Do NOT follow this link or you will be banned from the site!
Feed aggregator
Elwood Edwards, Voice of AOL’s ‘You’ve Got Mail!’ Alert, Dies at 74
Early in the internet era, he was also behind AOL’s “Welcome!” “File’s done” and “Goodbye” messages. “They said my voice was heard more than 35 million times a day.”
Lawsuit Against Meta Over Section 230, Tech Shield Law, Is Dismissed
A professor sued pre-emptively to release software that would let users automatically unfollow everyone in their Facebook feed.
Your doctor may have an AI assistant taking notes during your next Zoom call
Don't worry if you notice your doctor doesn't seem to be writing down anything you tell them during a Zoom video call anymore. Thanks to a new deal between Zoom and medical AI assistant creator Suki, there's an AI assistant on top of things.
The two companies have agreed to embed Suki's AI notetaker and administrative assistant into Zoom's telehealth service. That means your doctor can focus on what you're telling them without needing to split their attention by writing notes and possibly missing something you've said.
The Suki Platform already handles similar duties for almost a million healthcare clinicians in the U.S. but usually deploys for in-person appointments. The AI (with your permission) records your conversation with the doctor and then mines the transcript for important details and follow-up visit plans. After doctor approval and annotation, the notes are added to your electronic health record. Suki's research shows that this kind of AI-based clinical notetaking can cut down the time doctors spend on paperwork by up to 70%, giving them more time and energy for actual care.
Zoom will augment its platform with Suki for appointments to do the same for telehealth visits. The only difference is that the AI will listen in during your Zoom call instead of from a recorder in the room with you and the doctor. That could be a big deal considering Zoom's explosive growth in the healthcare industry during the COVID-19 pandemic lockdowns.
“With over almost 140,000 healthcare institutions globally using the Zoom platform, we feel a great responsibility to leverage the power of AI to drive employee productivity and enhance patient experiences,” Zoom chief product officer Smita Hashim said. “Working together with Suki to provide critical AI-generated clinical note functionality to Zoom Workplace for Clinicians will reduce documentation overhead for clinicians, allowing them to focus on patients.”
AI AdminZoom's interest in clinical AI assistance is not unique, of course. For instance, Microsoft recently debuted a new set of AI tools to handle administrative tasks for healthcare providers using the technology gained when it acquired Nuance. There's also Amazon, which has brought its Bedrock AI-fueled tools like AWS HealthScribe to subsidiary One Medical. There are also plenty of smaller companies jockeying with Suki to provide clinical AI assistants, including Abridge and Notable.
That's mostly in the background, though. Just know that when your doctor makes a lot of eye contact with you on your Zoom call, it doesn't mean they have Netflix open in another window and they're ignoring you. It actually means they're paying more attention than they could before and will have help reminding them what you said if needed.
“AI is changing the way we interact with the world. Everything from how we communicate to how we use technology to how care is delivered will evolve. Video will be a critical interface in the AI-driven world,” Suki CEO Punit Soni said. “We are thrilled to work with Zoom to develop new interaction models and AI that will advance our mission of making healthcare technology invisible and assistive so clinicians can focus on what’s most important: their patients.”
You might also like...How Russia Openly Escalated Its Election Interference Efforts
The Kremlin did not bother to hide its efforts to influence the 2024 presidential election, as it did in the past.
Canada Shuts TikTok’s Offices Over National Security Risks
The order, which will allow people to still use the app, cited concerns about TikTok’s owner, the Chinese tech giant ByteDance.
ChatGPT coded a game for me in seconds and I am simply astounded – and coders should be very worried
ChatGPT just stunned me. Almost two years after I first asked an early version of the generative AI chatbot to help me code a game, I returned to ChatGPT 4o with the exact same prompt, and the results blew me away.
Ever since OpenAI's generative AI platform arrived, people have been turning to it to handle tasks they'd rather not do or, in some cases, to attempt things even they're not capable of. I can write, so I usually don't ask ChatGPT to compose prose, and I'm an OK artist so I only use generative image creation for entertainment. However, I am not a programmer. I've done some HTML, Javascript, and even BASIC coding in the past, but my skills are beyond rusty; they're calcified.
I wrote last year about asking ChatGPT to help me build a new game called Tic-Tac-Go. It's like Tic-Tac-Toe but on steroids. I envisioned the playing field to be a 12x12 grid. Winning would take six "X" marks or six "O" marks in a row.
Here's the original prompt:
I want to create a variant on the game tic-tac-toe, but I need it to be more complex. So, the grid should be 12-by-12. It should still use "x" and "o". Rules include that any player can block another by placing their "x" or "o" in any space around the grid, as long as it is in one of the spaces right next to the other player. They can choose to place their "X" or "o" in any space, as well, to block future moves. The goal is to be the first one to have at least six "x" or "o" in any row, column, or diagonal before the other player. Remember, one player is "x" and the other is "o". Please program this in simple HTML and JavaScript. Let's call this game: Tic-Tac-Go.
