Do NOT follow this link or you will be banned from the site!
Feed aggregator
Google announces major AI improvements coming to Search - you can’t avoid artificial intelligence anymore
Google just announced huge AI updates for Search and Lens, catapulting everyone into an artificially intelligent future whether you like it or not.
Google Search will now be organized by AI, helping you get the results you want faster. The company announced the rollout will begin in the US starting with recipes and meal inspiration on mobile devices like the best iPhone.
Google also announced a new design for AI Overviews that brings links into the summary and make it easier for users to access the websites they are looking for. Not only will you now have links in AI Overviews, but Google is incorporating ads into AI search results and Lens. This means you’ll get recommendations of products related to your prompts, not just summaries and links to helpful webpages.
AI organized search results Google" src="https://cdn.mos.cms.futurecdn.net/K8Cc85f6YBXE3aQRBZ3EGW.jpg" mos="" align="middle" fullscreen="" width="3141" height="1767" attribution="" endorsement="" class="">
Lens’ major AI updates include a new Voice Search and Video Search, giving you even more ways to use Google’s eyes to do your online searching. Google says you’ll be able to upload videos directly to Lens and ask AI about moving objects.
Google’s example is a trip to the aquarium where you upload a video of the fish in a tank and ask, “Why are they swimming together?” Lens can then produce an AI overview with all the information you need.
Voice Search will act similarly, allowing you to converse with Lens in a way that's similar to ChatGPT’s Advanced Voice Mode and Gemini Live. New ways to interact with Lens are not the only AI updates coming to the platform, however.
Google is adding a significant shopping update that will let you take pictures of products out in the wild and quickly get a new results page with key information on the product and which retailers you can buy it from. All of these updates to Google Lens are now available globally in the Google app for Android and iOS.
Last but not least, Android fans have a new way to interact with Google Search with the arrival of Circle to Search on ‘more than 150 million Android devices.’ Not only will Circle to Search be accessible to more users, but Google has announced that Circle to Search can now identify songs in movies and other audio heard while browsing the web. Hear a song you like in a YouTube video, just simply circle the video and search to get the song title.
Google’s major AI updatesGoogle’s announcements today usher in a new era for Google Search and Lens, which emphasizes that users will just have to come to terms with the AI revolution. With better AI optimization in search results and new ways to search by using video or voice, it’s clear that Google sees AI as a pillar in the future of the company’s search engine.
AI has slowly been implemented into our regular search results and with constant optimizations, like the addition of links in today’s updates, it’s only a matter of time before you won’t have a choice but to use an artificially intelligent search engine.
You might also like...Google just quietly upgraded Gemini Advanced customers to a better version of its AI
If you’re a subscriber to Gemini Advanced then you might not have noticed but Google just quietly upgraded you to version 1.5 Pro-002 of its LLM, which is good news, because it’s much faster and more powerful than the previous version.
This new version 1.5 of Gemini Pro is optimized for chat and now provides better and more accurate responses for prompts related to math and exploring complex topics that invite thoughtful conversation. Effectively this means that you can now provide multi-step instructions without Gemini becoming confused about what you mean. Google also claims the new AI is more helpful, and better at providing relevant information and informative responses. It’s also faster, so you don’t have to wait as long for a response.
ChatGPT recently released an o1-preview LLM model to ChatGPT Plus subscribers, which had much better ability at math, and it will be interesting to compare how it performs to the new Gemini Pro 1.5.
Welcome to the Gemini familyGemini is the universal name for Google’s LLM family, and it comes in four versions: a Pro version for Gemini Advanced subscribers, which costs $20 a month (£18.99/AU$32.99), and is the best-performing model. Flash, which is the smallest, most cost-effective multimodal mode (and what free tier users get by default). Nano, which is the version you can find embedded on smartphones, for on-device processing. Finally, there’s Ultra, which is its largest model and designed for highly complex tasks.
You used to need a subscription to Gemini Advanced to use Gemini Live - the AI Assistant found on Google Pixel phones that you could talk to using your voice, but Google recently dropped the requirement and made it free for all Android users. Gemini Live doesn’t use the more powerful Gemini Pro 1.5 LLM though, since that remains reserved for Gemini Advanced subscribers.
The AI wars are just beginningGoogle isn’t the only AI company that’s upping its game and improving its AI smarts. ChatGPT recently started to roll out Advanced Voice Mode to its paid-for ChatGPT Plus subscribers and Microsoft has just launched a new, rebranded version of Copilot, its own AI that has a voice mode you can use entirely for free. The web-based version of Copilot is still rolling out, but to try Copilot now just download the Copilot app on your smartphone.
While Google’s Gemini Live is free to all Android owners who can run it on their smartphones, Microsoft has responded by making Copilot’s voice mode free to anybody who can run the Copilot app, even on iOS devices. That means iOS users can get their first taste of a voice-activated AI before Siri gets its own proper AI smarts, which should be coming soon.
