A unique BMX frame tests the limits of an electric bike’s capabilities. Sam Barker and Tom Stanton, two friends who enjoy tackling weekend projects, transformed a standard BMX frame into a speed machine. What they created is a beast that feels more like a dragster than a pedal assist bike, a tiny BMX frame with enough power to raise the front wheel and outperform more standard e-bikes.

The project began by simply plugging a 20-inch BMX wheel equipped with a 3kW hub motor into an existing bike in the garage, powered by a 72 volt, 20 amp-hour battery. Initially, some quick spins around the block revealed some potential acceleration, but pushing the throttle wide was all it took to send the back wheel spinning into the air and the front end flying off completely due to the narrow wheelbase.

Sale

VOLPAM Electric Scooter, 350W/500W Powerful Motor, 8.5″/10″ Solid Tires, 19Mph, 28/23/21/16 Miles Max…

• 350W Motor: Our electric scooter equipped with improved motor, you can experience a smoother, more powerful ride with our upgraded 350w brushless…
• Smart App: The electric scooter is equipped with an intelligent control system offering features like 4 speeds adjustment, cruise control, lighting…
• Portable & Compact: The lightweight e-scooter fits neatly in your car trunk or tight spaces , effortlessly foldable in seconds, its small size making…

They knew they needed to solve the handling issues, so they brought out the welder and cut the frame, adding a meter of 12mm mild steel tubing to give it some significant stability. They added some reinforced brackets to make the whole thing feel sturdy, almost like scaffolding. The longer frame kept the front wheel on the ground during those violent launches, and it also allowed them to finally tuck the battery inside the bike. They also changed the handlebars to a unique design to give them better control, and they adjusted the pedals to make them exactly perfect.

Stopping was accomplished with a borrowed rim brake caliper from a penny-farthing cycle. To be honest, it sort of works, but don’t expect to be able to stop on a dime soon. The machine overheats a lot under severe braking, and it won’t slow you down enough to be safe at the speeds it can reach. They did identify a workaround by shunting the motor wires to enable regenerative braking, but the main brakes remain marginal.

When they eventually took it out for a ride, they discovered the bike’s personality: it creates so much acceleration that the back wheel starts spinning before the rubber really has a chance to grip, and then you’re past 50 or 60 kph, feeling the speed wobble from the sharp steering angle. It’s not particularly stable, but it’s far superior to the original short version. While there is no fairing to protect you, and the low seat causes you to bounce around on the pegs, you must agree it’s a lot of fun.
[Source]

One way is to buy an optional $25 hardwire kit from Vantrue, which will connect the device directly to your fuse box. Note, however, that the camera draws about 25 to 35 milliamps while in parking mode (about 3 to 4 watts on a home outlet), and this will drain your car’s battery if hardwired. This isn’t a troublesome power draw, and should be fine for days or maybe weeks if your battery’s healthy.

But in general, I’d rather not drain my car battery while it’s parked. I’d instead nix the hardwire kit and invest in a good power bank. Even a $48 portable bank from Anker should be enough for about a week’s continuous operation while parked, without ever worrying about battery drain on your car.

Limitations and Faults

Vantrue app via Matthew Korfhage

For such a dinky device, the Vantrue does nonetheless offer a real-time screen. It’s doubtful you’ll use it to scroll through video, however. It’s most useful for verifying the framing on the cam when you stick it to your windshield. The screen is a bit busy with information, and lower resolution than the actual video.

If you’re reviewing footage, you’ll be doing so on your phone using the app. But the screen is still useful for adjusting settings manually, should you desperately need to—though controlling the phone via the on-camera interface will be a little fiddly and irritating. Again, you’re better off just connecting the camera to your phone to toggle settings.

There are, in fact, settings aplenty. Through the app, you’re able to track mileage, toggle GPS tracking, set the frame rate and resolution of the camera, and set whether you want the camera to use high-dynamic range settings or PlatePix. The former will be most useful when light is dim.

It’s also a simple, single-camera device. Vantrue has more elaborate (and more expensive, and larger) multi-camera options with similar camera specs that I’m in the process of testing, including a Nexus 4 Pro with a lower-resolution front-cabin and rear cam. A 4-channel N5 adds an extra rear-cabin channel for a wild amount of camera coverage, but at the expense of some image resolution.

