Marissa Adams, HSBC Regional Head of Global Trade Solutions for Europe and Americas, discusses the reasons for optimism among US companies about growing international trade versus their global counterparts. She speaks to Bloomberg’s Lisa Abramowicz and Dani Burger on ‘Surveillance.’ (Source: Bloomberg)
Source: Bloomberg Markets.
Health secretary tells MPs that government ‘will not look away as children get unhealthier’
Sweet-toothed drinkers face paying more for their milkshakes and pre-packaged lattes after the government confirmed plans for a tougher sugar tax.
Designed to tackle obesity the levy currently applies to drinks with a sugar content of 5g for each 100ml. However, after a public consultation this is being cut to 4.5g for each 100ml.
Source: World news | The Guardian.
to your child opens up a world of exploration through music and audiobooks. They can listen to tunes while playing, share their favorites with friends, and hold dance parties in their bedrooms. Stories and podcasts will take them to new worlds, both real and imagined. And when it’s time for bed, they can drift off to soothing sleep sounds, lullabies, or sleep stories. I’ve tried a string of different speakers with my kids over the years and found them fun, helpful, and a great way to cut down on screen time. These are our favorites.
Check out our guides to the Best Bluetooth Speakers, Best Smart Speakers, Best Kids’ Headphones, Best Wireless Headphones, and Best Headphones for $100 or Less for more.
Updated November 2025: We’ve changed our top pick, added Qubitunes, and refreshed existing entries.
How to Keep Your Kid Safe Online
There’s an element of risk with any device that can connect to the internet, so take the time to configure your child’s speaker and ensure that any related accounts are secure.
Discuss rules: Talk with your children to agree on rules about what they can use their speaker for, and, for smart speakers, how they should speak to it.
Secure your accounts and network: Make sure you select strong, unique passwords for your devices and Wi-Fi network, using a mix of lowercase, uppercase, numbers, and special characters. Always use two-factor authentication (2FA) where available (here’s how to activate 2FA on your Amazon account).
Use parental controls: You can set up parental controls with Apple, Google, and Microsoft accounts, and use them to lock down various devices, including speakers. You may also have parental controls on your router or mesh that enable you to schedule downtime for your child’s devices, so they can’t access the internet in the middle of the night, for example.
Consider privacy: You might want to limit how voice recordings from your child’s speaker can be used. We have guides on how to keep your recordings private, how to delete stored voice recordings and activity, and how to make smart speakers as private as possible.
The original Yoto Player (7/10, WIRED Recommends) was excellent, but Yoto has made significant improvements since then, adding stereo sound, USB-C charging, and better battery life, making the third generation our favorite kids’ speaker. It has a cute pixelated clock display on the front, with a day or night image that tells kids at a glance whether it’s time to get up yet. Content is mixed, and kids can play different books by inserting cards into the slot on the top. There are audiobooks from popular authors like Julia Donaldson and Roald Dahl, and prices are similar to Audible. You can also buy a pack of 10 blank cards and record your own content.
What elevates the Yoto Player is the inclusion of kid-friendly radio stations and a daily podcast. There’s even a sleep mode that plays a mix of lullabies, classical music, and sleep sounds, and it’s easy enough to use that your child can switch it on themselves if they wake in the night. It does need an internet connection, though you can play any content that has been downloaded (it has 32 GB of storage) when the Yoto Player is offline. There’s an optional circular wireless charging base, so there’s no need to fiddle with cables. The third-gen Yoto offers up to 24 hours of battery life. The additional Bluetooth support and 3.5-mm headphone port give the Yoto Player a little more versatility. The Yoto Mini ($80) is a smaller, more portable version that offers most of the same features.
For kids aged 3 to 12 years.
The durable Toniebox is a squishy cube with ears that double as volume controls. It’s built to withstand tumbles, spills, and hugs, and you can leave your kids alone with it without worry. The Toniebox plays stories and songs connected to different characters or “Tonies,” triggered when your child places the relevant figurine on top of the speaker. There are many Tonies to choose from and tie-ins with familiar favorites like Paddington Bear, Disney, and Pixar movies like The Lion King and Cars, and many popular characters.
