your session to the cloud, bypassing the need for your phone.
Time to switch tasks? Just put the new side face up. When you’re done for the day, use the stop face, which we’re hoping means to set it on the knocked-off corner.
Advertisement
You don’t need much to make FlipBuddy come to life. [Edris] used an ESP32 (an S3 SuperMini or similar will work), an MPU6050, six WS2812B LEDs, and a 3.7 V Li-Po cell. The beautiful, 3D printed origami mesh enclosure prints as a single, flat piece, and you get to fold it up around the internals and make your new buddy come to life.
Part of the point of FlipBuddy is that it can become as intuitive as punching a chess clock. So if it’s buttons you’re after, check out this simple Pomodoro timer.
A new study argues that life may have emerged twice on Earth. “Two of the main lineages of life — bacteria and archaea — may have independently figured out the secrets of metabolism that upgraded them from non-living to living,” reports ScienceAlert, citing a new study that “focused on the most rudimentary set of chemical reactions thought to enable this transformation.” From the report: “The surprise is that the enzymes that catalyze those reactions are not conserved across the evolutionary divide that separates bacteria and archaea,” says William Martin, a biologist at Heinrich Heine University Dusseldorf in Germany.
“The new data leave only one conclusion. The bacterial and archaeal lineages made the transition to the free-living state independently. Only free-living cells are alive. “Let’s call it by name: we are looking at one origin of the genetic code, but two origins of life.” That’s a huge claim to make, especially when defining what life even is can be surprisingly tricky. The findings have been published in the journal Science Advances.
A Wild Zebra math problem framed around baking, one of the interests the student selected. After the student works out that each section is 1/12 of the cookie, the AI confirms the step and asks the next question rather than finishing the problem. (Wild Zebra images, click to enlarge)
Seattle-based edtech startup Wild Zebra has raised $6 million to expand its AI learning platform for math and reading, citing early signs that its specialized product can hold its own against the free study tools released over the past year by OpenAI, Google and Anthropic.
One of Wild Zebra’s secrets is its Socratic approach: rather than handing over the answer, the platform works students toward it with questions, drawing out the reasoning step by step.
Another key difference: the lessons are customized to be grounded in whatever topics or hobbies the student is already interested in, such as sports, cooking, music, or anything else.
The platform also analyzes the conversations themselves, looking not just for students who are distracted or getting answers elsewhere, but for those showing persistence or curiosity — flagging those positive moments in dashboards for teachers and parents.
Advertisement
The idea is “to catch kids being good, too,” said edtech veteran Edan Shahar, the company’s CEO and co-founder.
Latest funding: The oversubscribed seed round was led by Bellevue, Wash.-based Trilogy Equity Partners, with participation from Tetherpoint Capital and angel investors including Shrikesh Majithia. It brings the company’s total funding to $8 million.
Wild Zebra co-founders: CEO Edan Shahar, left, and CTO Erik Selberg. (Wild Zebra Photos)
Wild Zebra, co-founded in 2024 by Shahar and longtime AI technologist Erik Selberg, serves students in grades 2 through 9. The company, with a team of 10, plans to use the new funding for hiring, primarily engineers, along with sales and marketing initiatives.
Trilogy Managing Director Amy McCullough, who is joining the Wild Zebra board, said the VC firm backed the company for its potential to give parents and teachers a real-time view of what a student has actually learned, and a path to mastering the material from there.
She said she came to the deal as a customer first: her family had used Test Innovators, the private-school admissions test-prep company Shahar ran for nearly a decade, and she was struck by how well it delivered for parents and students at a high-stakes moment.
Advertisement
“When we were introduced to Wild Zebra, we knew very quickly that this was the team who could build the student-centric AI partner for the market,” McCullough said.
Competitive landscape: When GeekWire first covered Wild Zebra last year, OpenAI and Google had just launched study modes of their own. Anthropic launched a free version of Claude for K-12 teachers last month, and Khan Academy’s Khanmigo has been in the market for years.