PlatePix does indeed help in capturing license plates, but it will do so at the expense of contrast on the rest of the image. This is a compromise that’ll matter most at night. Which is to say, you’re choosing between marginal license plate capture on a dark image, or worse, license plate capture on crisper night footage.

On March 8, 2026, in honour of International Women’s Day, a new edition of SheBuilds on Lovable invited builders from around the world to a 24-hour global, powered by Anthropic. On top of that, participants in the SheBuilds campaign receive $100 in Anthropic API credits and $250 in Stripe fee credits to kickstart their builds and remove early-stage barriers to creation.

This isn’t just another hackathon; it’s a campaign that blends community, technology, and real output. It asks a simple question of the tech industry and its marketers: do we want participation, or just talk about it?

If I’m honest, I have mixed feelings. We need a stronger community and louder voices for women in tech not just on a specific international day, but every day. I’d like to believe we are beyond the stereotype of what jobs women can or cannot do, and what is left for men. Beyond any single campaign, we should advocate for women as builders in the tech world every day.

Still, every time I see a campaign like this, it sparks some hope inside me. It reminds me that we are not reduced to specific skills on a specific day, and that progress is possible.

Returning to SheBuilds, which is rooted in a longer tradition: the 48-hour virtual buildathons Lovable has run for women founders and builders, where ideas become real applications over an intense weekend of collaboration.

In October 2025, 60 women joined a remote, AI-enabled sprint to design, prototype, and launch working products in 48 hours, and the resonant element wasn’t speed, it was agency. 

In 2026, the campaign took this idea further. Instead of being constrained to a single track, it anchored itself to a global cultural moment where women’s participation in tech isn’t just acknowledged but activated, with free build credits, peer connection, live sessions, and an open invitation to anyone who wants to make rather than debate what innovation should look like. 

We have been talking in tech about inclusion and diversity for years. Many initiatives generate reports, webinars, and packs of aspirational slides. SheBuilds does something different: it creates the conditions for output. In the world of software, execution matters. A founder funded or a product launched is a literal piece of industry change, unlike any promise that never materialises.

SheBuilds lowers barriers in two ways. First, it reshapes access to execution. Participants don’t need traditional engineering skills; they need a laptop, an idea, and the willingness to build it. Second, by connecting creation to a broader cultural moment, International Women’s Day, it reframes participation as both a personal achievement and a public conversation about who gets to shape tech. 

There’s good evidence that this model matters. Across the Lovable community, builders report that the ritual of building, not discussing, changes how they see themselves in technology. For many, shipping a live prototype in 48 hours meant moving from being stuck in ideas to being recognized as a doer; more than that, it positioned them within a network that values rapid iteration and real product agency. 

From the industry’s perspective, this matters because it accelerates user engagement in a genuinely participatory way. Rather than positioning a platform as something to use someday, the SheBuilds campaign positions it as something you use now. When marketers tie campaigns to output, not just signup counts, they create moments of meaningful interaction. That’s both a brand story and a user experience strategy.

Lovable’s move to embrace International Women’s Day isn’t a gesture; it’s a strategic campaign. By hosting SheBuilds events tied to a global calendar, the company amplifies its mission, and invites participation on a broader scale. It embodies a shift in how tech platforms engage builders: merging community activation with cultural moments, and showing that building isn’t reserved for a subset with coding degrees. 

If the tech industry wants to look past fancy words like diversity and equity, it needs campaigns that exchange words for workshops, panels for platforms, and theories for tools. That’s the real invitation SheBuilds extends to its participants, and the lesson it offers marketers who want to tell meaningful stories about making things.

The tech world still doesn’t reflect the people it serves. Women make up far less than half of technical roles and leadership positions, and the gap shows up in boardrooms, engineering teams, and the very products we use every day. Representation isn’t a nice-to-have statistic; it shapes what technology looks like and who gets to steer it. 

From where I stand, SheBuilds on Lovable isn’t just another event. It is a call to action that meets a long-standing need, a space where women are not only welcomed but empowered to create, ship, and be seen. It turns invisible potential into visible impact, inviting women to bring their solutions into the world rather than waiting for permission to do so. 

When women build, the whole industry becomes richer in ideas, perspective, and purpose. Moments like this, tied to real creative work, are not symbolic gestures. They are steps toward a tech ecosystem where women aren’t exceptions but equals. 