It is Wi-Fi–connected, with content downloaded to the speaker when a character is placed on top, but there’s no danger of your child accessing anything they shouldn’t, and there is an offline mode. You can also buy Creative Tonies to record up to 90 minutes of your own audio or have grandparents read stories to your kids. The hand-painted figurines are super cute, but they’re expensive, and you can expect tears if any go missing. Kids have the capacity to listen to the same thing over and over, and Tonies don’t offer a great deal of content. They are generally less than an hour and can be as short as 16 minutes. The lack of additional content or Bluetooth connectivity limits the appeal.
Tonies just released the Toniebox 2 ($140), adding Tonieplay games, a sunrise alarm, USB-C charging, and upping the storage from 8 GB to 32 GB, but also increasing the price. We plan to test it next.
For kids aged 3 to 7 years.
This screen-free play stage for kids has a wonderfully clever design with a circular sound stage where you can insert a cartridge and pop characters on top to change the audio that’s played. The stage looks like a face with ears that act as buttons and dials for power and volume. Everything is made from wood, and it feels durable. Using the sound stage is intuitive, with four different areas, and kids can experiment to build up different soundscapes, change stories, or play games by taking the wooden pieces on and off the stage or switching them around. You need to connect to Wi-Fi via the mobile app on your phone for the initial setup, content download, and updates, but there’s no need for a connection after that.
Qubitunes’ functionality allows for a wide variety of add-on packs, though you do get a basic set of characters with the speaker. There are also stories about quirky characters in the Groblies ($20), a fun cooking game called Bernard’s Kitchen ($20) with different ingredients, and my favorite, the Little Maestro set ($20) featuring players in an orchestra covering Mozart, Vivaldi, Beethoven, and Bach. The stereo speakers are quite good, there’s a headphone port, and it can be recharged via USB-C for up to six hours of play. My wife has this in the reward area of her classroom, and it’s the most popular option with 7- and 8-year-olds.
For kids aged 3+ years.
With a lovely mix of music, stories, and sounds, all available in eight languages, this speaker successfully combines fun and education. It is durable, has its own handle, and is simple enough for toddlers to operate. Timio is for kids aged two years and up. There’s no need for a screen, app, or internet connection, as your child can play content by selecting one of the plastic discs and slotting it on top. You get five discs in the box covering classical music, farm animals, vehicles, lullabies, and bedtime stories. Each disc looks like a clock face with pictures at each hour, and kids simply press the one they want. Additional sets of discs are $22 and cover all sorts of topics, from dinosaurs to learning colors.
The speaker sounds surprisingly good, and there’s a 3.5-mm audio port for kids’ headphones. Timio is a great way to introduce your kids to other languages or for multilingual households (you can press and hold the language button to switch between English, Spanish, French, German, Dutch, Chinese, Italian, and Portuguese). There’s also a quiz mode, with prompts like find the police car, identify numbers, or select a specific shape. Timio does require three AA batteries, and when you screw open the back panel, you will also find the SD card that holds the content. We recommend rechargeable batteries, as you will change them often if Timio proves popular with your kids. Although they are large and you get a bag to keep them in, there is a risk that discs will go missing.
For kids aged 2+ years.
This Alexa-powered smart speaker covers all the bases, with support for music and audiobooks alongside a world of kid-friendly extras included in the free year of Amazon Kids+ (usually $3 per month). Sound quality is solid, Alexa is responsive, and you can connect to your music streaming service. This Kids’ version comes painted as a tiger or panda, and Kids+ provides age-appropriate games, Audible books, and educational skills. The parental controls are great, and I appreciate extras, like the ability to broadcast when it’s dinnertime. You can configure everything by visiting the Amazon Parent Dashboard in your web browser or the Amazon Kids+ app (you must sign in using your Amazon account details).