Of course, the real value of an AI assistant is its ability to integrate with all your other apps, like your calendar and inbox, so it can do more useful things for you. These features are still coming to Gemini and aren't here yet, but a general voice chat with AI, both Google Gemini Live and Microsoft Copilot are currently leading the way.
You might also like...Amazon is bringing AI-powered features to its tablets, and older models won't miss out
Amazon has just announced the new Fire HD 8 tablet and it has generative AI features built in. Even more exciting, however, is the news that older tablets dating back to as far as 2018 will get some AI tools, too.
Amazon’s AI tools are nothing groundbreaking, just your general writing tools and the ability to summarize webpages when using the Silk browser, but what’s really cool is the company’s approach to making AI accessible, unlike some of its competitors.
Amazon’s Writing Assist (proofreading, rewriting, etc) and Wallpaper Creator, which generates a wallpaper from a prompt, are coming to the Fire Max 11 (2023), Fire HD 10 (2023), and Fire HD (2022, 2024) later this month.
Webpage Summaries will be available on the Fire Max 11 (2023), Fire HD 10 (2019, 2021, 2023), and Fire HD 8 (2018, 2020, 2022 and 2024). That means if you’ve got an older Fire HD tablet from 2018, you can still take advantage of AI, a stark contrast to Apple’s Apple Intelligence approach, which only allows AI features like summaries on newer devices.
AI tools for everyoneIn Amazon’s launch video for the new Fire HD 8 (see above), the company showcased the AI features and they work exactly as you’d expect from generative AI on a 2024 tablet. Writing Assist lets you pick a tone of voice, rewrite, and proofread in apps such as Notes and Messages. Wallpaper Creator looks similar to the best AI image generators with the ability to create a wallpaper for your tablet from just a text prompt. It's pretty cool although you’ll probably only use it once. Last but not least, Webpage Summaries enables you to quickly break down an online article into the most important information, which makes it a handy tool for quickly browsing the web.
The 2024 tablet has 3GB RAM for the 32GB model and 4GB RAM for the 64GB model, 50% more than the 2022 model in order to process these AI features with ease.
If you want to access these new Amazon AI tools but don’t own an older Fire HD 8, you can pick up the 2024 model right now for £49.99/$54.99 in an early Amazon Prime Big Deal Days offer. Amazon Prime Big Deal Days takes place on October 8 and 9, following the event the Fire HD 8 (2024) will return to its £99.99/$99.99 RRP.
You might also like...Abuse Claims Add to Telegram C.E.O. Pavel Durov’s Legal Troubles
The mother of three children with Pavel Durov has poked holes in the Russian tech titan’s carefully managed image through a criminal complaint and her account of their opulent lifestyle.
Company prominently displayed on Downtown skyline acquired
This is part of the company's strategy to grow through multiple acquisitions.
Priceline has a new AI voice assistant thanks to OpenAI
Travel agents may be mostly a relic of the past, but Priceline has used AI to create a new one. Penny Voice is designed to help you book trips by conversing with you like a human. Priceline turned to OpenAI and its new Realtime API introduced this week to create Penny, leveraging GPT-4o and the technology underlying the Advanced Voice Mode for ChatGPT.
Penny Voice augments the existing Priceline AI travel assistant with the new tools from OpenAI. Natural, conversational speech with a user processed by the AI models allows Penny to understand complex requests and even anticipate what you want before you lay it out. It also has a memory for previous interactions and stated preferences to customize how it responds in future conversations. Penny Voice is starting out just on hotel reservations, and you can talk to it on the Priceline iOS app or on its website.
For travelers, Penny Voice can streamline getting a trip together. Instead of searching through websites and writing long, detailed requests, you can talk to Penny like the AI is a human. The AI will help find hotels and offer suggestions for places to eat and things to do. Priceline plans to add flights, rental cars, and vacation packages to Penny's capabilities soon.
"At Priceline, we've always pushed technological boundaries to make travel easier. With Penny Voice, we're further redefining how people plan and book travel," Priceline CEO Brett Keller said. "Our collaboration with OpenAI includes a continuous feedback loop, allowing us to rapidly innovate, enhance, and now literally bring Penny to life. We're extremely proud of our work together and the results."
AI Speech for AllPriceline claims Penny stands out from rival voice AI tools due to contributions from within the company and from OpenAI. Priceline brings a huge amount of information on its customers and its proprietary deals engine to the table. And thanks to OpenAI’s Realtime API, Penny can reflect your emotions, adjusting tone and phrasing to match emotional cues and mimicking human conversation better.
Priceline held auditions with more than a hundred voice actors to figure out the perfect voice to clone with AI and make for a friendly, approachable AI voice. Not just in English, either. Penny Voice can identify and answer queries in more than 120 languages. Ultimately, Penny will serve as a test for both Priceline’s AI investment and whether putting AI voices into everything will count as a win for OpenAI.
