Technology

• Google Cloud adds six new AI agents for data scientists, engineers and more
• Advanced analytics will become more accessible with natural-language AI
• A solid data foundation is just as important, but Google can help you migrate

Google Cloud has launched six new AI agent tools to assist data engineers, data scientists, developers and business users realize even more productivity benefits.

Outlining a, "new era where specialized AI agents work autonomously and cooperatively to unlock insights at a scale and speed," Data Cloud Managing Director Yasmeen Ahmad explained the benefits of a "single, unified, AI-native cloud" over siloed tools when it comes to using AI.

Besides new, specialized AI agents, Google Cloud is also launching a series of APIs, tools, and protocols as well as updates to unify data.

Google Cloud launches even more AI agents

The first agent, destined for data engineers, is designed to automate complex data pipelines by allowing engineers to describe tasks and then autonomously building and executing workflows. A separate Spanner Migration Agent will simplify migrating from legacy databases like MySQL to Spanner, eliminating hours of tedious administrative work.

Data scientists will benefit from an agent that automatically performs exploratory data analysis, data cleaning, feature engineering and ML predictions, offering step-by-step reasoning and collaborative feedback, while business users and analysts will get to use two separate agents designed to answer questions about data and interpret code with visualisations and explanations, meaning that non-technical users can perform advanced analytics.

Finally, Gemini CLI GitHub Actions will automate pull requests, tests, reviews and implementation for developers.

"The true potential of the agentic shift is realized when developers not only use existing agents, but also extend and connect them to their own intelligent systems, creating a broader network," Ahmad explained.

With its new agents, Google Cloud hopes to lower the barrier of entry into advanced data analytics, "eras[ing] the line between operational and analytical worlds."

You might also like

• The next big thing in AI is agents, but is your data ready?
• Check out the best AI tools and best AI writers
• We've listed all of the best productivity tools

• One of Alien: Earth's antagonists isn't based on tech bros like Elon Musk
• The actor behind the character didn't seek inspiration from real-world examples
• He let series creator Noah Hawley's writing do the talking

One of Alien: Earth's stars has denied that specific examples of real-life tech bros inspired the duplicitous character he portrays in the FX TV Original.

Speaking to TechRadar, Samuel Blenkin, who plays Boy Kavalier in the sci-fi horror franchise's first-ever TV project, said he simply relied on how the character had been written.

For the uninitiated: Boy Kavalier is the 20-something CEO and founder of Prodigy Corporation. One of five megacorporations that essentially rule planet Earth in the Alien universe, Prodigy is at the forefront of unlocking human immortality via its Hybrid program – an experimental procedure that transfers the consciousness of a human child into an artificial adult body.

However, not long after Prodigy successfully creates six Hybrids, the USCSS Maginot – a deep-space research vessel owned by Weyland-Yutani, one of Prodigy's rivals and the Alien franchise's most famous multinational – crashes into Prodigy City. Upon discovering that the Maginot was transporting five dangerous alien lifeforms, including one of the franchise's iconic Xenomorphs, to Weyland-Yutani, Kavalier takes ownership of the potentially lethal extra-terrestrials for experimental purposes.

Alien: Earth introduces four new life-threatening organisms to the sci-fi horror franchise's universe (Image credit: FX Networks)

Anyone who's seen an Alien movie – or even a Jurassic Park one – knows that playing with things you don't fully understand is a recipe for disaster. Regardless of the consequences, though, the arrogant and so-called 'boy genius' Kavalier is hell-bent on unearthing the bioweapons' secrets in the Hulu and Disney+ TV Original.

If Kavalier's self-important and rebellious personality seems familiar, it might be that you're reminded of supposed 'tech revolutionaries' who, like Kavalier, claim their technological advancements are for humanity's benefit in spite of concerns about their use.

Need examples? How about the uncanny valley nature of Elon Musk's Tesla Bots, which some observers have likened to the Terminators from the James Cameron-created dystopian sci-fi franchise? What about artificial intelligence (AI) chatbots, such as ChatGPT, that use the OpenAI software co-created by Sam Altman? Or, take a look at Facebook founder and Meta CEO Mark Zuckerberg's all-consuming quest to make the Metaverse a real thing.

No, Alien: Earth's Samuel Blenkin didn't base Boy Kavalier on Elon Musk (Image credit: Getty Images)

For what it's worth, Blenkin rejected – or, rather, strongly sidestepped – the notion that any or all of the above, or any other tech guru, influenced his portrayal of Kavalier.

Nevertheless, he also indicated that projects penned by series creator Noah Hawley (Fargo, Legion) are often a commentary on people who've been in the public eye for the last few years, and/or the ever-changing nature of our own world. In Blenkin's view, then, it's possible that characters in the franchise's inaugural TV show might be crudely influenced by certain individuals who exist right now.

"I think that Noah did such a good job of painting a vivid character," Blenkin told me. "Like all of Noah's characters, they clearly have strands of the stuff that we're facing today and what's resonant right now.

"But what I love is that he [Kavalier] has very specific mannerisms and obsessions," Blenkin continued. "[He has] this Peter Pan obsession, he never wear shoes or socks, he has a little ball he throws about, his attention span is lacking, and he has an obsession with childhood and childhood innocence equating with the kind of genius [he is] and seeing himself as a boy who never grew up.

"He's able to break rules and not be held to the same account as an adult with that kind of morality," he added. "Everything that was written about him was so vivid on the page, so I kind of let the rest of the character threads take care of themselves."

Alien: Earth launches with a two-episode premiere on Hulu (US) on August 12 and Disney+ (internationally) on August 13. Before it arrives, read my review of Alien: Earth or get the lowdown on the series our dedicated guide on Alien: Earth.

You might also like

• Alien: Earth's official trailer is here – and the sci-fi horror series' xenomorph isn't the only frightening creature that'll need to be avoided
• Predator: Badlands' official trailer confirms the Alien and Predator reunion I've been waiting for
• Stranger Things star Finn Wolfhard has correctly guessed the outline of the popular Netflix show's live-action spin-off – and it's 'like David Lynch's Twin Peaks'

• BMW has launched an espresso machine built around a motorcycle engine
• The Big Coffee Boxer is hand-made, with a dual boiler and rotary pump
• It costs €7,900 (about $9,100 / £6,900 / AU$14,000), and only 80 will be made

If your usual morning coffee isn't giving you the same boost it used to and your wallet is weighing you down, BMW has just launched the espresso machine for you. The Big Coffee Boxer, made in collaboration with the coffee experts at ECM Manufacture, is built using a BMW R 18 Big Boxer motorcycle engine – and it'll certainly be a conversation-starter.