Let the world know the builder you are! Celebrate your voice here! Isn’t this lovable?

OPPO has started rolling out its February 2026 update for ColorOS 16, bringing the PopOut feature to more devices beyond the recently launched Reno 15 series. The update focuses on creative tools and productivity improvements, as well as refinements to AI-powered image editing.

PopOut Feature Expands to More OPPO Devices

Originally introduced with the OPPO Reno15 series, the PopOut feature is now available on a wider range of OPPO smartphones, including:

The PopOut feature creates a 3D-like depth effect by lifting the main subject forward from the background. The result is a more dynamic and eye-catching image that appears to extend beyond the frame. With this update, OPPO is making the creative tool accessible to more users across its premium and upper mid-range lineup.

Productivity and AI Enhancements

Alongside PopOut, the February update also introduces several system-level improvements.

PowerPoint Translation now allows users to translate presentations into multiple languages while preserving the original formatting. It also supports side-by-side comparison and easy export.

OPPO has further refined its AI Eraser tool, aiming for cleaner edits with fewer visible traces after object removal. The Reflection Removal feature has also been upgraded to reduce glare in document photos and portraits, including images where subjects are wearing glasses.

The Private Safe feature has also received performance improvements, with faster file transfers and a new option to add files directly under the “Other files” category.

Rollout Timeline

The ColorOS 16 February update is being rolled out in phases between February 6 and February 28, 2026. Users on eligible devices can check for the update via Settings once it becomes available in their region.

On Tuesday, Anthropic published tools that let Claude read, analyze and translate legacy COBOL into modern languages like Java and Python. By the end of the trading day, investors had wiped roughly $40 billion from IBM’s market cap — the company’s biggest single-day drop in 25 years — pricing the announcement as an existential threat to IBM’s mainframe business.

The reaction was swift. It was also built on a fundamental misreading of why enterprises run mainframes in the first place.

IBM’s COBOL is 66 years old. It was designed in 1959, runs on IBM mainframes, and continues to power transaction processing systems with an estimated 250 billion lines of COBOL in active production, according to the Open Mainframe Project.

The engineers who wrote it are retiring; the ones replacing them largely cannot read it. For decades, that skills gap has been one of enterprise IT’s most expensive unsolved problems — and one IBM has been working to fix with AI since at least 2023, when it launched watsonx Code Assistant for Z to help migrate COBOL to modern Java.

Claude Code, Anthropic says, can now analyze entire codebases, map hidden dependencies, and generate working translations of code that most engineers today cannot read. For enterprises running COBOL on distributed platforms — Windows, Linux and other non-mainframe environments — that capability is genuinely useful and increasingly practical.

The actual barrier was never technical

“Modernizing COBOL has been a technically solved problem for a while,” Matt Brasier, analyst at Gartner, told VentureBeat. “The real problem is that the costs of modernization are high and the ROI is low.”

Amazon and Google have been offering AI-powered COBOL migration tools for years. AWS Transform and a comparable Google Cloud Platform service both targeted the same problem: reducing friction for customers looking to move mainframe workloads to the cloud.

“This is basically one more source of competition,” Raj Joshi, senior vice president at Moody’s Ratings, told VentureBeat. “IBM has always lived in a very competitive domain. On the margin, this thing is basically negative, no question about that. There’s one more powerful competitor. But IBM has coexisted with these threats.”

Steve McDowell, chief analyst at NAND Research, cuts to the structural argument: “Applications don’t run on mainframes because they’re written in COBOL,” he said. “They run on mainframes because mainframes deliver a class of determinism, scalable compute and reliability that general purpose servers can’t match.”

The issue runs deeper than market positioning. “GenAI tools are helpful, but their non-deterministic nature means the resulting code is not consistent — the same operation will be implemented in different ways in different parts of the code,” Brasier said. “Leading tools combine deterministic and non-deterministic approaches. None of this solves the ROI problem, though.”

What COBOL translation leaves unsolved

“Translating COBOL is the easy part,” IBM communications director Steven Tomasco told VentureBeat. “The real work is data architecture redesign, runtime replacement, transaction processing integrity, and hardware-accelerated performance built over decades of tight software and hardware coupling. That is the problem IBM has spent decades learning to solve, and AI is the most powerful tool we have ever had to do it.”

