It feels as though Google is playing catch up at the moment when it comes to the ChatGPT-powered AI that Microsoft has introduced to Bing – but Google CEO Sundar Pichai says that his company's own Bard bot is going to quickly get more capable.

In an interview with the NYT's Hard Fork podcast (via The Verge), Pichai said that Bard was currently like a "souped-up Civic" taking on "more powerful cars" – but also that Google has "more capable models" that are going to get deployed in the coming days.

"We knew when we were putting Bard out we wanted to be careful," Pichai said. "Since this was the first time we were putting out, we wanted to see what type of queries we would get. We obviously positioned it carefully."

According to the Google CEO, more capable PaLM (Pathways Language Model) versions of the Bard chatbot will be rolled out "over the course of next week". That will mean Bard gets noticeably better at reasoning, and coding, and in other areas.

Slow and steady

Pichai's overall tone was a mixture of caution in terms of experimenting with what Bard could do, and excitement about where it could eventually end up. These "very, very powerful technologies" can be personalized to companies and people, Pichai said.

The Google executive also addressed data privacy concerns and worries about the pace at which AI engines like Bard and ChatGPT are progressing. Some of the most prominent voices in tech have called for a six-month pause on artificial intelligence development.

Pichai welcomes these sorts of discussions and wants to see governments setting rules: "AI is too important an area not to regulate," he told the podcast. "It's also too important an area not to regulate well. So I'm glad these conversations are underway."

The interview also touched on a variety of other areas, including how AI might impact jobs ("we all may need to course-correct in certain areas") and the content put out on the web ("we'll be committed to getting it right with the publisher ecosystem").

Analysis: lots of big questions

This latest podcast interview highlights just how many big questions there are over AI right now: how it will affect data privacy, the sorts of jobs it might make redundant, the impact it will have on publishers if Google and Bing are one-stop shops, and so on.

To be fair to Pichai, he dealt with those questions very sensibly – but that's not necessarily a guarantee that some of the worries we have over AI are going to go away. We're facing a gigantic shift in the way we live our lives and get our information over the web.

Pichai admitted that the tech is "going to be incredibly beneficial" but also "has the potential to cause harm in a deep way". It's good to acknowledge that, but companies such as Google are driven by profit and money making over any sense of moral obligation.

At least there's a conversation happening. "This is going to need a lot of debate," Pichai said. "No one knows all the answers. No one company can get it right. Am I concerned? Yes. Am I optimistic and excited about all the potential of this technology? Incredibly."

• Google Bard can't really help you play Wordle yet

Microsoft’s own Defender antivirus program has erroneously labeled a number of safe links as malicious, sowing confusion among dozens of users. 

After one of the affected users posted about the problem on Reddit, others quickly chimed in, confirming they had seen the same issue. For some, Zoom links were classified as malicious, while for others, Google’s links, as well.

Soon after being tipped off, Microsoft took to Twitter to acknowledge the problem and to say that its engineers were working on a fix.

Trouble viewing alerts

"We're investigating an issue where legitimate URL links are being incorrectly marked as malicious by the Microsoft Defender service. Additionally, some of the alerts are not showing content as expected," Microsoft said. 

"We've confirmed that users are still able to access the legitimate URLs despite the false positive alerts. We're investigating why and what part of the service is incorrectly identifying legitimate URLs as malicious."

A later update on the Microsoft 365 Admin Center portal stated that admins can expect an “increased number” of high-severity email message alerts saying “A potentially malicious URL click was detected”, and that they can also expect trouble viewing the details by pressing the “View alerts” link in the messages. 

"We're reviewing service monitoring telemetry to isolate the root cause and develop a remediation plan," Microsoft said. "Impact is specific to any admin served through the affected infrastructure."

A few hours later, Microsoft issued yet another update, saying the false positive issue has been addressed. Apparently, the problem was in the SafeLinks feature, and its engineers fixed it by reverting recent updates.

“We determined that recent additions to the SafeLinks feature resulted in the false alerts and we subsequently reverted these additions to fix the issue,” Microsoft said in a tweet. “More detail can be found in the Microsoft 365 admin center under DZ534539.”

• Check out the best firewalls

Via: BleepingComputer

Researchers from SentinelLabs have uncovered a new toolkit cybercriminals are using to breach email and web hosting services. 

The malware toolkit, called “AlienFox”, is being described as “highly modular” and getting regular updates. Most of the tools in the kit are open source, and with the speed at which it’s being updated, the researchers concluded the devs are becoming “increasingly sophisticated”.

As per SentinelLabs’ report, hackers are shilling AlienFox on Telegram groups, claiming it can be used to compromise misconfigured hosts on cloud platforms and steal sensitive data.