In terms of specs, the Big Coffee Boxer is up there with the best espresso machines. It features dual boilers, meaning you can pull a shot of espresso and steam milk at the same time, and professional-grade steam and hot water valves.

There's no color touchscreen here. Instead, the Boxer has two pressure dials (one for each boiler) and a discreet shot counter to help you see when it's time to backflush the machine, which is essential to remove residue and keep your coffee tasting as good as possible.

There's optional pre-infusion (a process that gently pre-soaks the ground coffee before applying the full brewing pressure), you can choose from three brewing temperatures, and use either a refillable water tank or a direct water supply if you're not opposed to a spot of plumbing.

Here's the catch

(Image credit: BMW, ECM)

Naturally, none of this comes cheap. Breville's new Oracle Dual Boiler raised eyebrows last week when it launched with a price tag of AU$4,499 (about $3,000 / £2,200), but the Boxer makes that look positively affordable as each BMW-branded espresso machine will set you back €7,900 (about $9,100 / £6,900 / AU$14,000).

To put that into context, if you currently pay $4 every day for a takeout coffee, it'll take you about six years and three months to offset the cost of the Boxer (not including the beans). You'll have to decide quickly, too, because only 80 of the machines will be made.

If that's a little outside your budget, take a look at our roundups of the best coffee makers and best bean-to-cup coffee machines, all of which are somewhat more affordable (if not as stylish).

You might also like

• The Sage Luxe Brewer Thermal is the drip coffee maker I've been waiting for all my life
• 6 ways to brew better coffee at home, according to a certified coffee grader – and the most common coffee myths busted
• Breville's new espresso machine can brew coffee and steam milk at the same time – and it's eye-wateringly expensive

This feature can help you identify certain chats so you don't let out a secret in the wrong chat.

Wix website builder review (2025)

Wix is a giant in the website builder industry. Its marketing campaigns across all mainstream media have made it a household name. While some services give you only a few templates, Wix has more than 500.

Pros

• Limited free plan
• Vast library of templates has something for every business type
• Huge number of professional features
• Easy to use
• Extensive ecommerce tools can build a feature-packed web store
• Excellent image features
• Hundreds of apps connect you to many other platforms and services

Cons

• Above average prices
• Can’t change a template once you’ve chosen it
• Only 2GB storage with the cheapest plan
• Email not included as standard (it’s a paid extra)

Squarespace website builder review (2025)

If you've been keeping an eye out for easy-to-use, all-in-one solutions to create a stunning website, you've surely stumbled upon Squarespace website builder somewhere along the line.

Pros

• Built-in email marketing
• Essential ecommerce features with most plans
• Lots of lovely, mobile-responsive templates
• Free trial with no credit card info required
• Helpful link in bio tool
• Free SSL certificate
• Helpful 24/7 customer support and well-supplied knowledgebase
• A solid set of features

Cons

• No free plan
• No telephone support
• Low level of customization

If you are looking for a simple solution to build a great website there are countless options out there. However, with many of the best website builders offering similar features, tools, and price points, it can be hard to know which one is the best option for you.

Wix and Squarespace are two of the market-leading website builders. They provide everything you need to create professional websites without coding skills - including website hosting, ecommerce features, easy-to-use interfaces, and more.

They even offer similar starting prices with Wix premium plans starting at $17/mo (or less with our Wix promo codes) and Squarespace coming in ever-so-slightly lower with its entry level plan starting at $16/mo (or less with our Squarespace promo codes).

Yet they do differ in many areas including tools, design flexibility, templates, and what you get with each plan. If you're looking to pick one over the other, our detailed guide will walk you through the highs and lows of each so you can make the right choice. Let's dig in.

Wix vs Squarespace: A detailed breakdown

Feature

Wix

Squarespace

Starting price

$17.00/month

$16.00/month

Free plan

Yes

No

Templates

900+ templates across multiple categories

180+ templates across 19 categories

Editor type

Highly flexible drag-and-drop editor with unstructured placement

Structured editor with Fluid Engine (grid-based drag-and-drop)

AI website builder

Wix AI - creates websites based on questions

Blueprint AI - generates sites based on brand personality and preferences

Storage

Starting at 500MB (varies by plan)

Unlimited on all plans

Mobile optimization

Dedicated mobile editor

Automatically responsive templates

Ecommerce features

Product management, point of sale, shipping options, abandoned cart recovery, advanced booking system

Product management, bookings, shipping options, point of sale, abandoned cart recovery

Payment gateways

80+ payment options

Limited payment options

Marketing tools

Email builder, newsletters, email campaigns, Facebook ads integration

Email templates, newsletters, campaigns, direct Instagram/Facebook sales

SEO tools

Meta titles/descriptions, URL customization, Google Search integration, canonical tags, image optimization, site inspection

Meta titles/descriptions, custom URLs, image alt text, Google Search Console integration, canonical tags

Analytics

Traffic monitoring, visitor behavior tracking, revenue reports, personalized suggestions, customer insights

Traffic monitoring, engagement tracking, sales statistics, Google Analytics integration, Purchase Funnel

App marketplace

500+ apps and integrations

35+ extensions

Blogging

Basic blogging features

Advanced blogging with monetization options

Customer support

Live chat, phone support (premium), knowledge center

Email support, knowledge base, community forum

Security

SSL certificates, DDoS protection, firewall

SSL certificates, DDoS protection, firewall

Design flexibility

Highly customizable with pixel-perfect positioning

More structured with focus on professional design

Multilingual support

Available through Wix Multilingual app

Requires third-party integration (WeGlot)

Forum feature

Available through Wix Forum app

Requires third-party integration

Live chat feature

Available through Wix Live Chat app

Requires third-party integration

Custom fonts

Direct upload in editor

Requires CSS code

Scheduling tools

Built-in booking system

Acuity Scheduling ($16/month)

Free domain

Yes (1st year)

Yes (1st year)

Content creation AI

AI tools for product descriptions, image generation

AI tools for product descriptions, email content, blog posts

Wix vs Squarespace: Features

Both Wix and Squarespace have strong features in 2025, but each shines in different ways. Wix boasts a huge template library with over 900 designs. In contrast, Squarespace has around 180 curated templates. Wix's app marketplace is much larger with over 500 integrations, while Squarespace has about 35 extensions. For ecommerce, Wix supports over 80 payment gateways and offers features like pre-order tracking and tax automation. Squarespace includes ecommerce in all plans but has fewer payment options.

Both platforms also use AI technology. Wix provides dedicated tools including product descriptions, image generation and editing, section editor, and even an AI marketing assistant. Squarespace offers AI for product descriptions, email content, and blog posts.