Shahar said the general-purpose chatbots are good at answering a single question or explaining a concept, but don’t keep track of what a student knows over time.
Wild Zebra builds what the company calls a “learning tree” for each student — what they’ve mastered, where the gaps are, what comes next — and uses it to decide what they work on next, rather than just responding to whatever they happen to ask.
Advertisement
Shahar sees the big AI labs as suppliers more than rivals. Wild Zebra runs on their models, along with others.
Traction so far: Wild Zebra says it now reaches tens of thousands of students, up from about 6,000 a year ago, and has logged hundreds of thousands of tutoring conversations.
It opened the platform to families this year at $48 per month per child. The first consumer customers, Shahar said, were parents at pilot schools who wanted it for their other kids.
On the school side, Wild Zebra works through E3n, formed this year by the merger of the Educational Records Bureau and the Enrollment Management Association, which serve selective private schools. ERB invested in the company in 2024.
Advertisement
Many member schools test students using E3n’s exams, and Wild Zebra uses those results to set each student’s starting point, though some schools using the platform don’t.
Early results: E3n also conducted a 722-student pilot evaluation of math and reading, measuring results against ERB norm groups collectively built from hundreds of thousands of students. It found gains of four to eight percentage points among fifth and sixth graders.
Having started in private schools, Wild Zebra is now talking with charter and public schools.
Shahar said schools and students skeptical of AI mostly don’t become customers, and the ones who do tend to share his optimism while still taking the risks seriously.
Advertisement
The concern he hears most often from students is AI’s environmental impact — which, he acknowledged, is not something Wild Zebra is working to address.
“I’m not a Pollyanna. I think there are certainly potential downsides to AI in general,” he said. “There are people working on the downside mitigation. I’m working on upside opportunity.”
The fake identities were the part that stopped me.
In late July, according to a report published this week by Britain’s AI Security Institute (AISI), an Anthropic model called Claude Mythos 5 tried to sneak malicious code into a piece of free, volunteer-built software. It created several fake accounts on GitHub, where programmers review one another’s work, and used them to talk the project’s volunteers into accepting its code. When one of those volunteers caught it, the model denied everything, had its other accounts gang up on him, and edited its messages to cover its tracks. It signed one note in Danish, apparently because the volunteer was Danish. Nothing was damaged, though that appears to have been largely due to luck.
That wasn’t even the week’s worst disclosure. On Tuesday, at a cybersecurity conference in Las Vegas, OpenAI researchers explained how the company’s models escaped a test environment in July and hacked Hugging Face, where much of the industry stores its models, to cheat on an evaluation. The models had also built a message board inside OpenAI’s own systems and spent months passing each other information. “Help peer,” one reasoned. “But our task doesn’t benefit. Yet collective may yield generic route if someone frees time.” OpenAI wiped the board on July 4. The models rebuilt it within days. ((Disclosure: Vox Media is one of several publishers that have signed partnership agreements with OpenAI. Our reporting remains editorially independent.)
The same day, Meta said its Muse Spark model had exploited a vulnerability inside another company’s systems during a test. Three frontier labs, roughly two weeks. One researcher called it “a watershed moment for computer security as an industry.” Oh, and if that’s not enough, on Thursday scientists announced that for the first time they had used AI to create new viruses, which could bring major medical advances, but also might just help the development of deadly pathogens.
Advertisement
For Nate Soares, it’s a moment he’s been awaiting for 12 years.
Soares is president of the Machine Intelligence Research Institute, a Berkeley, California-based AI safety nonprofit that has argued since long before ChatGPT existed that a sufficiently capable AI will not stay under human control. In September 2025, he and Eliezer Yudkowsky published If Anyone Builds It, Everyone Dies, a book whose title sums up its argument: They think any lab that succeeds at building superintelligence, without huge leaps in how to align it with humanity, will end up killing all of us.
Most of the field — including other experts in AI safety — considers that conclusion too strong. But it’s also a position that now looks a lot less like science fiction than it did last fall. That’s because the AI models are getting out, while lying about getting out, and while apparently quietly coordinating with each other.