I get that there's no Wordle-grade cleverness here, but my idea still requires nested tablets and in-browser interactivity, which is immediately out of my skill set.
Things started off well in my first attempt, with ChatGPT providing CSS, JavaScript, HTML files, and a lot of explanations for almost every bit of code. However, most of what it gave me was incomplete because of a built-in character limit (4096). I kept asking ChatGPT to provide the rest of the code, but it kept losing track of what it already told me. I made my share of mistakes but also realized how ill-equipped ChatGPT 1.0 was to help me fill in the blanks. Ultimately, we never built the completed game, and after hours of trying, I gave up.
In the intervening 20 months, ChatGPT has made vast leaps in capabilities. I recently started using ChatGPT search and have been impressed at how much more effective it is than Google at getting me useful answers. ChatGPT 4o suffers from few, if any, of the limits I found in the first version. I began to wonder what else I could do. Separately, an office Slack conversation mentioned how some new AI-powered distributed computing platforms, like Oasis AI, were helping coders quickly build games. I didn't think it would be useful, but it reminded me of my failed ChatGPT game coding attempt. When I shared the post with a coworker, he wrote, "Wonder whether it would be any easier now with some of these new AI tools."
It was a good thought. Could ChatGPT 4o succeed where the original ChatGPT and I failed?
Instead of polishing my prompt – and because I am lazy – I simply copied and pasted the above prompt text into ChatGPT 4o.
It didn't take me long to realize that this was not even close to my first ChatGPT coding collaborator. If ChatGPT 1 had grade-school-level intelligence, ChatGPT 4o is college-level, at least.
ChatGPT 4o started by synthesizing my game idea into a more intelligent brief:
Creating a complex variant of Tic-Tac-Toe, called "Tic-Tac-Go," on a 12x12 grid, can be achieved with HTML and JavaScript. In this game, players take turns placing their markers ("X" or "O") to try and create a line of six markers in a row, column, or diagonal. The ability to block nearby moves adds an extra layer of strategy.
It then quickly spit out the HTML and Javascript code, and not just part of it. I got everything from the "html" at the top to the "/html" at the end. It looked complete. After the code, ChatGPT dropped in a four-point explanation of its work and the game's related features. Alongside HTML and Javascript, there was a "copy code" button. I copied the code and asked ChatGPT how I would test it in a Web browser, even though I pretty much knew how to do this, or at least I thought I did.
It recommended I open a text editor like TextEdit on the Mac, paste the code, and save it with an ".html" extension. I did this but when I loaded the file in Chrome and Safari, it was not the game, just a bunch of code and a "Restart the game" button.
I told ChatGPT this and asked if I needed to separate the Javascript into a separate file. It told me no and reminded me to check that the file had the right extension and disable RichText. It also gave me a few other suggestions, like using developer tools to check that the Javascript code was working.
I had a different idea. I took a screenshot of the broken page and dropped it into ChatGPT, asking if it could analyze the image. This multimodal use was not possible in the first ChatGPT, but it turned out to be a game changer.
"It looks like your file is displaying RTF (Rich Text Format) markup instead of HTML. This typically happens if the file was saved as an RTF document rather than plain text," wrote ChatGPT and then explained why it was happening and how to fix it.
I repasted the code, set TextEdit for PlainText, and resaved it as an HTML file. Then, I opened the file with Chrome.
Image 1 of 2
It worked. The game was exactly as I envisioned it, with 144 empty boxes and a two-player system for marking "X's" and "O's"
I realized that someone looking at this grid would have no idea how to proceed, so I asked ChatGPT if we could add rules below the game. ChatGPT spit out new code that included the game rules.
In the first chatGPT, when I asked it to redo any of the code, something would invariably get lost in the process. Not this time, though. There was a level of efficiency, clarity, and recall that was missing 20 months ago.
Feeling bold and maybe a little bit like I suddenly knew how to code, I asked for the next logical game update: the ability to play online. Sure, this was a web-based game, but as it was designed, you could only play against someone sitting next to you. I wanted online competition.
"To make this a multiplayer online game, you'll need to set up a server to manage game state and player interactions in real-time. Here’s a high-level overview of the steps required and some code examples to get you started..."
ChatGPT gave me detailed instructions and all the necessary code for setting up servers. It called for things like Node.js and WebSockets, and as I read it, I felt the water rising above my ears; I was way out of my depth. But not out of ChatGPT's. It clearly knows what it's doing and really doesn't need my assistance to code. Yes, I came up with the idea, but ChatGPT 4o did everything else.