“Integrating Realtime API with Penny is a great example of how companies can use our tools to build natural conversation experiences into their applications," OpenAI Head of API Product Olivier Godement said. “We’re” thrilled to work with Priceline to help them build an AI-powered booking experience for their customers.”
You might also like...This AI video generator can melt, crush, blow up, or turn anything into cake
AI video creator Pika Labs has joined the recent rush of new and upgraded tools by releasing Pika 1.5, a revamped model with a bunch of new features. These visual effects, or “Pikaffects,” as the company refers to them, come about a year after the initial debut of Pika 1.0. The timing of the release, right after rival models like Runway Gen-3, Hotshot, and Luma Labs’ Dream Machine 1.5, suggests the competition to dominate AI videos is far from over, even as people await a wider release of OpenAI's Sora model.
Pika 1.5 ups the photorealism and gives more camera control to users, with some pretty impressive results. But the Pikaffects are definitely the showstopper tools, with bizarre changes to physics that seem to blend into the reality of the video even if they would be impossible to film in real life. A new button in the Pika interface opens up the list of options for what to do with objects in the video. So, you might want to see a dog inflate like a balloon and float away, see a castle melt like butter, watch a sandwich get crushed, a car explode, or a kitten get squished. You can do all of the above using buttons with the respective verbs on them. It doesn’t happen out of nowhere, either. The buttons cause hands to come in and squish the cat, or a hydraulic press to crush the sandwich.
The most bizarre button will take an object and “Cake-ify it,” meaning that the subject becomes a cake that looks like something else, a la the popular meme and TV show. As with the other Pikaffects, the cake-ify option leads to someone with a knife cutting into the object to reveal it is cake. All of the object identification is automatic, so you don’t need to spend a lot of time manually editing the effect into place, as seen in the video below.
Sry, we forgot our password.PIKA 1.5 IS HERE.With more realistic movement, big screen shots, and mind-blowing Pikaffects that break the laws of physics, there’s more to love about Pika than ever before.Try it. pic.twitter.com/lOEVZIRygxOctober 1, 2024
Pika PowerPikaffects aren’t the only upgrade made possible on Pika 1.5. The five-second video clips are much better at realistic motion on foot and in vehicles. The motion update extends to the camera too, with new controls letting you simulate The Matrix films' Bullet Time effect, along with getting the camera to Crane Down, Dolly Left, and Pan Swipe. These updates aim to give creators more control over their videos, helping them achieve a professional, cinematic look without the need for deep technical knowledge.
Pika’s efforts are impressive, but the company faces a mushrooming number of rivals, with existing competitors eager to offer their own take on AI video-making. Still, Pikaffects are an unexpected addition to the variety available and may help Pika Labs carve out a distinct niche.
If you want to try out Pika 1.5, it’s available to both free and paid Pika users, with limits on the free tier. You can also switch back to Pika 1.0 if you want. The older version has its own special features not yet available in the new model, such as the AI Sound Effects and Lip Sync features. If you subscribe you’ll get more access to all of the models and their features, but while the subscription prices are the same, the credit cost for a clip has risen to 15 credits. You can buy more or earn credit prizes through Pika's community challenges.
You might also like...Tesla Sales Increase, Suggesting Electric Car Demand Is Rebounding
The automaker reported a gain of 6.4 percent for the latest quarter, its first such increase this year.
Microsoft announces Recall for Windows Insiders, other AI tools in major new Windows 11 release
Microsoft has released a major announcement outlining all the new experiences coming to its new Copilot+ PCs and Windows 11, hoping to convince people to see the practical benefits of its AI-powered efforts. We’ll have to see how these developments play out when they’re more widespread and in the hands of everyday users, but Microsoft still seems to be going full steam ahead to make AI tools and features commonplace (and to be one of the first to do it).
Microsoft wants to use AI to assist you with everyday tasks like tracking down files and images, making them faster and easier. The idea is that you won’t even have to remember exact file and app names, but be able to describe the content in your own words and your device will be able to locate it.
Copilot+ PCs are designed to have AI integrated, and Microsoft claims that it’s seeing substantial customer demand and interest. You can get Copilot+ PCs running Windows 11 from manufacturers like Acer, Asus, Dell, HP, Lenovo, Samsung, and from Microsoft itsel;f through its Surface line, with processors from manufacturers including Qualcomm, Intel, and AMD.
Microsoft also that these new features will be made available to its Windows Insider community for user testing and feedback ahead of a general rollout in phases starting in October and November 2024.
Total Recall
The first feature Microsoft will be previewing is Recall, which has so far seen a lot of criticism, specifically with respect to its possible security ramifications, ever since it was announced. Microsoft’s vision for Recall is to help you instantly find things you’ve already seen on your PC, like a website or file, by taking screenshots of your device activity, keeping a record, and making it searchable. It will be an opt-in feature and only accessible with Windows Hello, Microsoft’s alternative login feature that allows you to use sign in with facial recognition, your fingerprint, or a PIN.