There’s a physical button to turn the microphone off, but you still might be uncomfortable with the idea of letting kids loose with a smart speaker. The content filters worked just fine, but if you want to be certain your kids won’t access something you don’t approve of, you’ll have to micromanage. If you’re a Prime subscriber and your kids have Fire tablets, this speaker is easy to recommend, but you can also use it as a regular Bluetooth speaker and stream any content from your phone. Older kids will be better served by the Echo Dot Max ($100 at Amazon), which offers superior sound quality and still supports parental controls.
For kids aged 3 to 12 years.
If you’re uncomfortable with the idea of an internet-connected device for your child, this incredibly cute, wooden speaker is styled like a radio but doesn’t connect to anything. It has two dials on the front: One has animals painted around it (like a lion, elephant, and koala bear), and the other has environments (like the jungle, castle, or desert). The idea is that your kids can turn the dials and get a matching story. There are 128 “meditative journeys” in all, alongside some gentle music tracks and nature sounds. It’s designed for kids 3 to 10.
The focus here is on relaxation, and these meditations are suitable for falling asleep or just for a calming break. The battery life is around three hours. Sadly, you have to plug in the Micro-USB cable to charge it, which isn’t as versatile as the reversible USB-C port for young kids. Aside from charging, it’s easy enough for kids to operate, and it’s a nice option for kids who wake at night and need a little soothing to drift off again. Sadly, this seems to be out of stock in the US, but it’s available in the UK.
For kids aged 3 to 12 years.
Parents seeking a safe way for their kids to listen to audiobooks might consider Voxblock. Refreshingly, there’s no need for a screen, app, or internet connection. Each audiobook (sold separately) resembles a real book and contains the audio. Your kids simply select the title they want and slot it into the speaker to play. There are controls to play or pause, change the volume, and skip chapters. Pause or remove a book, and it remembers your place. You can also plug in headphones via the 3.5-mm audio port and recharge the Voxblock via the USB-C port (cable included, but you need your own wall charger).
Voxblock is very easy to use. The wee books, made from recycled card, are super cute. My daughter also loves how portable Voxblock is. On the downside, the battery life is limited (we had to charge every couple of days). There’s a handy bedtime mode that automatically turns off after 20 minutes, but it’s a shame it can’t run for longer. You get three audiobooks and a protective bumper with the starter pack. There are more than 100 audiobooks available, with prices ranging from £6 to £14 each, but they are relatively expensive. For example, Kensuke’s Kingdom is £14 on Voxblock, but costs half that on Audible.
Sadly, Voxblock is only available in the UK. US-based customers looking for an audiobook alternative might want to consider the Kindle Kids ($130), which, along with the Echo Dot Max, can access books via Audible on Amazon Kids+.
For kids aged 2+ years.
For kids who struggle to relax or drop off to sleep at night, this cute turtle-shaped speaker can help. The Zenimal Kids+ is one of our favorite sleep gadgets because it contains nine guided audio meditations to calm your child and three sleep soundtracks to play as they nod off. It’s shaped like a turtle, covered in soft and durable silicone in teal or purple, and is small enough for kids to hold or place next to their pillow. The Zenimal Kids+ does not require an internet connection, app, or screen.
Even young kids can press the buttons to trigger meditations and sleep sounds. The meditations come on removable SD memory cards, and you can buy different collections at $20 a pop. The speaker isn’t powerful, but that fits with the intention, and there is an audio port for plugging in kids’ headphones. Battery life is around seven hours, and you must charge via microUSB (you get a cable but no wall charger in the box).
For kids aged 5 to 12 years.
Small but tough, this mini speaker delivers a surprisingly punchy sound, making it one of our favorite Bluetooth speakers. It’s a good solution if you just want to choose audio on your phone and stream it for your kids. The carabiner clip is easy to attach to a backpack strap or a car seat. It’s also durable and waterproof (IP67 rated), so there’s no need to worry about spills.
Battery life is solid (up to 15 hours from a single charge), and the JBL Clip 5 has more controls (volume, play, Bluetooth, and power) than other micro speakers. It’s a shame the kid-focused JBL Jr. Pop no longer seems to be available, but this is the next best thing.