According to IBM, Royal Bank of Canada, the National Organization for Social Insurance and ANZ Bank have all used watsonx Code Assistant for Z to accelerate modernization of COBOL code without moving off IBM Z.

That does not mean Anthropic has no competitive foothold. For enterprises running COBOL outside the mainframe — on distributed systems, Windows and Linux environments — Claude Code enters a space where IBM’s vertical integration is less of an advantage. “IBM understands mainframe technology at a level that others can’t match. If I’m only looking at COBOL, I’m using IBM’s watsonx,” McDowell said. “Anthropic, however, has a broader footprint within a lot of development teams, where a single vendor makes it worthwhile.”

What enterprise buyers should actually do

Senior data and infrastructure engineers will spend the next few weeks fielding questions from executives who saw the headlines and assumed the hard problem just got solved. It did not.

“It’s COBOL, but there are numerous applications tied to it,” Joshi said. “It’s not like you transform millions of lines and somehow you are ready to go to cloud. It’s a massive risk assessment, dependencies and all those things.”

The more useful question for buyers is whether this week’s noise creates an opening. Braiser thinks it does.

“They should use the resulting board-level and shareholder discussions to review postponed modernization initiatives and see if any of them now have ROI,” Brasier said.

McDowell was blunt on the competitive question. “Will Anthropic take business from IBM’s tool? Yes, of course,” he said. “But I’d be surprised if that tool was making significant revenue for IBM.”

Chirag Mehta, analyst at Constellation Research, cautioned that IT leaders should not react emotionally or rewrite strategy overnight.

“Treat this as a reason to run a small, bounded pilot to measure outcomes, not as a reason to rip and replace vendors,” Mehta told VentureBeat.

Mehta suggests that enterprises pick one well-scoped application slice or workflow with clear inputs and outputs, and evaluate approaches apples-to-apples: quality of dependency mapping, quality of recovered business logic documentation, test coverage and equivalence checks, performance and reliability regressions.

In Mehta’s view, the bigger reminder is that modernization is more than converting code. The hard parts are extracting institutional knowledge, reworking processes and controls, change management, and containing operational risk in systems that cannot break. AI can compress the “analysis and translation” work, but it does not eliminate the governance and accountability burden.

“The teams that win will treat AI as an accelerator inside a disciplined modernization program, with measurable checkpoints and risk guardrails, not as a magic conversion button,” Mehta said.

“El Mencho” is dead.

This weekend, Mexican Army Special Forces killed Nemesio Rubén “El Mencho” Oseguera Cervantes, the head of the powerful Jalisco New Generation Cartel (CJNG) in Tapalpa, Jalisco, Mexico. Following confirmation of El Mencho’s death by federal authorities, experts anticipate a profound reconfiguration of the global drug trafficking landscape, a scenario that could lead to a new and dangerous wave of violence.

The focus will turn to the CJNG’s mechanisms of control, intimidation, financing, and recruitment that granted the cartel unprecedented operational capacity. Much of its strength stemmed from the weakening of long-standing rivals through the sophisticated use of social media and artificial intelligence, state-of-the-art specialized weaponry, and a flexible internal structure.

The US State Department says that CJNG maintains a presence and contacts in “almost all of Mexico,” the American continent, and countries such as Australia, China, and various Southeast Asian nations. The agency underscores the cartel’s criminal versatility: In addition to fentanyl trafficking, it is involved in extortion, migrant smuggling, oil and mineral theft, and illicit arms trade.

How Was the CJNG Born?

The CJNG traces its roots to the Sinaloa Cartel, led by Joaquín Guzmán Loera, also known as “El Chapo.” Around 2007, this group formed an armed wing in Jalisco under the command of Ignacio “Nacho” Coronel Villarreal. This was the Milenio Cartel, also known as Los Valencia. During this period, Guzmán’s operatives were vying for control of Jalisco territories against Los Zetas, a splinter group of the Gulf Cartel.

In its early years, the CJNG presented itself as “Los Mata Zetas” (The Zeta Killers). According to the BBC, its first documented appearance occurred in September 2011, when it claimed responsibility, through a video circulated on social media, for the discovery of 35 bodies in Boca del Río, a municipality in the state of Veracruz.