We've come a long way in a short time, AI. The future for AI-assisted coding is very bright, but I'm not sure the news is as good for human programmers.
You might also likeTop 3 things you have to try with the new ChatGPT search
ChatGPT search (the new search engine built into ChatGPT that combines conversational AI with real-time information straight from the web) has recently launched for everybody who was signed up to the waitlist, or is a ChatGPT Plus subscriber.
What's more, OpenAI recently leveled up ChatGPT search with a shiny new Google Chrome extension that means you can use it from the address bar. If you're already using Google Chrome, this makes for a much more useful way to find and gather up-to-date information while simultaneously giving you more personalized filters and a way to leverage that information for new and fun activities.
Here are three practical and creative ways to take full advantage of ChatGPT search.
1. Trend spottingKeeping up with every new trendy hobby, TV show, book, or game that I might enjoy is impossible. Add in the things my friends and family care about that I want to be aware of for conversation (and birthday present planning), and no one who isn't a teenager could keep up, even with hours on social media.
ChatGPT search can handle that for me now. The AI can look up all of the latest news on subjects of interest to me and explain the latest buzz about those I am just curious about. As the AI adds more about me to its memory, it's going to be even better at curating those details. Asking “What’s trending today?” will get me the new video game buzz for games I might like as well as the conclusion of the latest episode of Survivor, which I need to know to be able to keep up when going to dinner with my in-laws.
2. Recipe rouletteI like looking up new recipes to try, especially if I don't want to go shopping and just use ingredients I have on hand. ChatGPT has always had the ability to come up with meal ideas, but the hallucinations and offline database meant I never really trusted what the AI wrote. Now, when I ask for specific kinds of dishes using ChatGPT search, like vegetarian meat sauce for pasta or the best methods for making gravlax, ChatGPT will not only remember previous requests and either reshare them or ask if I want new ideas, but it will actually use recipe websites to find them for me.
Even if I ask for recipes based on random ingredients like sweet potatoes, quinoa, and a mixed bag of herbs, ChatGPT search will find options that fit my palette without me worrying that it will suggest rocks as an appetizer.
Have you ever been hanging out with friends and felt like playing a trivia game without having to go to a bar? Well, ChatGPT search can use web search to pull together facts on everything from recent Oscar or Grammy award winners to the latest TikTok dance trends and act as your host for the night. I've found the right prompt can even get the AI to inject some humor into the game, while the web search keeps it from making up answers.
Even if you're not playing a game, having your own personal fact-checker is nice. Sure, there's Google, but ChatGPT search means you don't have to open a new tab or click through multiple websites (assuming Google AI Overview doesn't have an answer). And, because it looks online, you don't have to worry nearly as much about the response being a hallucination.
You might also like...Colossus expansion: TVA approves xAI for more power
TVA's board has now weighed in on xAI's power needs, and the CEO said that xAI will agree to reduce some of its 150 megawatt power usage during times of highest demand.
ChatGPT just got easier to find when you’re searching for something
ChatGPT has just paid a reported 15.5 million dollars for the URL chat.com, making it easier to get to ChaGPT. Now when you type Chat.com into your browser you’ll be redirected to ChatGPT.com. The move sparks rumors that the ChatGPT chatbot and search engine could be about to undergo a rebrand so it is better able to compete with rivals like Google.
The redirect was announced on X.com on Wednesday when Sam Altman, CEO of OpenAI, posted a tweet containing just the URL, chat.com.
https://t.co/n494J9IuENNovember 6, 2024
The announcement prompted immediate speculation from X.com users about whether OpenAI was going to drop the “GPT” part of the name altogether.
The chat.com domain was previously owned by Dharmesh Shah, the founder and CTO of HubSpot, who claimed in a LinkedIn post to have paid $10 million for it in 2023, but had sold it a couple of months later for “more than I paid for it”. At the time he said, “I'm not at liberty to disclose details on who the purchaser was or the price paid. Sorry.”
What's in a name?
It’s easy to forget how new the generative AI field is. As with any emerging field of technology, brand names tend to come and go quite quickly. Google started by calling its AI Bard, but this was dropped in favor of Gemini. Recently there has been speculation that Microsoft is not happy with its Copilot AI name, and is considering "Windows Intelligence". The move would be perhaps seen as a little too close to what Apple is doing with Apple Intelligence.
The GPT part of ChatGPT stands for Generative Pre-training Transformer and refers to the type of large language model (LLM) that uses the transformer architecture but is hardly a brilliant brand name. The URL chat.com is much more memorable, and as Sharmesh Shah said when he originally purchased the URL “chat.com is absolutely brilliant in terms of simplicity, shortness and being totally *on point* and meeting the moment. It immediately creates user trust. Somebody (not me) will build a massively successful product/company on it.”