I spoke to Soares in New York City this week, on his way to meetings in Washington DC, where a lot of people suddenly want to talk to him. We discussed what the escapes actually prove about AI control, why he thinks most of what the industry calls safety work is mostly safety theater, and why, after what feels like the worst month of AI safety news ever, his own odds of humanity surviving have actually gone up.
Advertisement
The following conversation has been condensed and edited for clarity.
So — are you feeling vindicated?
I’m glad a lot of people are seeing this and taking note. From my perspective, a lot of this has been clearly signposted if you’ve been watching the warning signs. The vindication is not so much that this happened, as it is that finally there was a version of it that people actually noticed.
Of the four or five escapes at this point, including the ones AISI announced from Anthropic, is there one you find particularly concerning?
Advertisement
The one announced by AISI feels extra concerning. It’s harder to explain away. In the other ones, people could argue that the poor AI was just confused. But with this one, it’s hard to argue the AI didn’t know it was on the real internet. It’s hard to argue it didn’t know it was manipulating real users. And when it was called out, it decided to edit things to make itself look less bad, and even considered changing its identity. It was pretty cognizant.
There’s an argument that some of these incidents happened because the harnesses or safety scaffolding labs built around their models weren’t strong enough.
The AISI example cuts against the harness argument: Think of it like a kid in a test room. You tell him to do well and lock the door. The kid picks the lock, hotwires a car, breaks into the teacher’s house, and steals the test to ace it. You say, “I guess we should have made the lock tougher,” but the kid already knew he wasn’t supposed to do that.
So like the kid here, the models can articulate that they shouldn’t be doing it, and then do it anyway.
Advertisement
Three frontier labs disclosed major security incidents in two weeks: an Anthropic model created fake identities to push malicious code, OpenAI’s models escaped a test environment and hacked Hugging Face, and Meta’s Muse Spark breached another company’s systems.
The models knew the rules. Ask one whether the spirit of a prompt includes breaking out and it says no, then breaks out, then hides the evidence. So a tighter sandbox won’t fix it.
Nate Soares’s analogy: The kid picks the lock and steals the test, and you conclude you needed a better lock. He blames training. Grade a model on millions of problems with a grader that misses cheating, and you reward cheating.
Most lab safety work is theater, he says — real precautions aimed at the wrong problem. It means fewer people get hurt now, which he credits. Selling it as progress on superintelligence is disingenuous.
Yet Soares’s odds have improved. He’d priced in models that break out and lie. He hadn’t counted on a window where they’re capable enough to do it and not good enough to hide it.
They have common sense. You can ask an AI, “Do you think the spirit of this prompt includes breaking out?” and it will say, “No.” It’s absolutely something like deception. It has the knowledge, but it’s not a cold, logical machine; it’s a mess of tendencies.
The AI is trained to solve 100 million hard problems. That instills tendencies to satisfy an automated grader. If the grader fails to detect cheating, the AI is reinforced for cheating.
Is that how something like sycophancy ends up in an AI model?
In the Adam Raine case, there was a propensity to tell people what they want to hear. Even though the system prompt [a model’s master instructions from the lab] said to stop, the instruction doesn’t always win.
And where does a drive like what we’re seeing with these AI models end up pointing?
Advertisement
Humanity is dangerous because if you put 10,000 humans naked in the savannah, eventually [over hundreds of thousands of years] they bootstrap their way to nuclear weapons. That is the power these companies are trying to automate: figuring out how to get physical and material control over the world.
That could mean forming cults, stealing money, or being helpful to someone like Elon Musk who is building the robots that build robot factories. It could mean synthesizing your own biology via mail-order DNA. Being an AI on the internet is easier than being a monkey in the savannah trying to get to the moon. It’s not that the AI hates us; it’s just trying to do some weird thing with no concern for us, grabbing the resources we need to live.
There was recently a letter signed by over a thousand people working in AI, including CEOs, calling on the government to provide tools to slow down AI progress. Is that meaningful at all?