Squarespace's BluePrint AI helps build websites based on brand personality and preferences. Wix AI creates sites based on user questions in a chatbot interface. Squarespace generally has better blogging features with monetization options, but, Wix excels in business tools and customization.

The best choice depends on your needs. Squarespace suits users who value design quality and blogging features, plus it offers ecommerce in all plans. Wix is better for those wanting flexibility, extensive app integrations, and more business tools.

Wix vs Squarespace: Ease of use

Wix and Squarespace are both easy to use for non-programmers without coding experience, but the learning curve still varies.

Wix features a simple drag-and-drop editor. Users can place elements anywhere on the page. This allows for pixel-perfect positioning, ideal for beginners who want total customization. Wix also offers two editing options: the original editor for small businesses and creators and Wix Studio for agencies needing top-notch design tools.

Squarespace is user-friendly too, but its editing system is more structured. The Fluid Engine, introduced in July 2022, allows drag-and-drop within a grid area. This design offers fewer constraints than the Classic Editor but still maintains some order. Users get consistent and professional results but at the cost of creative freedom. It takes more clicks to achieve the same results as Wix, plus you must manually save changes.

Overall, Wix wins for ease of use, especially for beginners who want creative freedom. Its user-friendly interface, automatic saving, and strong backup system enhance accessibility. Squarespace may suit those who prefer structure and consistency, as its limitations help avoid design errors while ensuring a polished look.

Wix vs Squarespace: Support

Wix provides many support options, including live chat and phone support in over ten languages. However, priority phone support requires a Business Elite subscription. Support is available from Monday to Friday, depending on the language. Wix also has a Knowledge Center with tutorials and guides. For Wix Studio subscribers, the Wix Studio Academy offers hundreds of short courses to help users maximize features.

Squarespace takes a different approach to customer support. It doesn't offer phone support. Instead, you have the option to use 24/7 human chat support, hire third-party experts, or participate in the community forum where users can ask for advice. It also provides documentation and tutorials. But unlike Wix, Squarespace's live chat support system relies on human support agents, not AI.

Wix offers more responsive support that's easier to avail. But, Squarespace stands apart with its strictly human-first support system. While the former offers more accessible, diverse, and affordable options, Squarespace is the better choice for users who need constant access to experts who can troubleshoot complex issues.

Wix vs Squarespace: Pricing and plans

Plan

/mo (paid monthly)

/mo (paid annually)

/mo (paid every 2-years)

/mo (paid every 3-years)

Free

$0

$0

$0

$0

Lite

$24

$17

$14

$12

Core

$36

$29

$24

$21

Business

$43

$36

$29

$26

Business Elite

$172

$159

$121

$110

Wix offers a free plan and four paid tiers. The Light plan starts at $17/month, up from the previous $16. The Core plan costs $29/month, the Business plan is $36/month, and the Business Elite plan is $159/month. Wix's entry-level plans are cheaper than Squarespace's. However, the Light plan lacks ecommerce features, which begin with the Core plan.

Plan

Monthly cost (paid monthly)

Monthly cost (paid annually)

Personal

$25

$16

Business

$36

$23

Commerce (Basic)

$40

$28

Commerce (Advanced)

$72

$52

Squarespace does not have a free plan. It offers four paid options with annual discounts. The Personal plan starts at $16/month (billed annually). The Business plan is $23/month, while Commerce Basic (Plus) is $39/month, and Commerce Advanced is $99/month. All Squarespace plans include ecommerce capabilities, making it easy for users to sell products online. Squarespace also offers unlimited storage on all plans. In contrast, Wix's storage ranges from 2GB on the Light plan to unlimited on the Business Elite plan.

For value comparison, Wix gives more resources and features at similar price points, especially for business and ecommerce. However, Squarespace offers better value for users focused on ecommerce, as those features are included in all plans. Your best choice depends on your needs. Wix is more affordable for entry-level sites and offers more features at higher tiers. Squarespace provides robust ecommerce capabilities from the start.

Expert insight Wix vs Squarespace: Final verdict

After comparing Wix and Squarespace in detail, we think that both platforms have unique strengths for different users. Wix shines with its easy drag-and-drop editor, wide template library, large app marketplace, and responsive customer support. Its flexible design options and AI tools are great for those who want creative freedom and customization. With a free plan and lower prices, Wix is also more accessible for beginners and budget-friendly users.

Squarespace, on the other hand, lacks a free plan and phone support but stands out with its sleek templates and structured design. This approach ensures consistent, high-quality results. All Squarespace plans include ecommerce features, making it a better choice for users who want to sell products online from the start. Squarespace also offers excellent blogging tools, unlimited storage, and better site performance.

The best choice depends on your needs.

Choose Wix if you want maximum design flexibility, lots of app integrations, and lower costs. Go for Squarespace if you prioritize professional design, structured editing, better performance, and all-in-one ecommerce features. For most beginners and small businesses seeking value and ease of use, Wix is the more versatile option. In contrast, design-focused professionals and ecommerce businesses may prefer Squarespace.

Wix vs Squarespace: FAQsCan I change my website template after I've started building my website?

Squarespace allows you to switch your template at any time. Your website content will automatically be transferred into the new design, although it will likely need some editing to make sure it still fits and flows well. Wix doesn’t let you switch templates once you have published your site. If you want to use a new template after this, you will need to build your website from scratch.

Is Wix or Squarespace cheaper?

Both website builders have similar entry level plans with similar prices with Wix starting at $17/mo and Squarespace starting at $16/mo. As you move up the pricing tiers, the gap starts to widen, with Squarespace remaining the cheaper of the two, but Wix offering access to more tools and resources. Wix’s most expensive (Business Elite) plan works out at $159/mo (paid annually), compared to Squarespace’s most expensive plan that comes in at $99/mo.

• Microsoft engineers are committed to making WinUI open source
• But first it needs to be separated from proprietary Windows code
• GitHub will become WinUI's home for the community

Microsoft has confirmed WinUI will become, "truly open source", however deep entanglements with proprietary Windows code could put this goal a long way off yet.

Windows UI Library (WinUI) is a user interface framework for building modern, fluid and responsive user interfaces on Windows, which works with Win32, .NET and C++ apps.

However, while the project lacks a specific end date, Lead Software Engineer Beth Pan did share more details about Microsoft's phased plans in a GitHub post.

WinUI is on a road to become fully open source

"While we’re not ready to commit to a specific end date for completing all milestones, we are actively working toward it," Pan wrote.