By then, the alliance with the Sinaloa Cartel had already fractured after a confrontation with federal forces that culminated in 2010 with the death of Ignacio Coronel. Leadership fell to Oseguera Cervantes, who spearheaded a rapid expansion in methamphetamine production and trafficking.

Cartels Embracing Tech

In less than five years, the CJNG displaced the Knights Templar from southern Michoacán and expelled Los Zetas from northern Jalisco and parts of Zacatecas. After Guzmán Loera’s capture and extradition, the group strengthened its strategy by recruiting financial and chemical specialists to boost the manufacture of synthetic drugs and diversify its income through money-laundering schemes in sectors such as livestock, mining, agriculture, and construction, as well as expanding extortion of small- and medium-size businesses.

The US Drug Enforcement Administration (DEA) notes that the CJNG operates in more than 40 countries and has a financial structure known as Los Cuinis, headed by Abigael González Valencia, Oseguera’s brother-in-law. This network coordinates money-laundering operations through international trade, cryptocurrencies, and links with Asian networks.

Several investigations have documented the use of digital tools for recruitment and fraud. In 2024, Interpol warned that groups like the CJNG were involved in large-scale financial scams supported by AI, natural language models, and cryptocurrencies. It also detected the expansion of human trafficking for forced criminal activity in scam compounds.

A study by El Colegio de México, in collaboration with the Civic AI Lab at Northeastern University in Boston, revealed that TikTok has become a recruitment tool for Mexican cartels, including CJNG. The research identified 100 active accounts linked to illicit organizations and categorized their content as recruitment, border crossings, illegal businesses, prostitution, propaganda, and arms sales. Forty-seven percent of the accounts promoted the recruitment of new members, and 31 percent disseminated propaganda messages. The report highlighted that the CJNG accounted for 54.3 percent of the detected accounts, followed by the Sinaloa Cartel and the Northeast Cartel.

Swiss shoe company On—one of the fastest-growing footwear brands—has done very well for itself, growing into a $3 billion company on the giant puffy soles of its very garish shoes. Today, the company announced three new kicks and the opening of a new production plant in South Korea that will make many, many more—at least, that’s the goal.

Specifically, the company is focusing on the production of its newest, weirdest shoe—a giant soled laceless running shoe with a single-piece toe box made of “hyper-foam” plastics sprayed on by robot arms. The plastics are 40 percent biofoam, and the shoe is made of just eight pieces; On says its minimalist approach saves on the shoe’s carbon footprint.

The LightSpray Cloudmonster 3 Hyper is a switch-up from On’s first shoe that uses its sprayed-on toe box technique, the Cloudboom Strike LS. Along with adding the LightSpray branding to the thing, the primary difference is the switch to the Cloudmonster model, which, in On shoe lore, tends to provide more cushy soles. This new version has 20 percent more foam, for all those runners who like a squishy bounce. They’ll have a limited release on the company’s website and retail stores in North America starting March 5, with a global expansion on April 16.

The new LightSpray shoes weigh 205 grams apiece (less than half a pound), which puts them a little heavier than the 170-gram Cloudboom Strikes. They’re also cheaper at $280 versus the Cloudboom’s heftier $330 price.

While WIRED liked the first iteration of the LightSpray shoe, nobody else, except for a few hand-picked marathon runners, had a chance to use it. Production runs were very limited, and the shoes were prohibitively expensive. Now, On has beefed up production in South Korea, enabling it to expand its overall production of LightSpray shoes 30-fold in 2026.

“On’s strategy to expand the LightSpray technology from racing to running shoes is to make it available to a broader audience beyond elite athletes,” an On representative tells WIRED via email.

Whether laceless, slip-on shoes are going to find their gait in the shoe market isn’t a winning idea yet. While brands like Nike offer slip-ons, the laceless fixation tends to be reserved for one-off shoes like the Back to the Future II-inspired Adapt BB self-lacing shoes. Some serious runners prefer the pursuit of a more natural run offered by slim-soled shoes like barefoot shoes, but On is still doubling down on its doubly big cushions. People seem to like them, as On is also announcing two other Cloudmonster shoes. One is the laced-up Cloudmonster 3, and the other is a regular ol’ Cloudmonster 3 Hyper with laces and a non-LightSprayed toe box. (Ugh, laces. How passé.)