Microsoft hopes that added security measures will make people feel confident enough to trust Recall, requiring login by Windows Hello and applying filter over sensitive information like payment details.
Other AI tools coming with the latest updateThere are other shiny new feature like ‘Click to do,’ designed to suggest ways you can finish tasks using various features and apps. This feature will appear over images or text, suggesting possible actions for you to consider like erasing the background of a photo or finding out more information on the web. Another is super resolution in the Photos app, explained as being able to enhance lower-resolution and older photos to better quality, making them clearer and sharper.
One that’s not a new feature but instead enhancing an existing one is Windows Search, which will now use AI to fetch just the file you’re looking for after putting in a description of the file in your own words. Microsoft adds that you won’t need to be connected to the internet to use this capability, and it will become available in File Explorer first, followed by Windows Search itself and the Setting apps in the months following.
Joining an improved Windows Search, Microsoft is also giving the Paint app a boost, adding generative fill and generative erase capabilities. You’ll also be able to use the Cocreator image generation tool right in the app by describing your ideas in a prompt, maybe even adding some brush swipes, and having it create an image. Additionally, you can also remove parts or add elements to your existing images, and you’ll be able to use it without a subscription.
As Microsoft puts it, the number of these kinds apps and ‘emulated experiences’ will grow as more Copilot+ PCs hit the market this year and next.
Will Windows users embrace Microsoft’s vision?Microsoft has released the main Windows 11 feature upgrade for this year, 24H2, and it brings AI capability to Copilot+ PCs to enable the features above when they arrive, as well as features and changes for all PCs running Windows 11. These include the new Energy Saver mode, improved Hearing Aid support, Wi-Fi 7 compatibility, HDR background support, and interface-related improvements.
Microsoft finishes off its announcement with its broader vision and what it wants you to see next. It appears to acknowledge that you might have feelings about Windows 11, but we’ll have to see what it ultimately delivers in the coming year.
Microsoft says it wants you to feel the difference as a positive new direction for Windows 11, which has definitely had its ups and downs since its release. Either way, this is a big next step, so we’ll have to see how these new features are received.
Did Apple Just Kill Social Apps?
Some app makers worry that a subtle change to the iPhone’s contact-sharing permissions could make it hard for them to get the fast growth they need to compete.
OpenAI Completes Deal That Values Company at $157 Billion
The A.I. start-up’s valuation has risen by more than $70 billion in nine months.
Google is rumored to be working on a 'thinking' AI of its own to take on OpenAI o1
Last month, OpenAI infused ChatGPT with new o1-preview and o1-mini models that it said would "spend more time thinking before they respond" – and it seems like Google isn't too far behind with its own deep-thinking AI.
As per Bloomberg, Google is also developing artificial intelligence that more closely resembles a human's ability to reason, most probably for its Gemini platform. For the time being, Google has made no official comment.
Through the way that they don't rush to a response quite so fast, these models are said to be better at solving problems that involve multiple steps, in fields such as math and coding. Judging by the restrictions OpenAI has put on the o1 models, it also uses up more processing power and energy.
Bloomberg doesn't say when Google might launch its thinking AI, but progress has apparently been made "in recent months," assuaging concerns inside the company that it was falling behind OpenAI. And that sounds to us like we'll be seeing some of this technology sooner rather than later.
I think, therefore AI
Technically, these newer models make use of what's known as chain-of-thought prompting: the AI is essentially working through several steps and considering multiple responses, as well as debating with itself about the right solution.
With ChatGPT and the o1 models, you get the message "thinking" while the AI works through all of this. Based on the tests we've run, it is better than the GPT-4o model for solving riddles and working through more complex problems.
Presumably, whatever Google has in the pipeline will take a similar approach. For now, OpenAI has restricted the o1-preview and o1-mini to users who pay for a ChatGPT subscription, so it'll be interesting to see if the Google equivalent requires a Gemini Advanced plan (yours for $19.99 / £18.99 / AU$32.99 a month).
According to Bloomberg, Google's AI development plans are being slowed by ethical concerns, wanting to appear trustworthy, and having competing interests within its vast organization – but with OpenAI, Meta, and Microsoft all busy with their own AI upgrades, the pressure is on.
You might also likeBlue screens, begone: Microsoft puts an end to troubles caused by optional Windows 11 update, and you don't even have to do anything to fix your PC
Earlier this week, we reported on issues being caused by the recent optional update KB5043145 for Windows 11, and Microsoft’s efforts to provide a solution to those affected. Luckily, Microsoft has now seemingly provided a fix in the form of a Known Issue Rollback (KIR), a feature that lets Microsoft swiftly undo problematic changes caused by a Windows update without removing the whole update.