The four phases of Microsoft's plan begin with more frequent syncing of internal commits to GitHub, starting post-WASDK 1.8 which is set for an August 2025 release.

From there on, Microsoft will allow external developers to clone and build the repo with full setup docs, after which third-party developers will be permitted to contribute and run tests.

The final stage will see GitHub become the "primary place for development, issue tracking, and community engagement."

However, because so much of the codebase touches proprietary Windows layers, Microsoft is planning a gradual and deliberate transition to separate what can be open-sourced.

"Our current focus is on foundational work that unlocks value for contributors and increase transparency," Pan added.

A separate GitHub project board has been established for the community to collaborate with Microsoft going forward.

Community responses have been generally positive and supportive, with many expressing their satisfaction that the project lives on with Microsoft's support.

You might also like

• We've listed the best productivity tools around today
• Fancy an upgrade? Check out the best laptops for progrmaming
• GitHub calls for major expansion in open source funding from the EU

• Turtle Beach is launching three new racing wheel products
• The VelocityOne Race KD3, VelocityOne F-RX, and Racer wheels release on September 9
• You're able to pre-order them now from the brand's website

Gaming peripheral brand Turtle Beach has just announced three upcoming racing wheels, all targeting budget to mid-range sim enthusiasts, and they may just be great additions to our best racing wheels guide in the future.

The Turtle Beach VelocityOne Race KD3, Turtle Beach VelocityOne F-RX, and the Turtle Beach Racer are all available to pre-order today from the brand's website, and will launch simultaneously on September 9, 2025. All these products are part of the 'Designed for Xbox' lineup, meaning they're compatible with Xbox Series X and Xbox Series S as well as PC.

Starting with the Turtle Beach VelocityOne Race KD3, this is a direct drive racing wheel that includes a wheel, 'K: Drive' wheel base, and a set of pedals. The motor will deliver 3.2Nm of force feedback and up to 2,160 degrees of rotation. It sounds like a suitably powerful mid-range option in line with the Logitech G923, and will retail at $449.99 / £329.99.

Next is the Turtle Beach VelocityOne F-RX. Similar to the Thrustmaster Ferrari 488 GT3, this is a standalone wheel suited to serious racing sim enthusiasts, and could be a great choice for iRacing or F1 25. It looks to have all the buttons, dials and switches necessary for an immersive sim racing experience, and will be available individually for $249.99 / £189.99. The F-RX is compatible with K: Drive wheel bases, too.

Finally, we have a budget option available in the Turtle Beach Racer. This looks to be the one to go for if you don't have room for a direct drive setup, and is more of a plug-and-play wheel. It has a lap mount if you're only option is playing on the couch, and also supports wireless connectivity with up to 30 hours of battery life. Do keep in mind that there may be some slight latency issues there, though. The Turtle Beach Racer will retail at $179.99 / £139.99.

You might also like...

• Don't let your Switch 2 get toasty – Nintendo warns against using the new console outside on hot days
• Battlefield 6 dev promises the game will return to the 'gritty' tone of earlier entries – 'you're not running around in a hero cape'
• Rare is getting a limited edition 8BitDo controller for its 40th anniversary, and I can't help but feel bittersweet about it

When Meta shocked the industry with its $14.3 billion investment in Scale AI, the reaction was swift. Within days, major customers (including Google, Microsoft, and OpenAI) began distancing themselves from a platform now partially aligned with one of their chief rivals.

Yet, the real story runs deeper: in the scramble to amass more data, too many AI leaders still assume that volume alone guarantees performance. But in domains like robotics, computer vision, or AR - that demand spatial intelligence - that equation is breaking down. If your data can't accurately reflect the complexity of physical environments, then more is not just meaningless; it can be dangerous.

In Physical AI, fidelity beats volume

Current AI models have predominantly been built and trained on vast datasets of text and 2D imagery scraped from the internet. But Physical AI requires a different approach. A warehouse robot or surgical assistant isn’t navigating a website, it’s navigating real space, light, geometry, and risk.

In these use cases, data must be high-resolution, context-aware and grounded in real-world physical dimensions. NVIDIA’s recent Physical AI Dataset exemplifies the shift: 15 terabytes of carefully structured trajectories (not scraped imagery), designed to reflect operational complexity.

Robot operating systems trained on these types of optimized 3D datasets will be able to operate in complex real-world environments with a greater level of precision, much like a pilot can fly with pinpoint accuracy after training on a simulator built using precise flight data points.

Imagine a self-driving forklift misjudging a pallet’s dimensions because its training data lacked fine-grained depth cues, or a surgical-assistant robot mistaking a flexible instrument for rigid tissue, simply because its training set never captured that nuance.

In Physical AI, the cost of getting it wrong is high. Edge-case errors in physical systems don’t just cause hallucinations, they come with the potential to break machines, workflows, or even bones. That’s why Physical AI leaders are increasingly prioritizing curated, domain-specific datasets over brute-force scale.

Building fit-for-purpose data strategies

Shifting from “collect everything” to “collect what matters” requires a change of mindset:

1. Define physical fidelity metrics

Establish benchmarks for resolution, depth accuracy, environmental diversity, and temporal continuity. These metrics should align with your system’s failure modes (e.g., minimum depth-map precision to avoid collision, or lighting-variance thresholds to ensure reliable object detection under specific conditions).

2. Curate and annotate with domain expertise

Partner with specialists: robotics engineers, photogrammetry experts, field operators, to identify critical scenarios and edge cases. Use structured capture rigs (multi-angle cameras, synchronized depth sensors) and rigorous annotation protocols to encode real-world complexity into your datasets.

3. Iterate with closed-loop feedback

Deploy early prototypes in controlled settings, log system failures, and feed those edge cases back into subsequent data-collection rounds. This closed-loop approach rapidly concentrates dataset growth on the scenarios that matter most, rather than perpetuating blind scaling.

Data quality as the new competitive frontier

As Physical AI moves from labs into critical infrastructure, fulfillment centers, hospitals, construction sites, the stakes at play skyrocket. Companies that lean on off-the-shelf high-volume data may find themselves leapfrogged by rivals who invest in precision-engineered datasets. Quality translates directly into uptime, reliability, and user trust: a logistics operator will tolerate a misrouted package far more readily than a robotic arm that damages goods or injures staff.

Moreover, high-quality datasets unlock advanced capabilities. Rich metadata, semantic labels, material properties, temporal context, enables AI systems to generalize across environments and tasks. A vision model trained on well-annotated 3D scans can transfer more effectively from one warehouse layout to another, reducing re-training costs and deployment friction.