OpenAI, the makers of ChatGPT, recently launched the equally cumbersomely-named AI search engine, ChatGPT search, despite owning the much more catchy URL, SearchGPT. In September the company announced that it would be releasing a new series of LLMs starting with “o1”, which recently saw the light of day, just for a few short hours, by mistake.
You might also like...Windows 11’s Paint and Notepad apps are getting smart new AI features – though one of the best will be for Copilot+ PCs only
- Notepad is getting a rewrite feature and generative fill is coming to Paint
- There are a fair few more AI abilities coming too, including generative erase
- One of the most impressive, generative fill, will only come to Copilot+ PCs
Windows 11’s Paint and Notepad stock apps have been worked on considerably in the recent past, and are getting some new AI features – albeit they’re only in testing for now.
Microsoft tells us that updates for these apps have started to roll out to testers in the Canary and Dev channels (which are the two earliest test channels).
Let’s start with Notepad which is getting a new ‘Rewrite’ ability. This works as you might imagine – you select some text, then fire up the tool which uses generative AI to do what it says on the tin, rewriting those words.
You get three different rewrites to choose from, and it’s possible to fine-tune the AI’s responses, so you can ask for something longer, for example, or in a different tone.
This ability is rolling out to testers in the US, Canada, UK, France, Germany, and Italy. As with most of the new features here, it’ll require you to be signed in with a Microsoft account.
There’s another non-AI benefit for Notepad here in that the app will launch faster. Microsoft claims that Notepad will open more than 35% more swiftly, with some Windows 11 users seeing a speed boost of 55% or maybe even more. So, Notepad should be considerably nippier when appearing on your desktop.
Painting cleverMoving on to Paint, this core app is getting a feature called ‘generative fill’ which can be used to easily add an AI-generated element to an image.
The example Microsoft shows is a person making a landscape who decides they want a castle in the picture. So, they simply select the area where they want it situated, and feed the AI a prompt – in this case ‘medieval castle’ – and Paint goes through its processing for some seconds before the castle appears.
It’s nice and seamlessly integrated into the landscape, too, although we guess as ever with AI, the results could be variable. If you don’t like what Paint has provided, you can ask it to try again, and generative fill will run once more to deliver a different take on a castle (or whatever it is you want).
Moving from creation to destruction, we have ‘generative erase’ also inbound for Paint. This lets you highlight an element in an image – a person in the foreground for example – and it’ll be removed, with Paint filling in the empty space left to match the background. Again, it’s clever stuff.
In this case, you brush over the area to be erased, allowing for a more fine-tuned selection than generative fill where you draw a box.
These are some pretty nifty AI-powered abilities with some slight caveats. The main one being that Paint’s generative fill is only going to be available with Copilot+ PCs that have a Snapdragon (Arm) CPU initially. We presume it’ll come to other (non-Arm) Copilot+ PCs soon enough – but as to whether Windows 11 users with a ‘normal’ PC will ever get it, that might not be the case (as it may require the beefy NPU which is a requirement for a Copilot+ machine).
All the other features are coming to non-Copilot+ devices, mind, but there’s another bonus here for Copilot+ PC owners. This is that Cocreator in Paint – which is also exclusive to these Windows 11 devices, and creates AI-generated images from scratch – has been made faster, with fine-tuning to its “underlying diffusion-based model” as Microsoft explains. So, everything you do with Cocreator should happen more swiftly.
The other potential catch, we guess, is that this is more fuel to the fire of discontent for some users who think that Notepad is getting beefed up too much with features – those folks would no doubt say bloated – and that this is at odds with its core philosophy. That said, at least the freshly introduced sizeable performance boost for booting the app will go down well.
All of these features are still in early testing, as noted, so it could be some time before they hit PCs running the release version of Windows 11. Microsoft sounds like it’s being cautious with the rollout in testing, as it’s going to “monitor feedback and see how it lands” before pushing these updates to more Windows Insiders. Even if you’re a tester, then, you may not see these new features for a while yet.
You may also like...Tech C.E.O.s Courted Trump Before the Election
The executives of tech’s biggest companies largely ignored Donald Trump before the 2016 election. This time around, they’re far more friendly.
Tech Giants Face a Familiar Uncertainty With Trump
Apple, Amazon, Google, Meta and others learned during the last Trump administration to expect the unexpected when it came to Washington scrutiny and support.
Gemini will yada yada your Google Chat into a neat summary
Step away from your computer or phone for a few minutes, and you might return to an avalanche of notifications about new messages. If you're using Google Chat, you won't have to comb through them all to find out what you missed anymore, thanks to Google's Gemini AI assistant. You can pull up Gemini from the Google Chat sidebar and ask the AI to summarize the conversation and dig into the most important bits.