Abusing scanning platforms 

"AlienFox tools facilitate attacks on minimal services that lack the resources needed for mining," the researchers said in their report. "By analyzing the tools and tool output, we found that actors use AlienFox to identify and collect service credentials from misconfigured or exposed services. For victims, compromise can lead to additional service costs, loss of customer trust, and remediation costs."

To generate a list of misconfigured hosts, the toolkit uses security scanning platforms, such as LeakIX, or SecurityTrails. Then, it uses multiple scripts to pull sensitive information such as API keys and secrets from configuration files, the researchers explained. Some of the versions analyzed for the report were able to establish AWS account persistence and escalate privileges, as well as collect send quotas and automate spam campaigns through victim accounts and services.

So far, attacks against cloud-based services were limited mostly to cryptominers. Threat actors would use compromised cloud servers to run XMRig or similar cryptocurrency miners, generating tokens without needing to pay for electricity, internet, or compute power. With AlienFox, SentinelLabs claims, opportunistic cloud attacks are no longer confined to cryptomining. 

“For victims, compromise can lead to additional service costs, loss in customer trust, and remediation costs,” the researchers concluded.

• Here are the best firewalls 

Via: The Register 

ChromeOS has been getting plenty of upgrades to its video editing since 2022, and now it's getting even more from Google Photos. 

According to a new Google support page post, Google Photos will be empowering the best Chromebooks everywhere with even more new features. Jumping off from Chromebook’s Gallery app integration in 2022, Google Photos has added more support for Gallery and Files by allowing you to “seamlessly” use photos and images saved on the Chromebook and edit them in Google Photos. 

Another major edit comes in the form of revamped movie creation tools on the Photos app, which lets you create a movie from scratch using a suggested theme. After you choose a theme from the Photos app, as well as the people or pets you want to be included in the movie, the video editor uses video clips and photos you have stored that match that theme and strings them together into a custom movie. If you want a more hands-on approach instead, Google Photos has a search feature that lets you find and arrange all that media yourself.

Google has a full how-to on creating movies, as well as editing photos and videos, through the app on ChromeOS. But it’s a simple process that involves the following:

• Install the Google Photos app 
• Open up the media you want to be edited and select Edit. 
• From there you choose whichever editing tool you want to use. 

If you want to create a movie:

• Select Creations on the Google Photos app
• Choose either the suggested movie option or New movie

Chromebooks are getting more recognition

ChromeOS, once the oft-ignored child of Google that's been the platform for the best cheap laptop option for school kids and nothing more, has been growing in both popularity among users and in support by its own creator. 

Last year, we saw the debut of the HP Elite Dragonfly Chromebook, considered by our standards to be the best Chromebook and one that can stand toe to toe with plenty of the best Ultrabooks out there.

More features like the aforementioned updates to Google Photos and the addition of more Apple support via one of the best video editors, Apple’s LumaFusion app, have become commonplace in the last several years. We even have gaming Chromebooks like the Acer Chromebook 516 GE that are specifically designed to handle high-end gaming via the cloud.

And, unlike even the best Windows laptop, ChromeOS simply doesn’t see the kinds of viruses and malware plaguing other operating systems, meaning you don’t need to worry about constantly protecting and scanning your machine for threats (or cleaning it out if something nasty does worm its way in).

It’ll be interesting to see how much more support Chromebooks get in the future, but it’s an exciting time for Chromebook fans.

From afar, the ROG Flow Z13-ACRNM from Asus looks like a rugged tablet similar to the one magnesium-coated ToughPad models from Panasonic. However, come closer and the industrial design, the CNC-milled aluminum chassis, the reinforced corners and the bundled accessories just ooze “schmexi-ness”, to paraphrase our sister publication, Laptopmag.

Produced in partnership with revered artist Errolson Hugh, the RMT02 (as it is otherwise known), is pitched as a gaming tablet but its price (£2999 in the UK, $2500 in the US) and form factor make it a match in heaven for businesses, organizations and professionals eager to get a top of the range slate without an outrageous price tag.

The fact that it runs Windows 11 Pro by default, supports Microsoft Windows Hello, has a stylus and a full array of accessories (accessory bag, keyboard strap, main strap and detachable keyboard) makes it a very enticing proposition. Could it be used as a business tablet? We wouldn't say no. Mobile workstation? Definitely given its specs.

A great alternative to boring business tablets

And the RMT02 delivers value-for-money in spades but let’s start with the bad news though: it is not IP-rated so you won’t be able to use it in dusty or moist environments. It’s not been tested to MIL-STD-810G standards either although Asus says that it has robust protection with plenty of rubber around the chassis to mitigate any falls. There’s also no fingerprint reader which is a shame.