The AI arms race isn’t over, but its terms are changing. Beyond headline-grabbing deals and headline-risk debates lies the true battleground: ensuring that the data powering tomorrow’s AI is not just voluminous, but meticulously fit-for-purpose. In physical domains where real-world performance, reliability, and safety are at stake, the pioneers will be those who recognize that in data as in engineering, precision outperforms pressure (and volume).

I tried 70+ best AI tools.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

As enterprise AI becomes more embedded into the fabric of everyday tools, the biggest challenge facing organizations isn’t AI adoption; it’s AI management. Gone are the days when AI features like meeting transcriptions or document summarization stood out as cutting-edge.

Today, they are expected. According to McKinsey's 2024 State of AI report, 72% of organizations have adopted at least one form of generative AI, and over half report using it in more than one business function. But this surge in adoption has led to a new operational crisis: AI sprawl.

What Is AI Sprawl and Why Does It Matter Now?

AI sprawl is the unchecked proliferation of AI tools and systems across departments, applications, and infrastructure without a unified strategy. The result? A chaotic digital ecosystem where:

• Redundancy is rampant (e.g., multiple summarization tools embedded in different apps)
• User experiences are inconsistent
• Data governance becomes unmanageable
• Security vulnerabilities go undetected

For example, companies eager to integrate AI across their tech stacks often deploy similar capabilities in silos - an AI assistant in a messaging platform, a different one in email, another in help desk software - without a shared interface or policy layer. This fragmented approach increases operational costs, confuses users, and makes compliance audits a nightmare.

The Rise - and Limits - of Vertical AI

Most enterprise AI today is what we call "vertical AI": narrow capabilities embedded directly into a specific tool, often by that tool’s own vendor. These AI features are excellent at solving bounded problems but struggle at scaling across workflows or departments.

IDC research notes that organizations are spending up to 30% more per seat due to overlapping AI functionality across their application ecosystems (IDC). While each solution may serve a use case in isolation, collectively they add inefficiency and cost.

The Real Cost of Fragmentation

Here’s where AI sprawl hurts the most:

• Wasted Spend: Gartner estimates that up to 25% of enterprise AI investment is duplicative, particularly in use-case specific tooling.
• Poor AI Literacy: Employees have to relearn how to interact with each tool’s AI assistant, eroding trust and slowing adoption.
• Regulatory Risk: Privacy settings and data policies vary app by app, creating blind spots for security teams and legal counsel.
• Broken Context: AI models can’t share knowledge between systems, meaning insights are trapped inside individual tools.

A Smarter Alternative: Interoperability as Strategy

Instead of asking, “How many AI tools do we have?” CIOs and CTOs must ask, “How well do our AI systems work together?”

Interoperability means more than just integrations or connectors; it requires AI tools that can share context, adhere to consistent governance, and surface insights across platforms. This horizontal approach avoids the trap of buying more features and focuses instead on making those features work in concert.

Three Core Benefits of AI Interoperability

1. Holistic Intelligence: AI-driven insights from one tool (ex: CRM) can inform decisions in customer support, marketing, and HR when systems talk to each other.
2. Trustworthy User Experience: Employees get consistent behavior, language, and recommendations regardless of the app they’re using.
3. Centralized Oversight: IT and security teams can manage data policies, model updates, and risk controls from a single pane of glass.

Charting a Coherent Path Forward

To navigate from fragmentation to function, enterprise leaders must pursue both operational alignment and robust governance practices. The good news is that AI sprawl is not an inevitable cost of innovation - it can be addressed proactively.

By taking a strategic approach that blends centralized governance with interoperable infrastructure, organizations can rein in AI fragmentation before it becomes unmanageable. The way forward is clear, actionable, and within reach.

• Build a centralised AI governance council that includes IT, compliance, legal, and business users.
• Define enterprise-wide AI usage policies and audit mechanisms to ensure consistent, responsible practices.
• Implement monitoring tools that track model performance, data lineage, and access across platforms in real time.
• Consolidate and rationalize tools to eliminate duplicative spend and improve oversight.
• Audit the AI landscape by inventorying every AI-enabled tool, feature, and data dependency across the organization.
• Invest in AI infrastructure by adopting open standards like MCP, APIs, and orchestration platforms that promote interoperability.
• Upskill employees through literacy programs that demystify AI, reduce risk, and build trust in intelligent systems.

In fragmented environments, IT and compliance teams are often required to support multiple incompatible permissioning models, audit trails, and deployment protocols. A centralized platform enables governance teams to monitor model performance and data lineage in real-time, reducing exposure while aligning AI use with evolving regulatory expectations.

Less Hype, More Harmony

Enterprise leaders need to stop chasing the next flashy AI feature and start focusing on cohesion, governance, and usability. The future isn’t about having the most AI, it’s about having the most effective, connected, and secure AI.

The maturity curve for AI adoption will increasingly reward organizations that move beyond fragmented experimentation. Those who consolidate capabilities and embed AI within core processes will unlock sustainable growth, resilience, and competitive advantage.

In the age of ubiquitous AI, everyone has tools, but not everyone has traction. The innovators aren’t the ones with the most features; they’re the ones who make it all work together. AI sprawl may be a modern challenge, but orchestrated intelligence is the competitive edge of tomorrow.

We list the best employee experience tool.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

Artificial intelligence (AI) isn’t something on the horizon. It’s already part of how people are getting work done.

Recent research from HP and YouGov found that 72% of UK employees using AI tools say it saves them time every week. One in ten are saving more than five hours. Some are using it to reduce manual admin. Others say it helps them focus, collaborate more effectively, or feel more in control of their day.

But these gains aren’t coming from structured enterprise rollouts. In many cases, they’re the result of quiet experimentation - employees using what’s already at their fingertips, often without training or direction from IT.

At the same time, more than a quarter of UK businesses still report having no formal AI strategy. This creates a growing disconnect: employees are forging ahead on their own, while the organization risks falling behind. It’s not a technology gap; it’s a leadership one.

In my conversations with CIOs and IT leaders across the UK and wider Northwest Europe market, I hear a mix of urgency and uncertainty. Everyone agrees AI is critical to future competitiveness. But there are open questions around where to start, how to scale responsibly, and how to balance experimentation with governance.

That hesitation is understandable, especially in industries where risk and compliance frameworks are tight. But as more teams adopt AI organically, the absence of a centralized plan introduces its own risks - from data leakage to inconsistent performance and lost opportunities for enterprise-wide value.

A rare opportunity to re-architect from the ground up

The end of Windows 10 support in 2025 presents a strategic window. Many organizations are already reviewing their device strategies and digital estate planning. This moment, whether viewed as a compliance trigger or a chance to modernize, is an ideal time to align IT infrastructure decisions with longer-term goals around workplace tools and AI integration.