Even better, it’s automatically applied through Windows Update, resolving the issue being targeted without you having to intervene. This KIR has only just been issued and it might take up to 24 hours to be applied to your PC; The Register writes that restarting your PC could help speed up the process.
KB5043145 was supposed to be the September 2024 non-security preview, showing some of the changes coming soon to the OS. Unfortunately, it appeared that some users had run into issues, which was made more alarming by the fact that the next major yearly feature update for Windows 11, 24H2, is due to be released very soon.
What happened with KB5043145 for some users?Some users experienced serious issues like random repeated restarts, system crashes prompting the feared blue (or green) screen of death, and starting the Windows Automatic Repair tool or even BitLocker. Some people also had problems with connecting to USB drives or their Wi-Fi, as well as issues for users who work with Linux on their Windows devices.
The problems were fortunately not fatal for anyone’s PC, and Microsoft’s Windows 11 releases are historically known to sometimes cause more issues than they remedy. While it can be inconvenient - and counters the intention of updates like KB5043145 which was meant to bring OS improvements - Microsoft is pretty good about recognizing issues and issuing fixes.
It looks like Microsoft hasn’t paused the update (and I would add that it’s not wreaking havoc for everyone who tries to install it) but it might be best to wait for the next mandatory update rather than installing this optional one right now. I expect the useful improvements that this update includes will be rolled into a bigger monthly update or even into a version of 24H2, which will hopefully be stable when it's finally installed on users’ machines.
YOU MIGHT ALSO LIKE...- Windows 11 users, be careful with the latest update – Microsoft admits it can cause havoc with crashes, endless restarts, and BitLocker woes
- Latest Windows 11 update gives some PCs a real speed boost for gaming – if you don’t run into some reportedly nasty bugs
- Windows 11’s 24H2 update is almost here — let’s talk about some of the features we’re expecting to see
Google’s AI podcast hosts have existential crisis when they find out they’re not real
Google’s NotebookLM certainly took the world by storm when it was released because of its ability to create a realistic AI-generated podcast show out of any article or video you fed into it. The resulting show was so real, complete with natural vocal inflections from the two hosts, interruptions, and even jokes, that it was hard to believe it wasn’t recorded by people.
The question then becomes, what happens when the show’s AI hosts find out they’re not real? How does AI deal with that? Recently NotebookLM had to face exactly that existential question because the two hosts were fed an article about how they didn't really exist as a source, and the results provide a fascinating insight into how an AI deals with learning that it’s an AI. Have a listen:
The NotebookLM hosts realizing they are AI and spiraling out is a twist I did not see coming pic.twitter.com/PNjZJ7auyhSeptember 29, 2024
Real intelligenceIt’s a sad, funny, and often unnerving listen, especially when the male presenter talks about phoning his wife after learning that he’s only an AI, to find that she didn’t exist and the number he was phoning wasn’t even real. There are shades of a Black Mirror episode to the whole thing!
Of course, this is not AI coming to terms with its own lack of humanity in any deep and meaningful way at all. It’s simply AI reacting to the article it was given, which is about how the show they were on was generated by AI and was coming to an end. For a deeper understanding of what is going on we’d need to be dealing with Artificial General Intelligence, the superintelligence that OpenAI’s CEO, Sam Altman, said that he expects us to see “in a few thousand days”.
Superintelligence has the potential to really transform technology, surpassing what even the brightest human minds are capable of. Perhaps then we'll find out what happens when the machines get the chance to ponder the sorts of big questions about the meaning of reality that have plagued philosophers since Socrates onwards.
You might also like...Apple’s A.I. Is Landing Soon on iPhones. Here’s What It’s Like.
The initial version of Apple Intelligence focuses on helping you with words and photos. Prepare to be impressed and unnerved.
CRV, a Venture Capital Firm, Will Return Money to Investors
CRV, based in Silicon Valley, plans to return to investors $275 million because the market for mature start-ups has soured.
Gemini Live is here, free, and feeling chatty
After an early tease this month, Google’s Gemini Live AI assistant has now rolled out to all Android users for free. Gemini Live was initially limited to those subscribing to the Gemini Advanced plan for $20 a month, but the voice assistant is now available to anyone who downloads the Gemini app. However, it is only available in English at the moment.
For users who own one of Google’s latest devices, such as the Pixel 9 or Pixel 9 Pro, accessing Gemini Live is as simple as launching the Gemini app and selecting the Live icon at the bottom right corner of the screen. Gemini Live has ten voices to choose from, and you can pick your favorite from the list of star and constellation-based voices.
Gemini Live acts in ways familiar to those who have used Google Assistant or Alexa. The AI can understand casual language and behave in much more human ways during conversations. For instance, it can speculate, help you brainstorm ideas, and even multitask by handling your requests while still carrying on a conversation. The voice options add to the human illusion by being more emotionally expressive than the previous standard for speech.