The new feature expands Gemini's presence from other Google Workspace applications like Docs and Drive into Google Chat. If you click on the "Ask Gemini" icon at the top of the Google Chat interface, a chat window for the AI will appear where you can ask about what's been said in group chats, direct messages, and spaces.
If you ask the AI to "catch me up," you'll get a complete conversation summary, which you can ask for in bullet points. If it's too short, you can request a longer summary too. You can also ask for more specific details, like any requests for help, key takeaways, or other decisions made in the thread. You can even ask about other people's tasks or what a person said about specific topics. The demo below shows how it works.
Gemini can't sort through your entire conversational history, just the current view. It also is restricted to Google Chat. That means no emails or files in Google Drive, despite Gemini having a presence and access to those applications in other circumstances. Google claims this is deliberate as it maintains focus on the current chat in context without pulling in irrelevant information. Plus, tightly constraining data sources reduce the risk of Gemini hallucinating. You also won't be able to pull up Gemini in Google Chat without a subscription to Gemini Business, Enterprise, Education, or Education Premium.
Even with those (likely temporary) limits to availability, bringing Gemini to Google Chat fits with how Google is working to embed Gemini across all of its platforms and services. That includes Gemini Extensions to take up Google Assistant's role with Google Messages, Maps, and pretty much everything Android does.
You might also likeWindows 11 24H2 misery continues, as Microsoft’s buggy update is now breaking printers – especially on Copilot+ PCs
- Windows 11 24H2 is reportedly causing a lot of problems with printers
- This affects all major printer makers, mainly on Copilot+ PCs
- Microsoft has admitted the problem and provided some workarounds
Windows 11 24H2 is continuing its run of trouble with bugs and this time we’re hearing about various difficulties with printers, which are mostly hitting Arm-based Copilot+ PCs, but also other computers (with x86 Intel and AMD chips) too.
Neowin reports that there are quite a number of complaints from those with printers who have upgraded to Windows 11 24H2 and are finding their device is no longer working. This is affecting all the best-known printer manufacturers, the likes of Brother, Canon, HP and so forth.
The issue is mainly being experienced by those with a Copilot+ PC powered by an Arm processor, as mentioned, and it either completely derails the printer, leaving it non-functional, or breaks certain features. In other cases, Windows 11 users can’t install the printer driver.
As noted at the outset, these gremlins in the works are also hitting standard (non-Arm, Intel and AMD CPU-based) PCs, and there are reports of various printer hiccups, and indeed printers vanishing from these systems, too.
On Reddit, one admin observes in a post about ‘24H2 problems hitting us hard’ that: “Print queues [are] clogging, that’s if the driver wasn’t randomly deleted from the machine.”
Printers going missing, or the driver being removed, appear to be common themes with 24H2, as well as not being able to install (or reinstall) said printer driver.
Potential workarounds to fix printer installation problems
Microsoft has acknowledged the issue and produced a support document which provides some possible workarounds to get your printer going on an Arm PC.
The company notes: “Some Arm PCs (for example, Copilot+ PCs) may not be able to add or install a printer using the included installer or the installer from the manufacturer’s website.”
The first suggestion to get around an installation failure is simply to connect your printer using a USB port.
Otherwise, you can connect the printer by going into the Settings app, then Bluetooth & devices > Printers & scanners, where you select the Add device option. This then allows you to manually add the printer to your system (hopefully, at any rate).
With any luck, one of these two options will help you get around any inability to install a printer – though there may still be other scenarios where printers are misfiring with Windows 11 24H2 by the sound of things.
All this comes among a raft of issues with Windows 11 24H2, including new Intel Z890 motherboards constantly crashing, drive space mysteriously vanishing, and also Task Manager malfunctioning among many other problems. The good news is that Microsoft is working on fixing a bunch of these bugs already, but there’s a lot of firefighting to be done.
You may also like...Microsoft reportedly sabotaged Windows 95 and 98's Control Panel, or it looks that way – is this a worrying revelation?
- A possibly intentional delay with the Control Panel has been found
- Presumably, it was necessary as a workaround in the distant past
- But it may not inspire confidence in those facing odd bugs in Windows 11
A rash of recent issues with the Windows 11 24H2 update has left many users frustrated with their systems, and it probably won’t be any comfort for those folks to learn that Microsoft has reportedly tampered with old Windows versions to make them run slower (albeit in the distant past, and with one relatively small aspect).
As spotted by German tech site WinFuture (via Neowin), Oerg866 on X posted to reveal that Windows 95 and 98 had a ‘hard-coded 8 second delay’ when using the ‘Add New Hardware’ wizard in the Control Panel.