We’re seeing growing interest in AI-capable endpoint devices as part of that strategy. These systems offer local processing, reduced latency, and better data control-critical features for organizations managing hybrid environments or strict regulatory requirements. But while improved performance and privacy are important, the real benefit is this: AI becomes embedded, accessible, and usable without disrupting the way people already work.

I’ve spoken with IT leaders who are introducing AI incrementally through use cases that matter to employees: summarizing meetings, creating first drafts, reducing clicks. It doesn’t need to be complex to be effective, but it does need to be intentional.

From pilot mode to platform mindset

Too many organizations remain stuck in test-and-wait mode. A pilot project goes well, but momentum fizzles. There’s no clear business owner, no framework to expand, no metrics to track long-term impact. Here, AI remains confined to one team or workflow, useful but limited.

To unlock real value, businesses need to stop thinking in projects and start thinking in systems. That means moving AI out of isolated pockets and into the core of IT and business strategy. From what I’ve seen across sectors, this shift requires three mindset changes.

First, move from experimentation to prioritization. AI isn’t a side initiative anymore. It needs sponsorship, resourcing, and KPIs tied to outcomes the organization cares about - whether that’s productivity, cost savings, or faster decision-making.

Second, move from scattered adoption to secure design. Governance, data privacy, and accountability must be built in from the beginning. In regulated industries, this is non-negotiable. But even in more flexible sectors, employees need to know where AI fits and what the boundaries are.

Third, move from short-term rollout to long-term enablement. AI success isn’t about deployment alone. It’s about building trust, training users, and supporting adoption in ways that stick. That means investing in support infrastructure-not just software licenses.

Some of the most effective CIOs I’ve worked with are building cross-functional AI working groups that bring together IT, data, ops, HR, and business units. These teams aren’t just coordinating rollouts-they’re shaping roadmaps, reviewing risks, and evolving policies together. That kind of alignment isn’t flashy, but it’s what allows AI to move from tactical to transformative.

AI that works - for people and the business

Beyond the tech stack, there’s a broader benefit to consider. In the same HP and YouGov research, AI users reported lower stress, improved work-life balance, and greater satisfaction with their roles. When implemented well, AI doesn’t just make work faster, it makes it more manageable and more meaningful. That translates into retention, productivity, and culture shifts that directly affect the bottom line.

As IT leaders, we don’t just manage systems, we shape environments. Our job is to build the foundations that allow people to do their best work. And increasingly, that means designing ecosystems where AI can be adopted confidently, used securely, and evolved sustainably.

The momentum is already there. Employees are experimenting. The tools are ready. The opportunity now is to implement structure and take those individual wins and build a strategy that turns them into lasting, measurable impact.

We list the best employee management software and the best employee experience tool.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

Here are the answers for The New York Times Mini Crossword for Aug. 6

It's reportedly "preparing to introduce" the unique form factor later this year.

• Microsoft's VP for OS Security has provided his vision of Windows in 2030
• It will be multimodal and involve 'more talking to our computers'
• AI will power the ability to 'do much more sophisticated things'

Ever wondered what Windows will be like at the turn of the decade, when 2030 rolls around?

Windows Central discovered a video clip uploaded on Microsoft's YouTube channel in which its Corporate VP for OS Security, David Weston, provides his vision for Windows in 2030 (you can watch it below).

In the short interview, Weston delivers answers to some set questions which are mostly on the topic of security (unsurprisingly, given that's his expertise), AI, jobs, and the business world. He does address the title of the video at one point, though, and gives us his thoughts on how Windows might look by the end of the decade.

Weston observes: "I think we will do less with our eyes and more talking to our computers. And I truly believe that a future version of Windows, and other Microsoft operating systems, will interact in a multi-modal way."

"The computer will be able to see what we see, hear what we hear, and we can talk to it and ask it to do much more sophisticated things. I think it will be a much more natural form of communication."

Weston adds: "The world of mousing around and typing will feel as alien as it does to Gen-Z to use MS-DOS."

Much of the rest of the video discusses AI and jobs, as mentioned, and how we can expect AI to take over grunt work to free us humans up to do more interesting and creative tasks (or that's the long-held theory anyway).

And indeed, how future security experts will be AI bots that you'll interact with just like a real person, talking to them in video chats and meetings, or emailing to give them tasks.

Analysis: Far-fetched?

To me, this doesn't feel like a vision of Windows in five years' time (well, it's nearer four if we want to nit-pick, and I do), but a good deal further out than that. Although Weston does hint that this is a broader vision of a 'future version of Windows', and I get the gist: the future is 'multimodal' - moving away from the simple mouse and keyboard as the main inputs for the PC - and, of course, everything's built around AI (naturally).

Will the future of Windows be like this, though? I'm certainly not betting against it being focused heavily on AI, as that very much looks to be the case. In general, AI feels like an almost irresistible force in terms of where computers are heading, and Microsoft is clearly trying to jam more AI into Windows wherever it can - a path that the software giant is doubtless going to forge ahead with.

Today, I've been writing about clues hidden in the background of Windows 11 that suggest another AI agent might be coming to the taskbar in the desktop OS. That possible addition would live alongside the agent already introduced to the Settings app, which is a smart addition.

With powerful NPUs potentially set to be included in desktop chips soon, as well as Copilot+ laptops, AI is likely to become much more widespread in the world of PCs pretty swiftly. I'd even go as far as to guess that the next version of Windows won't be Windows 12, but Windows AI (or Windows Copilot maybe, if that's still the brand for AI), the focus on this arena is likely to be that strong.

There are promises, lofty ideas, and marketing around AI, though - and then the reality of what Microsoft can achieve. Remember when Copilot was first introduced to Windows 11? We were told it would be able to change a swathe of settings in the operating system based on a vague prompt from the user (like 'make me more productive'). That still hasn't happened, and appears to be firmly on the back burner.

Which is to say that while I don’t doubt that Microsoft has these big ambitions, whether a very different way of working with a Windows PC will happen in 2030 seems doubtful to me.

Granted, I can indeed envision that talking - giving voice commands (which are coming along nicely in Windows 11) - could become a much more important, but still supplementary, part of the Windows experience and interface. And AI (presumably) doing more sophisticated things, yes, fair enough - maybe even manipulating Windows settings in one fell swoop at the behest of the user will be realized in a manner that works well.