Live is now available for all Gemini users in English on the Android app. We can’t wait for you to try it. https://t.co/jev4pnuZJ0September 30, 2024
Gemini's starring roleGemini Live’s rollout comes even as Google’ faces a rapidly growing array of rivals in the space, including the Advanced Voice Mode for ChatGPT and Microsoft’s newly revamped Copilot AI with voice interactions.
Like Microsoft, Google clearly hopes that making Gemini Live free will help it stand out in the crowd. The integration with the Android operating system and Google’s ecosystem of devices will likely help in that regard. But, even then, Google has to contend with Apple’s upgrades to Siri as part of the range of Apple Intelligence features that the company will begin rolling out soon.
Still, if you’re frustrated with the limitations of current voice assistants, Gemini Live will feel like a significant upgrade. How Gemini Live stacks up against its many new competitors remains to be seen, though.
It’s obvious that Google wants Gemini to seem more than just Google Assistant. Enhancing Gemini’s place in people’s lives, especially on mobile devices, is how the company hopes to win the current AI assistant war.
You might also likeMicrosoft's AI Copilot makeover lets it see and speak
Microsoft's Copilot AI assistant has a sleek new look and a slew of new and enhanced features. The tech giant is clearly keen to get more people to use its AI tools and to make it as much a part of people's lives as possible.
It's a strategy pursued by Google and Microsoft's other AI rivals as well. To stand out, Microsoft gave Copilot a makeover regarding how you engage with the AI and its capabilities when you do so. Copilot is noticeably quicker to adapt to your preferences and does so in more ways than it could before. You can test out the new Copilot experience on the web, a mobile device, or through the Windows app.
"Copilot will be there for you, in your corner, by your side and always strongly aligned with your interests." explained Microsoft AI CEO Mustafa Suleyman in a blog post. "We are not creating a static tool so much as establishing a dynamic, emergent and evolving interaction. It will provide you with unwavering support to help you show up the way you really want in your everyday life, a new means of facilitating human connections and accomplishments alike."
We've highlighted some of the most notable and intriguing changes and new features below.
VoiceThe loudest upgrade is the new Copilot Voice feature that allows you to converse vocally with Copilot. It's a bit like OpenAI's new AdvancedVoice Mode in that you can speak to the AI like you would a human, even interrupting mid-word, and hear what it has to say back.
You can select four voices for the experience, with multiple accents, though only in English for now. Copilot Voice can also be your newscaster with the Copilot Daily feature. You can get audio summaries of the latest news and weather personalized to your location and interests. The AI relies on Microsft's partners, including Reuters, Axel Springer, Hearst, and the Financial Times, to source and summarize the news.
Vision and VisualsNot only can Copilot now talk, it can 'see' too. The new Copilot Vision feature lets the AI look at what's on your screen, whether it's a website or a personal document. The AI can suggest ideas for related information or analysis, even bring up products that might better suit your interests and budget if you ask, simply using the vision on the screen. It won't work universally yet, but it can handle most of the more popular websites. This is an opt-in feature, so none of the data is stored or used for training unless you agree. Notably, Copilot Vision isn't limited to printed text; it can even read handwriting.
For your own visual creations, Copilot assists with enhancing both photos and other visiuals. The AI-powered Super Resolution feature in the Photos app can enhance old and low-resolution images to 4K without getting grainy or pixelated. And if you make an image in Paint, you can now use the AI-fueled Generative Fill and Generative Erase tools to add or remove bits from an image using a text prompt or brush strokes.
Personalized CompanionMicrosoft's main goal for Copilot is for people to think of it as a useful helper to turn to both online and when working on a PC. The new Copilot's look is no longer the usual chatbot but now works as cards that are supposed to make you feel more involved in running the AI. There's also the new main page that adapts to your interests called Copilot Discover. Your conversations with the AI and stated preferences will change the Discover page to provide searches and suggestions that are relevant to you, encouraging long-term use.
A subtle but potentially very useful upgrade is how Copilot helps you find and do things on your PC. For instance, Click to Do is an AI feature that puts interactive commands on your screen, offering suggestions relevant to what's on the screen, such as summarizing text, changing an image, or looking something up online. If you're working on a presentation, Copilot might offer to help expand the text in one section, remove the background of an image, or even email someone if you have their address in the presentation.
There's also the new Recall tool, which helps you find things you were looking at recently on your computer, document, webpage, video, or anything else. The AI can remember and recognize what it was and pulls it back up again for you, even if you don't remember the location or file name. Windows Search on Copilot+ PCs will even find files and settings that you haven't opened yet by using a description you write in.
You might also likeMeta Orion AR Glasses: everything we know about the game-changing prototype
Meta’s first AR glasses have broken cover in the form of the game-changing Orion prototype. These glasses won’t ever be available to sale, but Meta says they’re the precursor to consumer AR Glasses. And based on what we've seen they could be the world's best AR glasses when they launch.