The rest though is spectacularly good, very good: there’s a 13th Gen Intel Core i9-13900H processor (that’s 14 cores and 20 threads), an Nvidia Geforce RTX 4070 GPU with 8GB GDDR6 (remember, this is a gaming laptop wannabe), 32GB LPDDR5, 1TB SSD and Wi-Fi 6E. Despite being a mere tablet, it has a TB4 port with two USB ports, an audio jack and even a card reader.

The screen certainly remains the highlight: a sumptuous QHD+ 13.4-inch display that is not only DCI-P3/Pantone validated (making it an ideal display for Photoshop) but also one with a high brightness (500 nits), perfect for the busy outdoors. Asus chose a 56WHr battery, powered by a 130W AC GaN adapter, to keep the tablet running. At 1.32Kg alone (add 390g for the backlit chiclet keyboard), the ROG Flow Z13-ACRNM is light enough to be carried around.

There’s plenty else to talk about: Dolby Atmos, two speakers, a trio of microphones, Dolby Vision HDR support and the ability to support external GPUs. Just be mindful that battery life will not be great because of the high refresh rate and the small battery size.

Keeping your business safe from cyberthreats is set to get a lot easier, thanks to a new update from Microsoft Teams.

The video conferencing service is set to benefit from a number of security boosts as Microsoft brings several Defender for Office 365 tools and services to the platform.

This includes full end-to-end protection for Microsoft Teams, but also a hopeful rise in awareness and preparation for users in businesses across the world

Microsoft Teams security

"Today, 270 million users use Teams to communicate and be productive, making it indispensable to effective collaboration," the company wrote in a blog post announcing the news.

"However, the growing number of users has also made a new target for adversaries trying to gain access to an organizations’ sensitive data and operations. Attacks like phishing and ransomware that for decades have primarily used email as an entry point, are now also targeting users on collaboration tools with growing frequency."

The launch follows recent Microsoft research which found nearly three-quarters (71%) of companies admitted to sharing sensitive and business-critical data via collaboration tools like Microsoft Teams. 

Among the new features is the ability for users to report suspicious messages directly in Teams, similar to what is already available now to report suspicious emails in Microsoft Outlook.

Security teams will be alerted whenever users report suspicious messages, with the ability to block or quarantine potential risks, and can view them in the Microsoft 365 Defender portal.

Microsoft is also adding to the existing advanced hunting features in Microsoft 365 Defender to support security for Teams, and is giving IT admins the chance to run attack simulations to raise awareness and protection against threats such as phishing.

Collaboration Security for Microsoft Teams is available for customers of Microsoft E5, Microsoft E5 Security, or Microsoft Defender for Office 365, with a public preview available now, ahead of a wider rollout scheduled for June 2023.

• Check out our rundown of the best online collaboration tools available

Microsoft’s efforts to help its customers work “smarter and faster” has been spotted in the wild as the company scales up the testing of its Copilot platform to select users.

Following heavy investment in OpenAI, the company behind ChatGPT, and the subsequent rollout of Bing’s ChatGPT integration and more niche artificial intelligence tools to other services, Microsoft believes it is on the cusp of getting generative AI into the hands of millions more daily users as it looks to integrate it into its Microsoft 365 office software.

Some users in the Insiders beta channel have now gotten their hands on the early version of Copilot for Word in Windows 10 and Windows 11.

Copilot for Word

Microsoft hopes that with its GPT-4-based generative assistant, users will be able to produce and edit work more quickly with a tool capable of handling context like attachments.

At its basic level, Copilot can help handle things like formatting and rewriting text. While it’s also capable of generating new text, it’s at its most impressive when handling other documents, where it can do things like create summaries and action plans.

Key to its success will be Copilot’s ability to seamlessly work across the various Microsoft 365 apps on offer to crossreference data across workloads that would otherwise typically be tedious and time-consuming for workers.

Microsoft has kept the details about Copilot undercover, however it’s likely that testing will begin to roll out across other services and to more users soon with a full launch expected reasonably soon.

That’s to help the company stay ahead of the curve, which it has done in previous months, beating Google to the mark in terms of AI. In the meantime, reports that Google is testing generative AI across Docs and Gmail only add to the pressure to get it right, and get it done quickly.

• Here’s our pick of the best online collaboration tools

Via Windows Latest

Businesses are putting themselves at risk of all kinds of cyber-attacks due to poor practices when it comes to educating and training the workforce.

A new report from Yubico, found less than half (42%) of UK businesses it surveyed held mandatory, frequent, cybersecurity training. 

There are many things employees could be taught, which would improve the cybersecurity posture of organizations, the report further suggested. For example, roughly half (47%) often write down, or share their passwords - which is one of the most common mistakes when it comes to safeguarding a password. 

Resetting the password 

Elsewhere, the report found that many workers (33%) allow other people to use their work-issued device, while more than half (58%) use personal devices for work. 