Power up with unlimited access to WIRED. Get best-in-class reporting and exclusive subscriber content that’s too important to ignore. Subscribe Today.
Source: Wired.
Filonte Vvf
Nella tarda mattinata odierna, una squadra dei vigili del fuoco della sede centrale del comando di Catanzaro è intervenuta in Via Indipendenza, nel comune di Catanzaro, per un incendio.
Che ha interessato l’unità esterna dell’impianto di refrigerazione di una cella frigorifera appartenente a un’attività commerciale dedita alla vendita di prodotti ittici.
L’unità, posizionata in prossimità dell’insegna del negozio, è stata avvolta dalle fiamme per cause in corso di accertamento.
Il tempestivo intervento dei Vigili del Fuoco ha permesso di estinguere completamente il rogo, evitando che le fiamme si propagassero all’interno dell’esercizio commerciale e limitando i danni alla sola apparecchiatura esterna e all’insegna.
Sul posto è intervenuta anche la Polizia Locale per gli adempimenti di competenza e per la gestione della viabilità, che ha subito temporanei disagi durante le operazioni di soccorso.
Non si registrano danni a persone.
Source: Irriverentemente.
TSMC initiates legal action against Wei-Jen Lo, who retired from TSMC this summer, signed a non-compete agreement, but subsequently joined Intel as Executive Vice President.
Source: Latest from Tom’s Hardware.
I used to have seven browser tabs open just to write code. ChatGPT for generating functions. Claude for code reviews. Gemini for research. Stack Overflow for edge cases. GitHub Copilot in my editor. Perplexity for documentation lookup. And somehow, despite all this AI assistance, I was slower than before.
The problem wasn’t the tools. The problem was the context switching.
Every time I moved between tools, I lost mental state. Copy-pasting code between interfaces meant reformatting. Different tools had different conversation histories, so I’d explain the same context repeatedly. My workflow looked like “AI-powered” but felt like death by a thousand tab switches.
Then I tried something different: using multiple AI models within a single interface. Not replacing all my tools with one tool, but accessing different AI capabilities without leaving my workspace.
The difference wasn’t incremental. It was transformational.
Most developers pick an AI tool and commit to it. You’re either a “ChatGPT person” or a “Claude person” or a “Copilot person.” It makes sense—learning one tool well seems smarter than juggling multiple mediocre tool proficiencies.
But here’s what that approach misses: no single AI model is best at everything.
GPT-5 excels at creative problem-solving and generating varied approaches to ambiguous problems. Claude Opus 4.1 crushes complex reasoning tasks and produces more careful, nuanced code reviews. Gemini 2.5 Pro handles research and synthesis better, particularly when you need to process large amounts of information quickly. Grok 4 brings real-time knowledge and conversational iteration.
When you commit to a single model, you’re optimizing for consistency at the expense of capability. You’re choosing to be limited by one AI’s weaknesses instead of leveraging multiple AIs’ strengths.
The developers moving fastest aren’t the ones who’ve mastered one AI tool. They’re the ones who’ve learned to orchestrate multiple models for different parts of their workflow.
Using multiple models effectively isn’t about asking the same question to five different AIs. It’s about routing different types of work to the models that handle them best.
Initial code generation: Start with speed. When you’re exploring approaches or generating boilerplate, use fast models like GPT-5 mini or Gemini 2.5 Flash. You don’t need perfect code—you need fast iteration to find the right approach. Speed matters more than depth in the exploration phase.
Code review and refinement: Shift to precision. Once you have working code, route it through Claude Sonnet 4.5 or Claude Opus 4.1 for review. These models excel at catching subtle bugs, identifying edge cases, and suggesting more robust error handling. Their caution is a liability in exploration but an asset in refinement.
Documentation and explanation: Use synthesis models. When you need to explain complex code or generate documentation, Gemini 2.5 Pro excels at synthesizing information and creating clear explanations. It handles the “translate technical complexity into human language” task better than most alternatives.