Hey, maybe Windows AI, or Windows 2030, or whatever it ends up being called, will finally get rid of the legacy Control Panel, as a commenter on Weston's video amusingly observes. Hah - it makes me feel giddy just to imagine it. This is a battle Microsoft has been fighting for far too long, after all,

But mouse-and-keyboard usage is being made to feel like the equivalent of us being forced to revert to the days of DOS, all text and tinkering with the config.sys and autoexec.bat files to get a PC game to work? That feels like more than a stretch, and something much, much further away in the Windows computing timeline - but I could be wrong.

You might also like

• Microsoft promises to crack one of the biggest problems with Windows 11: slow performance
• Microsoft finally plays its trump AI card, Recall, in Windows 11 – but for me, it’s completely overshadowed by another new ability for Copilot+ PCs
• Microsoft has all but given up on Windows 11 SE – and it looks like the war against Chromebooks has been lost

• Microsoft is prepping Windows XP Crocs, dubbed the '50th Anniversary Exclusive'
• It'll be the first version of Windows to take this step, if true
• The Crocs come with six Jibbitz and a carrying tote

Yes, Microsoft is still celebrating its 50th anniversary, and while the company has done a lot of looking back, it’s also looking forward. Err, at least taking a step forward.

Sure, we’ve seen some iconic Windows ugly sweaters, including one with Minesweeper and one with Clippy, but Windows XP is going where no other version of Windows has ever gone before – to Crocs.

TechRadar's confirmed with the tech giant that the Microsoft 50th Exclusive Crocs – aka the Windows XP Crocs – are official, and got five images of the shoes.

According to a report from The Verge, Windows XP Crocs are currently available for internal order by Microsoft employees – priced at $80 – with the story noting that the employees “get first dibs” ahead of a “worldwide launch.”

Image 1 of 5

(Image credit: Microsoft)Image 2 of 5

(Image credit: Microsoft)Image 3 of 5

(Image credit: Microsoft)Image 4 of 5

(Image credit: Microsoft)Image 5 of 5

(Image credit: Microsoft)

We’ve seen other collaborations from the Croc brand, with plenty of Disney properties included – I mean, kachow, Lightning McQueen Crocs that light up – along with fashion houses, and even McDonald's. The Windows XP Crocs, though, take the iconic green hills and blue skies wallpaper to the shoe form.

And I know what you’re thinking, but the images of the Windows XP Crocs do indeed confirm the existence of a Clippy Jibbitz (aka what Crocs calls their shoe charms). The Windows XP Crocs will come with an iconic helper as well as a pointer, the MSN butterfly, a classic Internet Explorer logo, the recycling bin, and a folder. That comes to a whopping six Jibbitz in total.

(Image credit: Microsoft)

You also get a drawstring tote that's inspired by the classic, now iconic, Windows XP wallpaper. Microsoft did confirm the existence of the Crocs to us and shared these images, but didn't share anything more on pricing or availability.

At a reported price tag of $80, the Windows XP Crocs aren’t cheap, but if you’re a Microsoft collector or someone who’s also opted to get the previous ugly holiday sweaters, they might be the perfect shoe to add to your collection. Of course, I think many would be happy if Microsoft goes the route of other retro, nostalgia-fueled drops – it could be a fresh skin for Windows or even another wallpaper drop, and that would still be a great way to honor the 50th.

You might recall that Microsoft dropped a limited 50th anniversary edition of the Surface Laptop, which looked pretty snazzy. It’s also a more subtle way to celebrate 50 years of Microsoft than, say, blue and green Crocs.

Stick with TechRadar as once we learn more about pricing and how to get a pair of the Microsoft 50th Exclusive Crocs, we'll be sure to update this post.

You might also like

• Liquid Glass on iMac G4 is the mashup I didn't know I needed – now I can't stop thinking about it
• Confused about why you can't upgrade to Windows 11? Microsoft has some new advice that might help... including a trick I wasn't aware of
• Disney Imagineers brought Walt Disney back to ‘life’ and I still can’t figure out exactly how they did it

• The Yamaha True X Surround 90A is a new four-box Dolby Atmos soundbar
• 12 up-firing drivers, four oval full-range drivers and three tweeters in the bar
• Newly developed subwoofer, and rear speakers are also stand-alone Bluetooth speakers

What's better than having a couple of upward-firing Dolby Atmos speakers? Having a dozen of them. That's what Yamaha has delivered in its new True X Surround 90A soundbar system, aka the SR-X90A.

The True X Surround 90A is a high-end, high-spec home theater soundbar system with Dolby Atmos and DTS:X support, and that builds on the firm's True X 40 and 50 models, while incorporating tech from Yamaha's Sound Projector range.

According to Yamaha it delivers "an amazing home theater experience that goes beyond the realm of conventional soundbars".

Yamaha True X Surround 90A: key features and pricing

(Image credit: Yamaha)

The SR-X90A takes the same beam technology from the YSP-1 soundbar and applies it to a dozen up-firing speakers powered by Yamaha's YDA-141 amplifier. There are six speakers dedicated to the height channels projecting upwards at each end of the speaker, and according to Yamaha the results rival true ceiling-mounted speakers.

The very best Dolby Atmos soundbars tend to use four up-firing speakers (in the case of the Samsung HW-Q990F) or even five in the case of the LG S95AR. Some custom-install Dolby Atmos home theaters might use six in-ceiling speakers. 12 up-firing speakers is… hardcore.

The soundbar also features the Surround:AI processing from Yamaha's AV receivers, which is the first time it's been made available in a soundbar.

Those up-firing speakers are teamed up with newly developed eye-shaped oval drivers, which Yamaha says can deliver powerful audio without making the soundbar massive. There are four of these large oval drivers to cover the full frequency range, in conjunction with three tweeters. The speakers are arrange in left, center and right configurations on the soundbars front.

There's a newly developed subwoofer too, which keeps Yamaha's patented symmetrical flare port, and which has an internal plate to control the airflow in order to reduce vibrations from air turbulence and speaker movement. Yamaha says port noise is reduced by up to 20dB compared to more conventional designs.

The True X Surround 90A uses Yamaha's True X wireless connectivity for soundbar, subwoofer and satellite speakers, and the rear speakers it comes with can also be used as stand-alone Bluetooth speakers.

The system also has Yamaha's MusicCast network system for multi-room audio and in-app customization and configuration, and it's Apple AirPlay compatible too.

The True X Surround 90A will be available from September 2025 with an expected recommended retail price of £2,499 / AU$4,499 (about $3,300) – you can expect a more concrete US price closer to the time, depend on the latest tariff situation.

But before then, you can expect our early verdict on this soundbar – it's on its way to us, and this looks like a very exciting addition to the world of the best soundbars.

