AI chatbots can help plan a vacation or suggest what to pack. But there are situations where advice from someone who actually knows what they are doing is the safer choice.

As reported by the Chicago Tribune, three novice hikers from Roseville, California, had to be rescued from Mount Shasta after becoming stranded overnight. According to the Siskiyou County Sheriff’s Office, they had used Google Gemini to help plan the climb. The sheriff’s office said Gemini advised the group to carry far less food and water than they needed.

An eight-hour hike turned into an overnight rescue

The group started toward the summit at around 3 a.m. on Sunday. Hikers are advised to turn around if they have not reached the summit by noon, but the three continued and eventually reached the top at around 7 p.m.

Things went wrong on the descent. They lost their route in the dark and ended up in Mud Creek Canyon instead of staying on the Clear Creek Route. One of the hikers had also injured his knee, forcing the group to spend the night in a steep drainage before rescuers reached them the following morning.

The sheriff’s office advised hikers to carry multiple methods of navigation, including an offline option, and recommended calling the local U.S. Forest Service ranger station before attempting the climb. It also specifically warned people not to rely solely on AI for trip planning.

AI can give advice, but it cannot replace experience

Similar problems have surfaced when people turned to AI for legal and medical advice. We previously covered a lawyer who used ChatGPT for legal research, only to discover that the cases it cited did not exist. He even asked ChatGPT to confirm they were real, and the chatbot said they were. A study testing ChatGPT, Gemini, Grok, Meta AI, and DeepSeek also found that roughly half of their answers to medical questions were flawed or incomplete, while some included fabricated references.

Chatbots have their uses, but getting a plausible answer is not good enough when a mistake could put someone’s health, safety, or livelihood at risk. Advice from a professional or experienced expert should come first. At the very least, AI-generated information should be checked against reliable sources before anyone acts on it.

Even the chatbots themselves tell users not to take everything they say at face value. Gemini carries the warning, “Gemini is AI and can make mistakes.” Cases like this are a good reminder of why that warning is there.

Apple’s Time Machine tool removes the need to store everything online.

Setting up a drive for Time Machine

Adnan Ahmed for Engadget

How Time Machine works

Adnan Ahmed for Engadget

How far back can you go with Time Machine?

Adnan Ahmed for Engadget

AI AND ML

15 Mustangs on the city streets, but the Nissan LEAF follow-up could dispense with the safety driver

Wayve and Uber have launched self-driving rides in London, but what’s more interesting is what is planned for the next few months and years.

As of today, the plan is for 15 electric Ford Mustangs, equipped with the Wayve AI driver, to roam the capital, picking up passengers with a safety driver on hand to keep an eye on things.

However, at a launch event on Thursday, Wayve showed off the follow-up to the Mustang: a Nissan LEAF, which will run the Wayve AI Driver using NVIDIA DRIVE Hyperion. This glimpse into the future was arguably more intriguing than today’s electric Mustang.

Unlike the Mustang, the LEAF’s self-driving equipment looked as though it might have been part of the car’s design, from a roof-mounted sensor pack that followed the roofline, through to sensors attached to the wings and bumpers.

A Wayve representative told us it had worked with Nissan on the integration. That said, the tech is relatively agnostic when it comes to platforms and sensor configurations.

Wayve was, however, a little less forthcoming at the launch event about the underlying technology changes between the Nissan-based prototype on show and the Mustang parked up in front of it. The hope is that the update will reach UK roads next year, although that depends on several factors, including technology and regulations. Right now, Uber stated, “The initial launch phase features supervised autonomous rides, meaning a trained and TfL licensed private hire driver is onboard to oversee the trip.”

So no San Francisco Waymo-esque rides with the driver’s seat spookily empty just yet. However, a Wayve representative told us the Nissan LEAF would eventually be capable of such a feat.

The rollout of the Mustang feels a little like a stunt to steal a march on competitors such as Waymo, which has also had vehicles rolling around the tricky streets and traffic of London. Waymo is presently in a testing phase, but Wayve and Uber are now taking paying passengers. Both services, however, must have a trained driver onboard, poised to intervene if the need arises.