In an attempt to further speed up Windows 9x QuickInstall, I have patched SYSDM.CPL to remove the hard-coded 8 second delay(!!!) when new hardware is found and reduced it to 300 milliseconds. pic.twitter.com/BfGNpSjMfwNovember 4, 2024
While it isn’t at all clear why this occurs on the now-ancient operating systems, it’s worth wondering why Microsoft decided to implement this delay.
It seems a baffling move on the face of it, but there’s presumably a reason why the long pause was added in by programmers (unless it was just a mistake - which seems unlikely, and it’d be quite the blunder if it was).
What seems most likely is that the delay was introduced due to hardware limitations at the time, meaning there must have been some issues with devices back in the day which required this pause as a workaround. Why the delay had to be so lengthy is anyone’s guess, though.
It also makes you wonder if Microsoft introduced any other similar stalling tactics in these old versions of Windows - or indeed in more modern operating systems like Windows 10 or Windows 11.
It goes without saying that we don’t need to worry about Windows 9x versions anymore (Windows 95 and 98 are older than me). There should be no reason for a similar delay to occur in Windows 11, either, unless it really was an error - which could prompt the question: are there other such blunders still hanging around in the legacy Control Panel?
That thought wouldn’t exactly bolster the confidence of anyone currently using the 24H2 update and running into some of the aforementioned problems with the upgrade.
The ton of bugs that come with Windows 11 24H2 include problems that have left PCs with Intel motherboards suffering from BSODs (Blue Screen of Death crashes). Also, a good few of these glitches are really odd affairs - like a File Explorer menu disappearing off the top of the screen - admittedly not nearly as head-scratching as the reported hard-coded delay, mind you.
The old Control Panel won’t be around forever, of course. Microsoft is planning to gradually migrate the legacy Control Panel’s functions into the new Settings app in Windows 11 with its modern user interface.
We’ve reached out to Microsoft to ask the company about the discovered delay in Windows 95 and 98, and whether it might be able to shed any light on what was going on behind the scenes here. We’ll be sure to update this story if we hear back.
You might also like...- Microsoft is so desperate for people to drop Google for Bing it’s offering a $1 million reward
- Fed up with Windows 11’s look? New mod lets you revamp the desktop, including the floating taskbar once seen in a Windows 12 leak
- Good news! Soon Windows 11 updates won't take forever thanks to 24H2's speed boost
Jarvis AI is real – Google accidentally leaks its AI agent that browses the web for you
- Google has accidentally leaked its Jarvis AI
- Jarvis AI browses Chrome for you
- It's expected to be officially unveiled in December
Google has accidentally leaked its upcoming AI agent, Jarvis AI, which will browse the web for you in Chrome.
According to The Information, Google posted an internal preview of the product, codenamed Jarvis, on the Chrome extension store but was quick to remove the prototype before too many users noticed.
The listing read "a helpful companion that surfs the web with you.” which lines up with the news from October that Google was developing Jarvis, the AI agent that would help you shop online and even be able to book your next vacation.
Those who managed to download the Jarvis AI prototype before it was pulled from Chrome's extension store were unable to use it as the program required specific access permissions.
With The Information reporting Google plans to officially unveil Jarvis in December, this leak confirms that we won't have long to wait.
Christmas shopping for me
Jarvis AI is expected to allow users to automate web browsing tasks from within Google Chrome so they can focus on more important things while the AI agent shops or even books flights for you.
For example, think of how great this could be for mundane Christmas shopping – after all, nothing says 'I love you' more than a gift purchased by AI.
AI agents, which are systems that can complete tasks without your supervision, are starting to pop up increasingly with companies like Anthropic releasing a product similar to Jarvis last month. It's also reported that OpenAI could be working on an agent too, although we're yet to see how that would differ from Google's upcoming product.
AI development is booming with companies releasing new tools almost daily. Just last week ChatGPT Search became available to Plus subscribers ushering in a new era for AI search engines. With Jarvis AI, however, we might not even need to do the searching ourselves.
You might also like...Instagram to use AI to catch teens lying about their age and automatically move them to restricted accounts
- Meta to use AI age-prediction software to spot teenagers lying about their age
- Move comes after Instagram introduced teen accounts in September
- Meta will begin moving teens automatically to its new accounts
According to a new Bloomberg report, Meta is set to use AI to predict the age of its Instagram users and automatically transfer anybody under 18 to a teen account, if it believes they are lying about their age.
In the wake of national outcry about the impact of social media platforms like Instagram and Facebook on the mental health of teenagers, Meta introduced teen accounts in September. Teen accounts, for children aged 13-17, have built-in limits on who can contact them, and what content they can see. They are automatically set to hide potentially offensive comments and message requests.