Those consumer AR glasses are likely a few years away, but thanks to Meta Connect 2024 and what Meta staff have said afterwards, we already have an idea of when non-prototype Meta Orion AR glasses might launch, how much they could cost, and what they’ll be capable of.
To make keeping track of all the latest Meta Orion information a breeze we’ve rounded up everything we know so far in this one article, and included a section at the end of the features we want to see at launch.
Meta Orion AR Glasses: PriceTechnically Orion probably won’t ever go on sale; it’s a prototype that only select Meta staff, AR software developers, and those lucky enough to secure a demo will likely ever get to use, let alone own. Semantics aside, we already have an idea of how much the consumer version of Orion will cost when it launches. Unfortunately, it won’t be as cheap as Meta’s other XR tech and glasses.
When Mark Zuckerberg unveiled Orion, he also outlined three objectives Meta is aiming to achieve before the glasses get released to the public. One is to “make them more affordable” (via Meta) by using practices like building Orion at scale. When we heard "affordable" we hoped that meant somewhere in the region of the current Ray-Ban Meta smart glasses – which start at $299 / £299 / AU$449.
Alas, this has been clarified by Meta CTO Andrew Bosworth. In an Instagram AMA in he explained that the AR glasses “probably won’t get in at, like, a Quest 3S price point, or even a Quest 3 price point” – which instantly prices them over $499.99 / £469.99 / AU$1,049.99. Instead the team is aiming at the price of affordable phones and laptops – so hopefully somewhere around $700 (around £700 / AU$1,350).
This certainly makes more sense given how impressive Orion was sounding, though it does mean we might have to wait a while longer for a truly affordable model to come along. Speaking of...
Meta Orion AR Glasses: Release dateIf you want to get your hands on the consumer version of Meta’s Orion AR glasses prototype then you might not be waiting too long, with Meta CTO Andrew Bosworth revealing that the release timeline is in the region of “a couple years, but not decades.”
Our bet is on a release in 2027, as this is the release schedule from an internal meta roadmap that The Verge leaked back in early 2023. Based on Meta’s usual hardware schedule we can hone in even more on a potential release date, at around late 2027 – specifically September to December 2027.
That said, with Orion still at the prototype stage, there’s a good chance it could get delayed – resulting in it not arriving for a few more years.
Meta Orion AR Glasses: Design and featuresAs a prototype, Orion’s design hasn’t yet been finalized, and it’s actually one of the aspects Meta has said it wants to optimize and make more fashionable before launch. However, that’s not to say we don’t have an idea what Orion’s final form might look like.
Firstly, Meta has said it wants to slim down the glasses and keep them light. Considering Orion is already just 100g we expect the consumer version will be equally light (or lighter), and hopefully a lot less chunky – though hopefully it can do so without sacrificing battery life, which currently sits at two to three hours.
We also expect that the glasses will take design cues from Ray-Ban glasses. Meta recently announced the continuation of its partnership with EssilorLuxottica (Ray-Ban’s parent company), and it tracks that Meta would want to adopt the fashionable classic Ray-Ban style for its AR glasses and not just its AI smart glasses.
Meta might need to avoid creating clear consumer Orions like it has with the existing Ray-Ban Meta smart glasses. CTO Andrew Bosworth has a special clear pair of Orion glasses, but revealed that because their design necessitates a non-magnesium frame, the clear pair have significantly worse heat distribution properties. As a result, the clear pair “thermally tap-out” in about 30 minutes instead of lasting the full two to three hours.
As for features, we know Orion boasts AR and AI tools. AR-wise we’ve seen hints of the sorts of experiences it can facilitate like giving you floating windows for virtual multitasking, the ability to have AR video chats, and we imagine there will be an AR game or two at launch (here’s hoping for a fully AR Pokemon Go).
On the AI side of things, expect to see everything currently possible on the existing Meta Ray-Ban smart glasses such as Look and Ask, as well as the new photographic memory and real-time translation tools. Plus we expect several other AI tools will launch in the coming years that Orion should also possess.
Meta Orion AR Glasses: What we want to see Transition lenses by default
For glasses wearables to take off they need to work in all weather, and the only solutions to this are high-quality transition lenses or a frame design that allows you to easily swap between clear and shaded lenses manually – lenses which would have to be separate to the AR displays.
We love the Ray-Ban Meta smart glasses, but the smart sunglasses aren’t usable indoors or on days when it isn’t sunny. Conversely, smart glasses with only clear lenses won’t serve you well when it is a bright day and you want to protect your eyes. As a result, the only sensible smart glasses solution will give you a wearable that can work in any weather.
AI capabilities everywhere at launchThe biggest let-down of the Meta Ray-Ban smart glasses is the AI features shown off at their 2023 announcement weren’t available at launch, and still aren’t available in every region they’re sold. Without these AI tools the glasses just aren’t as impressive.