The untried nature of the Wayve and Uber Mustang is evidenced by the words “TEST VEHICLE. KEEP A SAFE DISTANCE” printed on the rear of the car. Paying passengers are part of the AI Driver’s learning process, with lessons learned feeding into the next generation of autonomous vehicles, which might, or might not, still need that safety driver behind the wheel, just in case.

Wayve was founded in 2017 and is headquartered in London. It has trained its AI tech on the city’s roads since 2018. In 2024, Wayve partnered with Uber, which included an investment to deploy vehicles with Wayve’s technology in 12 markets worldwide, starting with London. The LEAF will first see action in Tokyo. ®

What just happened? Google introduced its Gemini Audio AI models earlier this year, during the latest developer-focused I/O conference. Now, the advertising corporation is going to bring these additional features to three of its most popular mobile apps for Workspace customers.

Google is significantly expanding Gemini Live’s reach to Gmail, Docs, and Keep, at least for paying customers of the enterprise-focused Workspace service. The Alphabet company highlights how these new capabilities will enrich its customers’ day-to-day lives with a new option to use their apps and personal data, but not all users seem ecstatic about the idea of talking to a mail client on the phone.

The newly announced features are based on Gemini Live’s conversational AI models, so they are conveniently named Gmail Live, Docs Live, and Keep Live. Thanks to Gmail Live, customers can quickly recall important details about a meeting or their kids’ school events. They will just need to ask the app what they want in natural language, and Gmail will (hopefully) bring the right details to the conversation.

Docs Live does not necessarily work on pre-existing data. The AI feature is designed to be a “co-writer” for new document projects, helping users lay out a first draft based on “rough” thoughts and ideas. Furthermore, Docs Live can also extract “relevant” details from other Workspace apps (Gmail, Drive, Chat, etc.) if the user grants specific permissions to do so.

Finally, Keep Live is conceived as a novel way to turn whimsical, undefined notes and ideas into more organized and therefore “actionable” notes. Google is inviting users to “dump” their mind’s content in the company’s note-taking app, and then ask Gemini Live’s AI engine to turn a stream of consciousness into a more useful list of data or ideas. Google added that Gemini should be able to understand everything being said.

Gemini Live’s new conversational features for Gmail, Docs, and Keep are already rolling out to eligible subscribers of Google AI Plus, Pro, and Ultra plans. Furthermore, they should arrive soon for business customers using Google’s Workspace productivity suite.

While some people consider talking to a mobile app a rather natural way to interact with computers and personal devices, others see it differently. Some early reactions to Google’s new conversational features highlight the ever-growing number of AI tools that are now being proposed to interact with Android apps, while one disgruntled user said that “chatting” with an email app is definitely not the most appealing thing to do on a phone.

In this week’s Sunday Reboot, the sudden top-table rule changes and how John Ternus will be a good Apple CEO, whether he likes it or not.

Sunday Reboot is a weekly column covering some of the lighter stories within the Apple reality distortion field from the past seven days. All to get next week underway with a good first step.

More than one job change

September is always a big month for Apple, but this one is bigger than ever. Not because of the inbound hardware, but because of the people making it.

The key change is at the top of the Apple executive food chain, as the multi-month transition process has almost concluded. John Ternus is now the CEO, taking over from Tim Cook, but that doesn’t mean the process has ended.

Of course, there’s some behind-the-scenes logistics happening, such as Cook moving out of his office and into a new one for the leadership-adjacent position of executive chairman of the board.

Ternus, meanwhile, will be watching some hardware lackeys shifting cardboard boxes full of circuit boards and unused tech from his old office to his new one. It’s a big move too, as he will be among his fellow executives instead of near his beloved hardware teams.

It’s a thing that has to happen, as Ternus takes the mantle. Though inevitably other unintended changes are also underway too.

Apple Fellow Phil Schiller

Advertisement

Apple Fellow Phil Schiller has decided to reduce his workload. He’s switched out of running the App Store for a more relaxed existence within Apple, presumably taking another step to retirement.

Then there’s Luca Maestri, who was the previous CFO but continued managing Corporate Services. Except he’s not doing that anymore and current CFO Kevan Parekh has picked up those duties too.

We know that seismic change can prompt other employment updates, and a new CEO would certainly count as one. But there was no serious expectation of a mass exodus as Ternus took his seat.