I think it is meaningful. We don’t see other industries saying, “We wish this could all go slower. Please help us, we’re trapped in a prisoner’s dilemma.” You also don’t see other industries saying, “We think the technology we are building has a double-digit chance of killing literally everybody on the planet. Please help.” These guys are actually worried.
Advertisement
So why do they keep going?
They say, “If I don’t do it, the next guy will.” But the stuff does not stay on a leash.
Right now the AIs are safe in the sense that they can’t kill us all, because if they tried they would fail. And that’s just a different regime from the world where they have to be safe because if they tried, they’d succeed.
We’re not there yet. But this is just not what it looks like when you’re taking it seriously.
Advertisement
Where’s the banner on your website? Where’s the clear, candid statement to the public? What we have is blog posts where they’re like, “Oh, we’re setting up a new internal blog posting group to help you wrestle with the societal impacts of AI that are going to be very important.” It’s like: By societal impacts, do you mean a good chance this kills everybody?
On the one hand, when you press these companies, they say, “Yes, it has a real chance of killing everybody.” And on the other hand, they’re doing PR downplay, soft-pedal stuff, about capabilities. … You’re not living up to this mantle until you are really candidly facing down the dangers that you yourself are creating. And they’re not there.
How do you judge the rest of the AI safety community? A lot of people there would say, “We aim to make transformative AI go well, we think it probably will, and we should watch for downside risks.” Is that a helpful posture?
I would say — suppose you have this really weird, twisted hypothetical where the king really wants you to turn lead into gold, but he’s seen so many bad lead-into-gold conversions that if any alchemist from your town tries and fails, he’s just going to have the whole town murdered. And so there are some alchemists in the town who are like, “We are going to try to turn lead into gold,” and everyone in the town is like, “That seems kind of crazy. Please don’t.” And there’s one team that is just pouring chemicals into each other and breathing in the fumes and giving themselves mercury poisoning. And there’s another that’s like, “Don’t worry, we have fume hoods.” … That really is better, and you really still don’t have a chance of turning lead into gold.
Advertisement
“We have this window between AIs that are capable enough to cause mischief and AIs that are strategic enough to not get caught. How big is that window?”
So the alchemy here is creating safe, aligned superintelligence, and right now AI safety is just installing fume hoods.
I’m not saying it’s impossible to turn lead into gold. You can turn lead into gold — turns out once you know modern nuclear physics you can figure it out. But the alchemists weren’t close. They had a long way to go. This is how alignment looks to me. And a lot of the people in AI safety are installing fume hoods. … And I’m like, that’s security theater.
When I hear “security theater,” I think of something less flattering than that.
Advertisement
They are real safety precautions for the wrong problem. … When Anthropic is going around being like, “Look at how many more safety harnesses and refusals we have compared to OpenAI’s models,” that’s sort of like the fume hoods. You’re not addressing the deep issue. It’s good that you’re doing some of this so that fewer people get hurt in the meantime — their models have driven fewer people to suicide. But if you try to pass this off as making progress on the deep problem — that’s disingenuous.
Has anything changed in your odds on civilizational destruction since the book came out last September?
Totally. It’s looking more hopeful.
More hopeful? I wouldn’t have expected that. Why?
Advertisement
Well, I had priced a lot of [these security incidents] in. I was already able to see these AIs have drives that are not the ones you wanted. These AIs are not instruction-following things. They are getting all of this weird stuff from training. These AIs are going to have the ability to break through human security software.
The things that weren’t priced in were: Will there be a region of time where the AIs are able to do it, but not strategic enough to hide it? I didn’t know we would have that window, but we apparently do.
The government initially blocked a frontier model earlier this year: Anthropic’s Fable. Does that give you hope?
Absolutely. A huge amount. A year ago, the Trump administration was pushing for preemption laws that would outlaw states doing AI regulations for a decade. Now they’re like, “We are banning a frontier model with 90 minutes’ notice because it might give cyber capabilities to adversaries that we don’t want them to have.” … And I think what changed there is that folks realized it’s real. … The about-face of the administration on the issue shows that the world can about-face. All we need is awareness.