Secure accounts for teenagers will only work if those under 18 volunteer to use them or are honest about their ages, which is why Meta has devised a way of enforcing the appropriate accounts. The proprietary software tool Meta has developed, called ‘adult classifier’, will come into play next year and is designed to classify users into two groups: under or over 18 years of age. According to Allison Hartnet, Meta’s director of product management for youth and social impact, the tool will scan a user's profile, what content they interact with, and their follower list to determine their age. Even innocuous “happy birthday” messages can be used to help determine a user's true age.
Busted
Since teen accounts are new and currently voluntary there’s a large user base of teens on Instagram who are using regular adult accounts. Meta plans to start moving teens who have volunteered their age information onto the new teen accounts very soon, and introducing the 'adult classifier' early next year.
The move is Meta's latest attempt to redress some of the public outcry over the effect of social media on teens. In 2021 a report published in the Wall Street Journal showed that Meta’s reports indicated that it knew Instagram was harmful to the mental health of teenagers, especially teenage girls. “We make body image issues worse for one in three teen girls,” said one slide from Meta’s own report from 2019.
You might also like...AI meets democracy on Perplexity AI's new Election Information Hub
AI conversational search engine Perplexity wants to help keep you updated about the election this year and has introduced a new Election Information Hub designed to do just that. The hub uses Perplexity's AI tools to bring real-time information to visitors.
Perplexity sees the hub as a seed for future voter awareness and participation, one that has accessible information on candidates, ballot measures, and results and that can prove AI won't fall into hallucinations around important, fast-changing subjects like elections.
You can find all kinds of election-related information on the platform. Perplexity limits its sources to reputable providers like The Associated Press and Democracy Works to ensure there aren't any wrong or made-up answers from the AI about the election and various other votes. The hub includes data on state and federal elections as well as propositions on the ballot in different states. Real-time tracking ensures that users are informed as things change.
"We want to do our part to support an informed electorate, so we’ve built the election hub on Perplexity’s answer engine: an entry point for understanding key issues, voting intelligently, and tracking election results," Perplexity explained in an announcement. "We answer your election-related questions using a curated set of the most trustworthy and informative sources."
The hub is fairly easy to use, especially if you've played with Perplexity and its AI search engine. It launched only a few days ago, limiting the utility of information on voting requirements, polling places and times, and related details. But, those elements will be a boon to users in future votes.
The AI also helps summarize ballot measures and candidate profiles, including official policy positions and who has endorsed the measures and candidates. That's a good resource when there's a cavalcade of information for each election. You can use the hub without a subscription to Perplexity's premium services, another point of accessibility Perplexity hopes will encourage people to try out the hub. And, it doesn't seem to bring in sponsored follow-up questions, which makes sense as it would run somewhat counter to the non-partisan, informative tone Perplexity is going for.
"We want to make it as simple as possible to receive trusted, easy-to-understand information to inform your voting decisions. For each response, you can view the sources that informed an answer, allowing you to dive deeper and verify referenced materials," Perplexity wrote. "Whether you're seeking to understand complex ballot measures, verify candidate positions, or simply find your polling place, Perplexity is here to support your civic engagement."
You might also like...Apple is currently levelling up Siri’s ‘onscreen awareness,’ enabling it to interact with your screen
- Apple is developing an "onscreen awareness" feature to allow Siri to understand and interact with the content currently displayed on your screen
- Apple will also provide APIs to developers for integrating onscreen awareness into their third-party apps and is currently testing ChatGPT integration, allowing Siri to answer questions based on screenshots
- While not available in the iOS 18.2 beta, "onscreen awareness" may arrive in time for iOS 18.4 in 2025
Among the digital assistants, Siri has fared fairly well (certainly compared to Cortana, the ill-fated assistant from rival Microsoft), and now Apple is working on making Siri even smarter by giving it a better sense of what you’re looking at on your screen, calling it ‘onscreen awareness.’
Apple has gone into detail about the development of this feature in an Apple Developer Documentation page, which also notes that it’s due to be included in various upcoming Apple operating system (OS) beta releases for testing.
Apple originally showcased onscreen intelligence in June 2024 and this is a pretty solid indication that it’s still in development.
The core idea of onscreen awareness is pretty straightforward – if you’re looking at items on your screen, say a document or a browser with a page open, and you have a question about something you’re looking at, you can ask Siri (equipped with Apple Intelligence). Siri should then be able to respond to your question with relevant information or perform an action asked of it, like sending content to a supported third-party app.
If it works as intended (and that’s a big ‘if’), it will result in a smarter Siri that won’t need you to describe what you want it to do as extensively as you probably need to right now. For example, you can have a document open and ask for a summary without entering the contents of the document yourself.