On says it has the further strategic goal of scaling up its global production of the LightSpray shoes, “within the next few years.”

Claude Code has become increasingly popular in the first year since its launch, and especially in recent months, as developers and non-technical users alike flock to AI unicorn Anthropic’s hit coding agent to create full applications and websites in days, on their own, that would’ve taken months and technical teams without. It’s not a stretch to say it helped spur the “vibe coding” boom — using plain English instead of programming languages to write software.

But it’s all been restricted to the desktop Claude Code apps and Terminal command-line interfaces and integrated development environments (IDEs) — until today. Now, Anthropic has added a new mode, Remote Control, that lets users issue commands to Claude Code from their iPhone and Android smartphones — starting with subscribers to Anthropic’s Claude Max ($100-$200 USD monthly) subscription tier.

Anthropic posted on X saying Remote Control will also make its way to Claude Pro ($20 USD monthly) subscribers in the future.

The mobile command center

Announced earlier today by Claude Code Product Manager Noah Zweben, Remote Control is a synchronization layer that bridges local CLI environments with the Claude mobile app and web interface.

The feature allows developers to initiate a complex task in their terminal and maintain full control of it from a phone or tablet, effectively decoupling the AI agent from the physical workstation.

Currently, Remote Control is available as a Research Preview for subscribers on the Claude Max tier. While access for Claude Pro ($20/month) users is expected shortly, the feature remains a high-end tool for power users and is notably absent from Team or Enterprise plans during this initial phase.

To access the feature, users must follow this guide and update to Claude version 2.1.52 and execute the command claude remote-control or use the in-session slash command /rc. Once active, the terminal displays a QR code that, when scanned, opens a responsive, synchronized session in the Claude mobile app.

Less screen time, more IRL time: philosophy of flow

The messaging behind the release centers on the preservation of a developer’s “flow state.”

In his announcement, Zweben framed the update as a lifestyle upgrade rather than just a technical one, encouraging users to “take a walk, see the sun, walk your dog without losing your flow.”

This “Remote Control” is not a cloud-based replacement for local development, but a portal into it. According to official documentation, the core value is that “Claude keeps running on your machine, and you can control the session from the Claude app.”

This ensures that local context—filesystem access, environment variables, and Model Context Protocol (MCP)servers—remains active and reachable even if the user is miles away from their desk.

Architecture, security, and setup

Claude Code Remote Control functions as a secure bridge between your local terminal and Anthropic’s cloud interface, which provides the Anthropic AI models, Opus 4.6 and Sonnet 4.6, that power Claude Code.

When you run the command, your desktop machine initiates an outbound connection to Anthropic’s API for serving the models — meaning you aren’t opening any “inbound” ports or exposing your computer to the open web. Instead, your local machine polls the API for instructions.

When you visit the session URL or use the Claude app, you are essentially using those devices as a “remote window” to view and command the process still running on your computer. Your files and MCP servers never leave your machine; only the chat messages and tool results flow through the encrypted bridge.

To get started, ensure you are on a Pro or Max plan and have authenticated your CLI using the /login command. Simply navigate to your project directory and run claude remote-control to initialize the session. The terminal will then generate a unique session URL and a QR code (toggleable via the spacebar) for your mobile device.

Once you open that link on your phone, tablet, or another browser, the two surfaces stay in perfect sync—allowing you to start a task at your desk and continue it from the couch while maintaining full access to your local filesystem and project configuration.

From brittle community hacks to official solution

Prior to this official release, the developer community went to great lengths to “hack” mobile access into their terminal-based workflows.

Power users frequently relied on a patchwork of third-party tools like Tailscale for secure tunneling, Termius or Termux for mobile SSH access, and Tmux for session persistence.

Some developers even built complex custom WebSocket bridges just to get a responsive mobile UI for their local Claude sessions.

These unofficial solutions, while functional, were often brittle and prone to timeout issues. Remote Control replaces these workarounds with a native streaming connection that requires no port forwarding or complex VPN configurations.

It also includes automatic reconnection logic: if a user’s laptop sleeps or the network drops, the session remains alive in the background and reconnects as soon as the host machine is back online.

The $2.5 billion-dollar agent

The launch of Remote Control serves as an “escalation of force” in what has become a dominant business for Anthropic. As of February 2026, Claude Code has hit a $2.5 billion annualized run rate — a figure that has more than doubled since the start of the year alone.