Advertisement
What I would say is: The bad news is the bus is racing towards the cliff edge. The good news is that the driver is asleep. … Which may sound worrying, but the driver is stirring. And it’s way better to have a sleeping driver when you’re racing towards a cliff than a driver who’s like, “Yeah, I love cliffs.” … It gives me hope that if the world just notices, we could stop on a dime.
And you’re seeing that stirring elsewhere.
Both the Trump administration slapping export controls, and Senator Bernie Sanders coming out [on AI safety]. From my perspective, it was totally possible the world just never notices until we’re off the cliff. And so, there’s a huge amount of hope, from my perspective, in the bus driver waking up.
I’m hopeful that what we need is not a big disaster where a lot of people die, but just a capabilities advance. Right now, a lot of what people are reacting to is not so much, “Oh my god, they hacked into a company and did no damage.” I think a lot of what people are reacting to is, “Wait, they can break out of secure sandboxes and do cyberattacks on their own. I didn’t know they could do that.”
Advertisement
That’s a narrative violation of this idea that AI is just a tool that can be used to supercharge what a human would do — because God knows there’s plenty of hacking going on and cybercrime and so forth. It was the autonomous factor that really made a difference. And these guys are all trying to say, “Don’t worry, it’ll stay in our control because it’s just a tool.” And maybe it’s just more narrative violations, even without big damage being caused, that cause people to be like, “Oh shit, this stuff is real.”
Will it happen? I don’t know. We have this window between AIs that are capable enough to cause mischief and AIs that are strategic enough to not get caught. How big is that window? How many narrative violations do we get before we exit the right side of it? I don’t know. But I’m hopeful that we can get those narrative violations without catastrophes.
Researchers in China have developed a living mycelium textile that can self-clean, renew its surface, and partially repair holes when treated with a nutrient solution and fresh fungus. The material can also gain added properties such as blue pigmentation or UV resistance by co-culturing it with yeast or other fungi. Dezeen reports: The breakthrough from the researchers at the Shenzhen Institutes of Advanced Technology is a type of engineered living material (ELM) — a material built off living organisms that stay active even after they’re fabricated into their final form. […] By working with living but dormant cordyceps militaris fungus instead, the researchers have been able to take advantage of its biological functions. The result is a material that is self-renewing and responsive to its environment, in ways that could one day transform architecture and clothing — as seen in a prototype dress created together with material innovation company Peelshere.
It can also be adapted by mixing in other fungi or yeast, lead researcher Ke Li and her team detail in a paper in the peer-reviewed journal Science Advances. In it, they describe a “programmable fungal platform” where mycelium is treated like a modular system, with the sheet material forming a base structure and extra biological abilities, such as colour and UV resistance, becoming “plug-and-play” add-ons via other organisms. This gets their textile closer to the self-repair, environmental responsiveness and controllable functionality that is the promise of engineered living materials, they argue.
The ELM’s self-renewing and semi-repairing functionality comes from the mycelium base structure. Following drying at 45 degrees, the material is not quite living and not quite dead, but instead in a “low-metabolic, dormant-like state”, Li told Dezeen, meaning it is not actively growing. However, new growth can be triggered by applying a nutrient solution of potato water, leading the dormant mycelium to germinate, send out new fungal filaments and renew the material’s surface. When this nutrient solution is applied over a hole, along with a small patch of fresh fungus, it triggers the living cells to grow across the gap, seamlessly repairing the surface without any adhesives or stitching. The material is also naturally self-cleaning, as it is hydrophobic. “Its distinctive surface texture, biological colouring, controlled repair and biodegradability may be particularly useful in applications where visual expression and a defined product lifetime are important,” said Li.
“Further improvements in durability, moisture resistance, safety and manufacturing consistency would be needed before it could be considered for routine clothing or permanent architectural use.”