There still isn’t that expectation, but it doesn’t stop things like this from happening. We do know that the Schiller change was in the last days of Cook’s management, but Maestri’s decision took place a few weeks ago, long before the CEO switch.

It’s probably safe to say that the two changes are not massively consequential to Apple.

At least, Schiller’s App Store withdrawal could be a bit worrying, until you consider it’s only part of a larger Services division. Maestri’s post-CFO role was also not what you would call critical to Apple’s success, or at least wasn’t as critical as managing a massive segment of the company.

Besides, this CEO transition has been going on for months. Anyone who would’ve seriously objected to Ternus being in control has had ample time to get out of Dodge.

The people who are still leading Apple are going to hang around for quite a while longer.

Ocean liner management

As the opening weeks and months of Ternus control continue, there may be the threat of more managerial shake-ups. But it certainly won’t be from radical moves by the CEO.

You can expect him to make small changes to how the company functions. I say small in a relative fashion, because Apple is a tech giant.

His changes need to be like a captain steering a cruise ship. It takes time to slow down and to adjust course, as rocking the boat of that size would be dangerous.

That is to say, anything he decrees will take a long time to become an actual problem for someone. If a problem does indeed manifest down the road.

Then there’s the corporate culture, since any decision Ternus makes will affect thousands of people. Even though the CEO’s word is effectively gospel, if there are enough people who disagree, there’s always the risk of a mutiny.

Apple CEO John Ternus [left] and previous CEO Tim Cook [right]

There is one more safety barrier in place to prevent Ternus from creating a Titanic-like situation. Cook is still hanging around.

Sure, he’s not going to have much of a say on what products Apple will develop. Nor should you expect him to appear during an Apple event video in the future.

But it’s hard to get away from the thought that Ternus won’t have free rein without Cook’s oversight or advice.

Cook has a wealth of experience as captain of the SS Cupertino, and it would be foolish for Ternus not to listen to anything Cook says.

You’re not going to get any insane company-destroying proclamations coming from Ternus. Anything that he does say will have been discussed and debated to an exhaustive degree to keep it sane and safe.

Ternus is a safe pair of hands. But even if he wasn’t and was hell-bent on hitting an iceberg, Cook would still be a sanity safety net.

Last week’s Sunday Reboot discussed Apple’s early Mac upgrades and John Ternus meeting Pikachu.

Disclaimer: Unless otherwise stated, any opinions expressed below belong solely to the author. Data sourced from the Ministry of Manpower.

Singapore’s Ministry of Manpower has just updated its probably most accurate statistical list of salaries in Singapore, which is used to benchmark foreigners applying for local Employment Pass.

These tables are based on salaries drawn by local PMETs, which are used to gauge how many points on the COMPASS scale foreign EP applicants can be granted.

As a result, we are provided with a very detailed breakdown of how much Singaporeans earn by age and industry.

I have extracted the incomes of the Top 10% of workers in each of them, to show you how much you could potentially earn if you’re among the elite in your particular business.

Salaries at the 90th percentile (top 10% of workers) among local PMETs

(Professionals, Managers, Executives, and Technicians)

At the bottom end, the best young PMETs in Accommodation can still earn nearly S$6,300 per month, which is above the national median of S$5,775.

The top end is set by 45+ year olds working in specialised Fund Management Services, within the financial sector, where they make over S$500,000 per year.

Interestingly, 1 in 10 of the 25-year-olds in that industry is already paid over S$15,000 per month. That said, Fund Management is a bit of an outlier, so let’s look at other industries.

Six-figure annual incomes are possible everywhere in your mid-30s, within a decade or so after graduation. If you’re really good, your financial situation should be quite comfortable by then.

At 40 years old, every single one of them pays S$10,000 per month or more (Food & Beverage Services come in at S$9,890, close enough to qualify). In Finance, Insurance, IT, Professional Services and Air & Sea Transport, you can aim for S$20,000 and above. In Wholesale Trade, you can get there by age 45.

The most revealing finding here is that there is decent money to be made in every sector of Singapore’s economy. None of them is significantly lagging behind the rest, although, of course, the most money is offered to those working in skyscrapers down in the CBD.

But you don’t have to be a banker to bring home more than S$100,000 each year. There’s potential for that everywhere (as a PMET), and you don’t have to slave away into your 50s to get there.