Debugging and troubleshooting: Leverage reasoning depth. When you hit a truly puzzling bug, Claude Opus 4.1’s deep reasoning capabilities help trace through complex interactions. It’s slower, but for problems where speed doesn’t matter because you’re stuck anyway, depth beats velocity.
Architecture decisions: Compare multiple perspectives. For important technical decisions, run the same question through three models simultaneously. GPT-5 gives you creative alternatives. Claude Sonnet 4.5 evaluates tradeoffs systematically. Gemini 2.5 Pro provides research-backed context. The synthesis of all three perspectives produces better decisions than any single model would.
The speed improvement from multi-model workflows isn’t theoretical. It’s measurable and dramatic.
Context preservation eliminates rework. When all your AI interactions happen in one workspace, you don’t lose context. The conversation history persists. You don’t re-explain your architecture five times. You don’t copy-paste code between tools and fix formatting. The friction just… disappears.
Model comparison reveals blind spots. When you can see how different models interpret your question side-by-side, you catch ambiguities in your own thinking. If GPT and Claude give completely different answers to the same architectural question, the problem isn’t the AI—it’s that your question was under-specified. Multi-model comparison forces clarity.
Right-tool-for-job reduces wasted time. Using Claude Opus 4.1 for simple boilerplate generation wastes time waiting for a slow model to solve a trivial problem. Using GPT-5 mini for complex code review misses bugs because the model isn’t deep enough. Multi-model access means you can route work optimally without leaving your workflow.
Parallel processing multiplies throughput. When you need multiple things—code generation, test creation, documentation, and deployment scripts—you can run them through different models simultaneously rather than sequentially. What used to take twenty minutes of serial AI queries takes five minutes of parallel processing.
The shift toward unified multi-model platforms is already happening. Instead of maintaining separate subscriptions to ChatGPT Plus, Claude Pro, and Gemini Advanced, developers are moving to platforms that provide access to all major models in one interface.
Crompt AI represents this evolution—14+ AI models accessible from a single workspace, with conversation history that persists across model switches and side-by-side comparison built in. You’re not just accessing multiple AIs; you’re accessing them in a way that preserves your workflow continuity.
The Excel Analyzer becomes valuable not just for analyzing spreadsheets, but for comparing how different models interpret the same data patterns. The Charts and Diagrams Generator lets you visualize system architectures with AI assistance, then use different models to critique and improve the design.
For documentation work, tools like the Grammar and Proofread Checker ensure technical writing is clear without losing precision. The AI Fact-Checker helps verify technical claims across documentation, particularly valuable when documenting APIs or system behavior.
When you need to quickly understand complex technical concepts, the AI Tutor adapts explanations to your level while maintaining technical accuracy. All accessible without leaving your primary workspace or losing context.
The productivity gain isn’t from any single tool—it’s from having the right tool immediately available without context switching.
Fast developers don’t just use better tools. They architect their workflows to eliminate friction at every decision point.
Reduce cognitive overhead by defaulting to patterns. Establish go-to models for common tasks: fast model for exploration, precise model for review, synthesis model for documentation. You stop deciding which tool to use and start executing based on patterns.
Build feedback loops with instant comparison. When you’re uncertain about an approach, don’t debate internally—run it past multiple models and synthesize their feedback. The five minutes you spend comparing perspectives saves hours of pursuing the wrong solution.
Use AI for the right abstraction level. Don’t ask AI to write your entire application. Use it for the parts where its speed creates leverage: generating boilerplate, suggesting architectural alternatives, reviewing error handling, writing tests. Keep the high-level design and critical logic human-driven.
Optimize for iteration speed over correctness. First draft doesn’t need to be perfect—it needs to be fast enough that you can iterate multiple times. Use fast models for rapid iteration, precise models for final refinement. The workflow is: explore quickly, refine carefully, ship confidently.
Multi-model workflows don’t just make you faster. They make you better at engineering judgment.