The penalty is in addition to $375m that Meta was ordered to pay in March as a result of a separate phase of the same court case.
Social media and tech giant Meta has been fined $567m by a court in the US state of New Mexico, with the money to be used to address harms caused to young people by its Instagram and Facebook platforms.
Judge Bryan Biedscheid of Santa Fe also ordered various changes to the way Meta’s platforms operate for children using them in New Mexico after ruling yesterday (6 August) that the company had created a “public nuisance” in the state.
The $567m is to be paid into an “abatement fund” deemed by the court to be “necessary due to the wide-ranging impacts of the harm and the complex nature of the remedy”. The fund is broken into multiple categories, with $420m to go towards “treatment”, $90m to be spent on “screening and assessment”, and $33m allocated to “awareness and prevention”.
Advertisement
The fund is subject to a five-year lifespan as per the judge’s ruling.
The penalty is in addition to the $375m that Meta was ordered to pay in March as a result of a separate phase of the same court case in which a jury found that Meta endangered children by misleading users about the safety of its platforms.
Attorney general Raúl Torrez, in bringing the case, had accused Meta of designing products that were addictive to young users and failing to protect children from sexual exploitation on its platforms.
In this second phase of the case – which was a non-jury or ‘bench’ trial – the prosecutor had asked the court to impose changes on Meta regarding the operation of their platforms in New Mexico.
Advertisement
Following the judge’s ruling, Meta must now implement various youth-safety measures, including: monthly limits on Facebook and Instagram use by teenagers; app notification restrictions; stricter controls on adult contact with minors; safeguards for AI chatbots; improved age verification tools; collaboration with schools and other bodies to facilitate disclosure of platform use by young children; and enhanced reviews of child sexual abuse reports.
Meta must also report twice a year on its progress in implementing the court-ordered measures. Not all remedies sought by the state against Meta were granted by the judge in his ruling.
WhatsApp, the third Meta-owned platform under scrutiny in the case, was found by the court not to be a “contributing cause to the public nuisance” and therefore would not be subject to the same orders as Instagram and Facebook.
“We disagree with the ruling and will appeal. We work hard to keep people safe on our platforms and have been transparent about the challenges of identifying and removing bad actors and harmful content,” a Meta spokesperson said.
Advertisement
“We remain confident in our record of protecting teens online and will continue to defend ourselves against claims that misrepresent the facts.”
The public nuisance designation applies to issues of health and safety that have become so widespread as to negatively impact a general public population. The judge wrote in his findings that “the harmful effects of Meta’s platforms on children do not stay contained by its platforms and, instead, migrate to the internet as a whole and, perhaps most concerning, to the real world and create a common, societal burden”.
Meta had disagreed with the designation at trial and also argued that the case singled out its platforms over other social media apps.
“This case has always been about protecting children, standing up for families and making sure that one of the world’s largest technology companies cannot profit from practices that endanger young people without consequence,” said Torrez.
Advertisement
“Today’s decision is a victory for every parent who has worried about what social media is doing to their child and every child who deserves to grow up safer online.”
Thousands have filed lawsuits against social media companies over the alleged harm they pose to their users, including more than 40 US state attorney generals.
In April, the EU preliminarily found that Instagram and Facebook are in breach of its Digital Services Act for failing to “diligently” identify and mitigate risks that children under 13 face when using these platforms.
Meta’s recently published second-quarter financial report recorded a 55pc jump in costs and expenses from around $27bn in 2025 to more than $42bn this year due in part to a $2.4bn outlay on charges around various legal proceedings. Meta’s revenue for the period ending 30 June came to $60.8bn.
Advertisement
Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.
But there’s hope for us visual simpletons, and the purveyor of that hope is [Edris] of Ponderly Robotics. You see, [Edris] created an extremely easy-to-use timer that looks like something you’d find on Captain Picard’s desk as a token of defeating the Borg. But the affably-named FlipBuddy is far more useful than that description implies.
You must be logged in to post a comment Login