Claude Code is currently experiencing its “ChatGPT moment,” surging to 29 million daily installs within Visual Studio Code. Its efficiency is no longer theoretical; recent analysis suggests that 4% of all public GitHub commits worldwide are now authored by Claude Code.

By extending this power to mobile, Anthropic is further entrenching its lead in the “agentic” coding space, moving beyond simple autocomplete to a world where the AI acts as an autonomous collaborator.

Future outlook: vibe coding everywhere

The move toward mobile terminal control signals a broader shift in the software market. We are entering an era where AI tools are writing roughly 41% of all code. For developers, this translates to a migration from “line-by-line” typing to “strategic oversight.”

This trend is likely to accelerate as mobile-tethered agents become the norm. The barrier between “idea” and “production” is collapsing, enabling a single developer to manage complex systems that previously required entire DevOps teams. This shift has already rattled the broader tech market; shares of major cybersecurity firms like CrowdStrike and Datadog fell as much as 11% following the launch of Claude Code’s automated security scanning features.

As Claude Code moves from the desk to the pocket, the definition of a “software engineer” is being rewritten. In the coming year, the industry may see a surge in “one-person unicorns”—startups built and maintained almost entirely via mobile agentic commands—marking the end of the manual coding era as we knew it.

Apple made a huge announcements today by revealing that some Mac Mini computers will roll off the line in Houston, Texas, marking the first time this compact desktop gets built on American soil. This move is a significant extension of an existing site in Houston where they have been building high-end AI servers since 2025.

A brand new plant dedicated to producing Mac Minis will effectively double the size of the campus there. Along with it, a separate 20,000-square-foot Advanced Manufacturing Center will open later in 2026. The idea is that everyone, not just Apple employees, but students, workers from suppliers, and so on, will receive hands-on training in the latest and greatest production methods, thanks to a special curriculum developed by Apple in collaboration with some experts from Michigan State University.

Sale

Apple 2024 Mac mini Desktop Computer with M4 chip with 10‑core CPU and 10‑core GPU: Built for Apple…

• SIZE DOWN. POWER UP — The far mightier, way tinier Mac mini desktop computer is five by five inches of pure power. Built for Apple Intelligence…
• LOOKS SMALL. LIVES LARGE — At just five by five inches, Mac mini is designed to fit perfectly next to a monitor and is easy to place just about…
• CONVENIENT CONNECTIONS — Get connected with Thunderbolt, HDMI, and Gigabit Ethernet ports on the back and, for the first time, front-facing USB-C…

This latest development fits into Apple’s overall ambition to invest $600 billion in the US economy, which they announced last year. We’ve already seen some success on this front; for example, last year Apple sourced over 20 billion chips manufactured right here in the good old United States of America from 24 separate facilities spread over 12 states. In 2026, they will rely even more on TSMC’s expanding factory in Arizona for their high-end chips. And it’s not just the chips; GlobalWafers is building a $4 billion plant in Sherman, Texas to make silicon wafers, Amkor is opening a $7 billion facility in Peoria, Arizona to pack chips, and Corning has fully committed their Kentucky operation to producing cover glass for Apple.

Assuming all proceeds as planned, this should result in thousands of good employment in Houston, both from plant expansion and from the training facility. The Mac Mini is already the most cheap desktop available, priced at $599. The Mac Mini accounts for less than 5% of total Mac sales, according to some estimates, thus this is a small but significant step toward producing more Apple goods in the United States.

It appears that all of Apple’s partners are on board, including Foxconn, which has worked with Apple on products dating back to the early iMac days and is running the North Houston site. So production isn’t halting in Asia; this is more of an addition to their product range than a complete move. All of this comes at a particularly fascinating time for discussions about supply chains, tariffs, and so on. Apple claims that the practical benefits of all of this are obvious: people are learning new skills, local economies are benefiting, and the company is able to continue producing even better goods by manufacturing close to home.

Tim Cook described the expansion clearly. “Apple is deeply committed to the future of American manufacturing, and we’re proud to significantly expand our footprint in Houston with the production of Mac mini starting later this year,” he said. “We began shipping advanced AI servers from Houston ahead of schedule, and we’re excited to accelerate that work even further.”