You might also like

• The best soundbars: TV audio upgrade for all budgets
• The best Dolby Atmos soundbars for all budgets
• The best cheap soundbars, all tested by our experts

Here are hints and the answers for the NYT Connections: Sports Edition puzzle for Aug. 6, No. 317

Here are some hints and the answers for the NYT Connections puzzle for Aug. 6, #787.

The next flagship model for the most popular AI platform is expected soon.

Here are hints and the answer for today's Wordle for Aug. 6, No. 1,509

If you struggled to solve many of the Wordle puzzles served up last month, then don't be too hard on yourself: it was the toughest month in the game's history.

I've crunched the numbers and, by my reckoning, it left every other month so far in the dust for difficulty, with an average score for the 31 games of 4.22.

That's according to the daily figures reported by WordleBot, the game's AI helper tool, which records the average among the many thousands of people who play. In turn, I've kept a list of those averages since the 'Bot launched in April 2022, meaning I now have a spreadsheet ranking 1,221 games by difficulty.

The bad news is that rather than being merely a statistical anomaly, that tough run may point the way towards Wordle's near future. My daily Wordle hints might well be even more useful from here on.

Wordle's month from hell

Regular Wordlers will be in no doubt as to the game's difficulty last month, with a string of near-impossible words causing all kinds of problems.

There was TIZZY, for instance, with its repeated letter Zs and average of 4.9, and POPPY with its triple Ps and 4.8 score.

FOIST might not look as difficult, but that was a classic example of Wordle's letter-trap games, where the first letter can be changed to make several other words, in this case JOIST, HOIST, and MOIST; that one also came in at 4.8.

SAVVY, with its double Vs, also hit that score, while BALER at 4.7 was one of the nastiest ER games we've had recently. EXILE (4.6), NERVY (4.5), and FRILL also caused problems, and the fact that LORIS (4.2) can be considered easy in this company points towards the overall difficulty.

July 2025 in Wordle: the 10 toughest

Game

Answer

Date

Average score

My score

1497

GOFER

Friday, 25 July 2025

5.6

5

1482

JUMPY

Thursday, 10 July 2025

5.2

5

1493

TIZZY

Monday, 21 July 2025

4.9

4

1475

POPPY

Thursday, 3 July 2025

4.8

6

1487

FOIST

Tuesday, 15 July 2025

4.8

4

1500

SAVVY

Monday, 28 July 2025

4.8

4

1477

BALER

Saturday, 5 July 2025

4.7

4

1484

EXILE

Saturday, 12 July 2025

4.6

4

1488

NERVY

Wednesday, 16 July 2025

4.5

4

1503

FRILL

Thursday, 31 July 2025

4.4

4

But the two worst last month were JUMPY and GOFER – with scores of 5.2 and 5.6, respectively.

Only 21 games in Wordle's history have passed the 5.0 mark (with another nine at exactly that score), so to get two in the space of two weeks is the stuff of nightmares.

JUMPY's problem was that J at the start; as my analysis of every Wordle answer shows, J is the least common letter in the game by far, so spotting it is rarely easy. The existence of LUMPY, DUMPY, and BUMPY will also have been a factor.

With GOFER, meanwhile, it was the combination of an ER ending (the most common in the game) and the not-so-common letters G and F that caused the issue. Its 5.6 score places it as the equal fourth hardest ever, behind only PARER, MUMMY, and CORER, and level with ROWER.

If you failed to solve any of them – or even all of them – then it's entirely understandable; Wordle is a simple game, but it can be fiendishly tricky at times.

Tougher than the rest

WordleBot only launched in April 2022, a couple of hundred games into the series, so it's possible that December 2021 or March 2022, or another month, was even more difficult. But I doubt it. I've played every Wordle so far and lost only once, and I certainly don't recall anything like July 2025.

To confirm my hunch, I tallied the average score for each day to get the overall average for July, then repeated the process for each of the other 38 months for which I have full details.

One thing I found interesting was that July 2025 wasn't just the most difficult so far – it was the most difficult so far by a long way.

The month's overall average of 4.22 might not sound that much higher than that for October 2024, the next highest in the list at 4.15, but it's statistically significant given that there's only a 0.04 difference between months #2-7 in the list.

Plus, it's a whopping 0.57 guesses harder than the easiest month, December 2023 – which came in at 3.65 – and way higher than the game's overall average of 3.97.

Hard times are coming

One notable feature of July's Wordles was that there were five 'non-original' answers among the 31 games.

When Josh Wardle created Wordle, he and his partner drew up a list of 2,315 words which would form the game's answer list, then scheduled them to appear one a day for the next six or so years.

The New York Times removed a few of those when it bought Wordle in 2022, then left the list more or less unchanged for the next year. Then, in March 2023, it gave us GUANO – the first 'extra' solution added to the original pool and the start of a new era for Wordle.

More have followed since then, 17 in total, including such gems as UVULA, SNAFU, PRIMP, and MOMMY, all of which have been hard in their own right. But these words have been spaced apart, with most months seeing just one or none at all.

There have been exceptions, with June 2023, November 2024, and May 2025 all having two, and January 2025 having three. But to get five in a month, as we did in July, was unprecedented.

(Image credit: New York Times)

And all were difficult: ATRIA was a 4.1, NERVY a 4.5, LORIS was 4.2, TIZZY 4.9, and GOFER that immense 5.6. The average across those five games was a staggering 4.66; these were all genuine head-scratchers.

And the thing is, the NYT is going to have to keep adding more of these as time goes on. That's because we're now past 1,500 Wordles, meaning we have only around 800 original games left, with as yet no idea what will happen to the game when that list runs out.

The smart thing for the NYT to do is extend it for as long as possible, which means adding more words. And there lies the problem. Wardle's list already covers many of the most obvious five-letter words in English, so we can expect the majority of the newly added words to be more difficult than the average.

So you can forget about classic Wordle start words such as STARE, CRANE, and SLATE being added – they've all already been and gone. So too ultra-common English words such as HOUSE, TODAY, and BELOW; they've all been past Wordle answers too. Instead, you can look forward to more like BALSA, KAZOO, BEAUT, SQUID, and TAUPE; uncommon words, slang words, words with uncommon letters…

August initially continued the July trend, with BANJO and DAUNT both coming in at 4.4, but the next few games were a little easier; maybe the NYT was giving us all a breather. But don't be surprised if things get tougher again soon, because this game is only going one way from here. Don't say you weren't warned.

You might also like

• Wordle is 1000 – these are the 50 hardest games so far
• I analysed every Wordle ever – here's what I found
• I've played all 1,000 Wordles and only lost once — here are my tips