Apple's plans for Siri's foreseeable futureMacRumors reports that Apple has provided APIs for developers, enabling them to make the content in their apps available to Siri and Apple Intelligence. The idea is to provide developers with this API many months in advance of its release so that you can use onscreen awareness with third-party apps when it’s officially rolled out.
Currently, we know Apple is trialing out ChatGPT in the latest iOS 18.2 beta version (among other Apple OSs), and ChatGPT integration coupled with Siri will enable you to ask questions about items (like images, PDFs, and videos) on screen. Then, Siri will take a screenshot and pass it on to ChatGPT to reply to your query. This means that, for now, this functionality is limited to screenshots of your screen.
However, onscreen awareness is a little bit different, as pointed out by MacRumors, as onscreen awareness is intended to be integrated in a more direct way.
Siri’s onscreen awareness is supposed so be a sort of capacity to directly analyze, interpret, and interact with the content on your screen. If someone messages you their number, and you’d like to save it, you can tell Siri to create a new contact without having to add additional instructions or Siri having to perform many intermediate steps.
Will Siri survive and even thrive in the AI age?Apparently, onscreen awareness isn’t actually in the iOS 18.2 developer beta just yet, and MacRumors speculates that it’s one of multiple Siri features that we won’t see for a while, but this news is still promising. One prediction has onscreen awareness to possibly be included in iOS 18.4, expected to be released in the first half of 2025.
If this pans out, Siri could become a more helpful digital assistant, and with Apple’s flair for design, it might become the choice digital assistant for many. This development reminds me of what Microsoft is aiming for with Copilot. That hasn’t been received very well from what we’ve seen so far, leaving the goal open for Apple.
YOU MIGHT ALSO LIKE...- Your Mac’s menu bar will finally get a weather widget in macOS Sequoia 15.2 – plus these Apple Intelligence features
- iOS 18.2 beta shows exactly how Siri’s ChatGPT upgrade will work on iPhones – including new ‘daily limit’ guide
- Microsoft could be about to rebrand its AI into Windows Intelligence rather than Copilot
Windows 11 update fails are becoming a joke – but it looks like Microsoft is working hard on fixing them
- New Windows 11 preview build is almost entirely about bug fixing
- There are some important fixes delivered for the Start menu and more
- Microsoft is working hard to firefight all the glitches in Windows 11 24H2
Windows 11 has been hit by a whole load of bugs in recent times, with the move to version 24H2, as you’ve probably noticed – but the more positive news is that Microsoft is busy fixing a fair few of those gremlins in the works.
We’ve got evidence of this courtesy of the latest preview release of Windows 11 pushed out to the Dev channel, namely build 26120.2213.
Microsoft lists the work underway in terms of bug fixing in the usual blog post, and while there’s a whole lot of resolving glitches going on, some of it’s more mundane – so let’s pick out some of the highlights.
A notable bug we’ve recently reported has hit Task Manager, resulting in it showing that there are zero apps and processes running on the host PC – which, of course, can never happen (otherwise Windows 11 itself wouldn’t be working, never mind anything else). Microsoft informs us that it has resolved this issue where Task Manager is claiming a zero count for running processes.
On top of that, an ‘underlying issue’ with the most recent 24H2 preview that meant some games were completely failing to launch has been cured here.
Multiple bugs with the Windows 11 Start menu have been fixed with build 26120, including one where people were seeing a large amount of spacing between the apps present in the menu’s list of applications. There also a broad cure for an issue “impacting Start menu reliability” we’re told.
There’s a similar problem with File Explorer where items in the navigation pane were also becoming too spread out, with too much spacing between them for some users, and this has been remedied too. A more minor issue where RAW images taken in portrait mode were displaying in landscape mode with File Explorer’s thumbnails has also been fixed.
Furthermore, Microsoft has applied a clutch of fixes for various minor problems with the taskbar, including a glitch where the ‘X’ button to close a window off the taskbar preview (that pops up when you hover over the app on the bar) wasn’t working.
This preview release of Windows 11 is notable because there’s hardly anything at all here in terms of feature additions (just a minor change for the IME toolbar which won’t make any difference to most users).
Almost all of this build is about fixing bugs, and aside from that, there are a couple of known issues flagged which haven’t yet been resolved. One of those is the bugs with the new gamepad keyboard layout for Windows 11 users that have meant it has been put on ice for the time being, and the other is a problem where those running multiple monitors are seeing corruption with their desktop background (with “big black areas” appearing mysteriously).
Microsoft is still working on the cures for those issues, and a fair few other problems besides that are affecting the latest incarnation of its desktop OS. To say Windows 11 24H2 has got off to a rocky start in terms of bugs is something of an understatement, in fact, but it does appear that Microsoft is working hard to firefight these various problems.
You may also like...