When you see how different models approach the same problem, you develop intuition for when creativity matters versus when precision matters. You learn to recognize the shape of problems where one approach dominates versus problems where synthesis is necessary.
You start asking better questions because you see how question framing affects answers across models. Ambiguous questions get wildly different responses; precise questions get convergent answers. The feedback is immediate and educational.
You develop a more nuanced understanding of AI capabilities and limitations. No model is magic. Each has strengths and blind spots. Knowing which model to trust for what kind of problem is a meta-skill that compounds over time.
The real test of any tool is how well it integrates into actual development workflows. Multi-model platforms succeed when they eliminate friction at integration points:
IDE integration matters. The best multi-model workflow isn’t abandoning your editor—it’s having multi-model AI accessible from within your development environment. Copy-paste should be minimal. Context should flow naturally.
Conversation continuity is non-negotiable. If switching models means losing context, you haven’t actually solved the tool-switching problem. The conversation history needs to be model-agnostic so you can shift tools without restarting from scratch.
Speed must be acceptable across models. If using the “precise” model means waiting thirty seconds per response, you won’t use it. The latency needs to be low enough that switching models doesn’t break flow state.
Mobile access extends utility. Some of the best debugging happens away from your desk—during lunch, on the commute, right before falling asleep. Having the same multi-model capabilities accessible via iOS and Android means insights aren’t lost to context switching between devices.
There’s a cost argument here that matters: maintaining separate subscriptions to ChatGPT Plus ($20/mo), Claude Pro ($20/mo), and Gemini Advanced ($20/mo) costs $60/month for access to three models, with zero integration between them.
Unified platforms like Crompt offer access to 14+ models starting at significantly lower price points—often with free tiers generous enough for serious usage. Even paid tiers that provide access to all premium models cost less than maintaining separate subscriptions while providing dramatically better workflow integration.
The ROI isn’t just cost—it’s the productivity multiplier from eliminating context switching and enabling true multi-model workflows.
If multi-model access saves you even two hours per week—a conservative estimate given the friction reduction—that’s 100+ hours annually. For a developer billing at $100+/hour, that’s $10,000+ in recovered productivity for a tool costing under $250/year.
Developers who successfully transition to multi-model workflows follow a predictable pattern:
Phase 1: Skepticism. “I’m already productive with ChatGPT. Why do I need other models?” The single-model approach feels sufficient because you don’t know what you’re missing.
Phase 2: Experimentation. You try comparing models for a specific problem. The differences in output quality and approach are immediately visible. You realize different models genuinely have different strengths.
Phase 3: Pattern Development. You start routing work based on task characteristics. Fast models for iteration. Precise models for review. Synthesis models for documentation. The patterns become intuitive.
Phase 4: Workflow Integration. Multi-model access becomes your default. You can’t imagine going back to single-model constraints because the productivity gain is too obvious. The friction of model-switching is gone.
Phase 5: Advocacy. You tell other developers about multi-model workflows because the speed difference is too dramatic not to share.
In five years, using multiple AI models won’t be a competitive advantage—it will be table stakes. The developers still constrained to single-model workflows will be at a measurable disadvantage.
But right now, we’re early enough that multi-model proficiency creates real differentiation. The developers who master orchestrating multiple models for different workflow stages are shipping faster, with higher quality, than their single-model peers.
This isn’t about being an early adopter. It’s about recognizing that the tools shape the work, and better tools enable better work.
The question isn’t whether multi-model workflows are worth adopting. The question is whether you can afford to wait until everyone else has already made the transition.
Developers work faster with multi-model tools because multi-model tools eliminate the fundamental friction of modern development: matching the right capability to the specific problem at the specific moment without context loss.
Single-model tools force you to use one AI’s approach for everything. Multi-model tools let you use the best AI for each thing. The difference compounds across hundreds of daily decisions.
The speed improvement isn’t magical. It’s architectural. You’re eliminating friction, reducing context switching, and routing work to optimal processors. It’s the same systems thinking we apply to distributed systems, now applied to our own workflows.