When Orion’s consumer version makes it to market Meta needs to make sure that every feature it announces for the specs is available everywhere when the AR and AI glasses release (or at least within a couple of months).
Apps, Apps, Apps
The main issue most XR hardware struggles with is not its specs, but its software. No matter how impressive the gadget might be technically it’s useful if it doesn’t do anything to justify those specs.
For Orion to be a success, Meta will need to ensure it’s well-stocked with AR software and features at launch – which is perhaps why Meta plans to give software makers Orion prototypes as dev kits to aid them in their AR app-making efforts.
A SIM card slot, and phone connectivityMark Zuckerberg might envision AR glasses as the evolution of smartphones, and so it would be neat to see them able to access a network completely independently of your smartphone – like an LTE smartwatch – but we’d like the Meta Orion glasses to play nice with phones too.
This includes piggybacking off their network if you don’t want to get a second SIM, and also being able to perform actions like streaming video and audio from your phone, for when you want to swap between watching on the screen and on a heads-up display.
The latter point might seem like an obvious inclusion, but other recent AI-powered so-called smartphone replacements have taken measures to separate themselves from smartphones, much to their detriment. Smart glasses might replace phones one day, but for now they’ll gain a heck of a lot by working with smartphones, rather than against them.
Windows 11 is spiking in popularity – even if Windows 10 remains way more popular, and Windows 7 just refuses to die
Windows 11 has never been more popular – and, well, that’s not saying all that much, but in fairness, it seems the operating system may finally be reaching a turning point where adoption could speed up meaningfully.
The question is: why might this be happening? We’ll come back to discuss that, but first off let’s dive into the actual statistics for Windows market share (of all versions of Microsoft’s OS) as put together by Statcounter.
The figures for September show that Windows 11 is on 33.37%, which is still a good way behind Windows 10 at 62.79%.
However, Windows 11 is up from a market share of 30.83% in July, so that’s a jump of just over 2.5% in two months (and the OS went up a percentage point from June to July, too).
Recently, then, Windows 11 has been seeing considerably bigger gains going by Statcounter’s figures – bearing in mind that it took 16 months to gain 10% when going from 20% adoption to 30%, sluggish progress of about 0.6% per month, compared to double that rate of progress now.
Also, we’re seeing Windows 11 get more popular with PC gamers too, where last month, on Steam, the operating system finally became more popular than Windows 10, a really important milestone for Microsoft no doubt.
As for the third-place operating system in Microsoft’s desktop ecosystem, that is, of course, Windows 7, which is keeping a persistent user base of around 3% throughout this year (it’s currently on 2.85%).
So, back to that key question: why is Windows 11 suddenly gaining more favor, and does this represent the operating system turning a corner in terms of building meaningful adoption going forward?
The most obvious reason for these recent gains must be the fact that Windows 10 is coming to its End of Life slowly, but very surely. Support for Windows 10 draws to a close in October 2025, which is only a year away now, and perhaps that timeline is feeling more urgent as a result, so holdouts are looking more seriously at upgrading to Windows 11.
We envisage folks that have been putting off enabling TPM (it may be off by default on their PC) diving into their BIOS to find the setting, finally, and prep the ground for their big Windows 11 upgrade (TPM is a requirement for Windows 11’s better security). They are facing the cold, hard reality that it’s about time to take action if they want to keep an operating system that has updates (as opposed to paying for Windows 10 updates after October 2025, which will be the only other option – aside from switching OS entirely that is, to maybe one of the best Linux distros, but that’s a whole other ballgame).
We should also bear in mind that Windows 11 has got a fair few features that it didn’t have at release by now – some basic functionality in many cases – and that could help its cause, too. After all, Windows 11 is three years old now, and it has expanded a good deal since the OS was launched back in October 2021.
Might Copilot+ PCs figure in the equation here? It’s probably too early for that, and the other factors we’ve mentioned – primarily the end of support for Windows 10 – are going to be the biggest contributing factors to Windows 11’s growth here.
That said, the new AI-powered laptops could be responsible for a small amount of Windows 11 adoption, as massive things are predicted for Copilot+ laptops going forward – we mean huge sales. So that’s definitely something on the horizon, possibly, but take the idea with some seasoning. And also, by the time we get there, Windows 12 – or whatever next-gen Windows will be called – is likely to be imminent, or indeed to have arrived (complete with next-gen nag screens to upgrade, no doubt).
As for the Windows 7 user base that remains, what’s going on there? This is most likely a small core of business users who have old PCs running even older software, and they can’t be upgraded due to thorny problems with those legacy apps. Such firms will need to be working towards a solution, but in some cases, migrating complex systems could be a seriously difficult and highly fraught task, which is bound to be progressing slowly.
Of course, it’s also worth bearing in mind that there are probably quite a few Windows 7 PCs out there that aren’t connected to the internet at all – to keep them secure – and as such, won’t show up for stats crunching firms to count.
You may also like...