A similar percentage (49%) do vice-versa, as well, by using a work-issued device for personal use, which is another cybersecurity red flag. Finally, half (48%) have been exposed to a cyberattack such as phishing, without reporting the incident to their IT and cybersecurity teams. 

Even when an employee gets exposed to a cyberattack, their organization does very little to amend the issue. “Very few” companies implemented phishing-resistant cybersecurity methods in response to being targeted, a third (28%) simply had their passwords reset, and just a quarter (28%) were made to attend cybersecurity training. 

“Cyber attacks, and how to prevent them, should be top of mind for every organization. However, our research reveals a remarkable disparity between the risks of cyber-attacks and businesses’ attitudes toward them,” commented Niall McConachie, regional director (UK & Ireland) at Yubico.

For McConachie, businesses should deploy multi-factor authentication (MFA) as soon as possible, and consider FIDO2 security keys. The latter “have been proven to be the most effective phishing-resistant option for business-wide cybersecurity”, he says. 

“By removing the reliance on passwords, MFA and strong 2FA are more user-friendly and can be used for both personal and professional data security. This is especially important as cyber-attacks are not limited to companies but can directly target customers and employees too.”

One of the most used-used passwords - “123456” - is still in use today, despite being known by virtually every cybercriminal out there, the report concluded.

• Here are the best firewalls 

Prusa Research has launched its much-anticipated Original Prusa MK4 3D printer, which promises to deliver high-speed printing with ease.

The company says the MK4 is more than just an upgraded version of the popular MK3 (despite the similar looks) with 90% of the machine being all-new and offering a host of features and improvements.

Prusa says this makes it look likely to be one of the fastest and most reliable 3D printers on the market today.

Original Prusa MK4

In a call with Josef Průša, I asked why the company had decided to stick with the iconic design of the Original PRUSA's that had come before. 

The answer was straightforward - the Original Prusa customer base is loyal, and to cut off the development for those existing users now is not an option, especially when the performance of the MK4 can more than compete with the latest Core XY machines when it comes to speed and speed reliability. A clear upgrade path must be part of the PRUSA model, as does ensure that the open-source community is maintained.

While the new MK4 machine may look familiar, the upgrades enhance almost every aspect of the build and technology. Most noticeably, the metal frame of the MK3 that was machined from sheet metal is replaced by an all-new aluminum composite injection molded frame, which provides greater rigidity.

Among the other most notable upgrades is the new high-speed extruder, which allows for faster printing speeds while maintaining high-quality prints. This extruder is compatible with many filaments, including PLA, ABS, PETG, etc.

This new print head has been developed by PRUSA Research and based on the Nextruder that we first saw on the XL, and now the smaller version that features on this new printer again features quick swap nozzles. 

For many, the most groundbreaking feature is the new fully automatic first layer calibration, thanks to a Load Cell sensor. This means that even if you swap out the default nozzle for any standard E3D V6 (with an adapter), the probe will still automatically set the first layer height accurately. This means that the need for fine Z-axis adjustment should be history. 

Another feature that will appeal to many users is the upgraded connection options with ethernet and wifi. This means there will be no more toing and froing from the machine to load or update files through production iterations. 

A new print bed features an improved surface texture, providing better adhesion and eliminating the need for adhesives or other surface treatments. Regarding materials as standard, there's wide compatibility, and specialist nozzles can be fitted with the quick swap system. In addition, the MK4 has a new filament sensor that detects when the filament runs out and automatically pauses the print, allowing users to easily swap out filaments without starting the print job all over again.

Prusa Research has also enhanced the printer's firmware to the next generation and fully used the 32-XBuddy control board, which now includes advanced thermal runaway protection and automatic skew axis calibration. This ensures the printer operates safely and efficiently, requiring minimal manual intervention.

But perhaps the most significant aspect of the MK4 is its commitment to open-source principles. Prusa Research has a strong heritage in the open-source community and is dedicated to keeping its printers and software open and accessible to everyone.

Open source backing

In a recent blog post, Prusa discussed the importance of open source in the 3D printing world and expressed his concerns about the current state of open source development.

"We stand by our roots in open-source and will continue to do so," he wrote. "Our desktop 3D printers will always be open source. We intend to continue publishing plastic parts and firmware source codes."

Prusa Research's commitment to open source is evident in the MK4, which has models of all plastic parts and firmware sources available for download. While the electronics plans have not been released yet, the company has indicated that it plans to release them under a new license that addresses some issues facing open-source development.

Overall, the Original Prusa MK4 is an impressive 3D printer that builds on the success of its predecessor while introducing a host of new features and improvements. With its commitment to open-source principles, it's a printer that will appeal to makers, hobbyists, and professionals alike.

• Also check out our list of the best all-in-one printers out there