And just like with distributed systems, the gains from proper architecture are dramatic and measurable.
-ROHIT
Source: DEV Community.
Fortunately, we live in a time when most information can be recovered from another device or the cloud. Now, iPhone users can add another app to that list. Signal has announced that secure backups are now available for iOS devices. It follows a similar launch for Android devices in early September. The company shared the news on X and linked to the previous Android blog post for all additional information.
At the time, Signal noted that secure backups would be an opt-in feature and available in Android users’ latest beta release. The company stated this was to test the feature on a smaller scale before it came to iOS and desktop — so clearly something worked.
The end-to-end encrypted backup offers no-cost storage of a user’s text messages and up to 45 days of media. Keeping older media costs $2 per month — Signal’s first paid offering. “The reason we’re doing this is simple: media requires a lot of storage, and storing and transferring large amounts of data is expensive,” Signal said at the time. “As a nonprofit that refuses to collect or sell your data, Signal needs to cover those costs differently than other tech organizations that offer similar products but support themselves by selling ads and monetizing data.”
Anyone who opts-in will have their messages backed up each day. The secure backups use a 64-character recovery key that is available exclusively on a user’s device. It’s required to access the backup and Signal is unable to recover it.
This article originally appeared on Engadget at https://www.engadget.com/apps/signal-extends-backup-feature-to-ios-users-133006884.html?src=rss
Anthropic’s Opus 4.5 appears to be a big breakthrough that slots into Anthropic’s enterprise strategy, while ChatGPT gets new consumer features, and Meta might use Google’s TPUs
Source: Stratechery by Ben Thompson.
The holidays might still be a little way off, but that hasn’t stopped retailers from offering deals ahead of time. Black Friday falls on Nov. 28 this year, but deals are already popping up at Amazon, Best Buy, Walmart and more. And with thousands of items to shop at some of the best prices we have seen all year, we’ve gathered the best early Black Friday sales now live so that you don’t miss a thing.
Beat the Black Friday rush and enjoy your Thanksgiving weekend, and get ahead of holiday shopping now. To help you save even more time, CNET’s shopping experts have hunted down the best bargains for you. Take a look below for our top offers on TVs, headphones, laptops and more. There are even a few finds from our 2025 Readers’ Choice list.
This gigantic 98-inch smart TV offers vibrant QLED colors, stunning 4K clarity and impressive sound quality, essentially providing a home theater experience at a price that’s within reach for most people — especially with this enormous $1,500 discount. Best Buy’s top TV deals have been selling out quickly, so you may need to act fast to snag this savings while you can.
This 45mm GPS smartwatch has an AMOLED display, fitness tracking, health metrics and a battery that lasts up to 14 days per charge. Grab it while it’s back down to the lowest price we’ve seen for this model.
This product has been independently, editorially selected by our expert deals team. The company has paid for highlighted placement on this page.
More amazing early Black Friday deals
This M4 MacBook Air is the latest addition to Apple’s most lightweight lineup and our overall favorite laptop on the market right now. The basic 256GB model is on sale for $250 off, or you can get double the storage and still score a $250 discount. The 15-inch model is also on sale for those who prefer a larger display.
Our top choice for artists and colleges students on our best iPads list, this lightweight tablet has a massive 13-inch display and delivers all the power most people need for basics like streaming, reading, games and social media.
Our favorite budget tablet of 2025, this affordable device is an even better value now at half-price.
Save big on this 16-inch gaming laptop with 16GB of RAM, a GeForce RTX 5060 and 1TB SSD. While it’s not top-end, it is a very solid choice for pretty much everyone who wants to game on the go and still enjoy solid specs.
More early Black Friday laptop and tablet deals
If you want the sound quality of the AirPods Pro with the comfort and security of over-ear headphones, the AirPods Max are for you. These headphones are some of the best high-end noise-canceling headphones and a popular option for holiday gifting.
Enjoy rich Dolby Atmos sound with these on-ear noise canceling headphones. The Beats Studio Pro are our top overall pick for the best Beats headphones thanks to their comfortable design, excellent voice-calling performance and advanced spatial audio.
These Bose Ultra Open earbuds are not only supremely comfortable, but also have some of the best sound quality you could hope for, along with an amazing battery life and a durable design.
More early Black Friday headphone deals
This 65-inch Sony model is discounted by a whopping $300 right now. Its beautiful HDR10, HLG, Dolby Vision, 120Hz frame rate, four HDMI inputs and Sony’s XR Contrast Booster make it a very appealing choice for your watching experience.
Brighten up your entertainment setup with this kit that adds lights to the back of your TV. It can sync with whatever you’re watching or playing, and there are kits for multiple TV sizes on sale right now.
One of our favorite Bose speakers available, this soundbar adds a whole new dimension of sound to your TV experience. It’s easy to setup and offers fantastic audio when in use.
This sizable 2025 TCL QM6K TV has a 4K QLED display with high-energy LED chips for enhanced brightness and Dolby Atmos spatial audio for immersive audio. It also impressed CNET reviewer TY Pendlebury, who praised it for featuring “some of the best picture quality, including great black levels, for an LCD-based TV.” Other features include a gaming mode, Bluetooth connectivity and four HDMI ports.
More early Black Friday TV deals
Egohome’s early Black Friday sale is offering massive discounts on select products. Right now, you can score up to 60% off, plus an extra 20% with the CNET exclusive code EGOHOME20 at checkout. The sale runs from now through Dec. 10.
This product has been independently, editorially selected by our expert deals team. The company has paid for highlighted placement on this page.
Make ice cream, sorbet and other frozen treats at home. This excellent Ninja Swirl lets you create soft serve, milkshakes and more with ingredients you love, allowing you to customize to your heart’s content. This bundle comes with extra pints so you can prep more flavors at once, which is a fun endeavor for everyone with a sweet tooth. Normally $349, it’s down to $280 right now. And if this is your first QVC purchase, you can take another $20 off your purchase of $40 or more with code HOLIDAY20.
This product has been independently, editorially selected by our expert deals team. The company has paid for highlighted placement on this page.
This 1,024-watt-hour power station is currently half-off on Amazon. It’s a great backup power source during inclement weather or when you’re traveling off the grid. It has six AC outlets, two USB-C ports, four USB-A ports and more. You can also connect it with compatible solar panels.
The T30S Combo is tied for first in our rankings of the best overall best robot vacuums available now. It’s an all-in-one system with 11,000Pa of suction power and dual mop pads, as well as advanced object avoidance and hands-free voice control. Plus, the base station automatically empties the dustbin, cleans the mop pads and refills the water tank so there’s minimal maintenance on your end.
Keep an eye on your whole property while you’re traveling for the holidays with this six-camera system can help you. It offers two-way audio, 1080p resolution and motion alerts. It also includes a Sync Module XR for an extended range.
Get hand-picked deals from CNET shopping experts straight to your phone.
More early Black Friday home tech deals:
This year, Black Friday is Nov. 28. It always lands on the Friday after Thanksgiving, which falls on the fourth Thursday of November. There are hundreds of price cuts and sizable sales from the biggest online retailers already available.
Technically, Black Friday sales tend to begin a few weeks ahead of Thanksgiving, with the pace increasing closer to the actual Black Friday. And with Cyber Week starting two days after Black Friday, you’ll have more time to enjoy the savings from both events.
Lots of CNET readers are concerned about tariffs and how they could affect holiday shopping. It’s certainly possible that tariffs could affect Black Friday sales, as we have seen some related price hikes, including on Nintendo Switch consoles. But that doesn’t mean you can’t find great deals on tech and other items. The best way to determine if you’re getting a good deal is to track the item’s price history before making a purchase. You can do this on sites like CamelCamelCamel and Price Tracker, or by using the Keepa browser extension. We’ve rounded up several useful shopping tools for you to check out, and we’ll highlight the best deals we find throughout the event.
-SOURCE-Ryan-Waniata.jpg)
.png)