You just have to be in the Top 10% in your industry.

• Read other articles we’ve written on Singaporean businesses here.

Featured Image Credit: Alamy

OLED has already conquered the premium gaming-monitor market. Now Alienware wants to take it somewhere it has historically struggled to crack: competitive esports. I got hands-on with the new Alienware AW2527HX at IFA 2026, and with a 560Hz fifth-generation Penta Tandem QD-OLED panel, a design shaped around professional players, and some wonderfully obsessive esports-focused features, Dell is making a pretty convincing case that OLED’s next battlefield is competitive gaming.

This is an esports monitor first, OLED monitor second

The timing couldn’t be more interesting. ASUS has just made its ROG Swift OLED PG259QWS Ace the official monitor for the upcoming PGL CS2 Major Singapore 2026. That monitor pushes 1080p OLED all the way to 720Hz on a glossy Tandem WOLED panel. Alienware, meanwhile, is taking a slightly different route with 560Hz QD-OLED. Either way, the message is pretty clear: OLED is starting to knock on the door of professional esports.

And this isn’t some budget 1080p monitor that happens to have OLED slapped onto it. Alienware is targeting serious competitive gamers, whether they’re already professionals or aspiring to become them. The AW2527HX is a 24.5-inch Full HD panel running at 560Hz, with a claimed 0.03ms gray-to-gray response time. Underneath that is a Penta Tandem QD-OLED panel with an RGB-stripe layout, bringing OLED’s infinite contrast and true blacks to a category that has traditionally been dominated by TN displays. Alienware also claims up to 1,500 nits of peak brightness, while the matte anti-glare low-reflectance (AGLR) coating is designed to keep the image clear even under bright tournament lighting.

At 25 inches, Full HD is still surprisingly crisp, too. And once you actually start playing, that 560Hz refresh rate makes everything feel ridiculously smooth. I tested PUBG: Battlegrounds during my demo and, despite being more of a Valorant player myself, I came away pretty impressed. OLED has long deserved a place in competitive gaming, and this monitor is one of the better arguments I’ve seen for it.

The only catch? Don’t expect it to be cheap just because it says 1080p on the box. I was told to expect pricing similar to its 32-inch OLED monitors, which puts it firmly into premium-monitor territory. Dell hasn’t officially announced the final MSRP yet, but that price alone tells you who this monitor is really for.

The stand gets weirdly serious

Alienware has made some very thoughtful decisions around the physical design, too. The base rotates smoothly and includes markings showing exactly what angle you’re using. That’s a tiny detail, but when your entire gaming setup is built around replicating a very specific competitive posture, tiny details suddenly matter a lot.

The arm itself is even more interesting. It’s essentially an inclined telescopic stand, allowing you to push the monitor all the way down and then tilt it backwards into that aggressively angled position many esports players prefer. Think of it like a tablet lying almost flat on your desk, except tilted upwards towards your face. That leaves plenty of space underneath for your keyboard and mouse.

You can obviously achieve something similar with other monitors, but usually that means buying a separate monitor arm and fiddling around with your setup. Alienware has simply built the functionality into the bundled stand. The monitor also gets a 155mm height adjustment range and up to 45 degrees of backward tilt.

And then there’s the front. Or rather, the complete lack of anything on it. There’s no giant Alienware logo staring back at you, no distracting gamer aesthetic, and no unnecessary visual nonsense. Dell told us that its work with Team Liquid directly influenced this decision: professional players didn’t want branding or anything else distracting them from the game. So all you really see is the screen.

True Stretch feels almost like cheating

My favourite feature, however, might be the one that sounds the least impressive on a spec sheet. There’s a dedicated eSports button on the back that lets you instantly switch to the monitor’s True Stretch resolution. During my demo, that meant jumping straight to 1280 x 960 without having to dive into Windows settings, GPU control panels, or configuration files.

On the Alienware, you just press a button, and you’re there!

If you’re into competitive shooters, you’ll know why that’s useful. Stretched resolution doesn’t actually change hitboxes, but it does make player models, and especially heads, appear wider on screen, which can make targets feel easier to track and hit. That’s a big reason why so many competitive players swear by 4:3 or 3:2 stretched setups. On the Alienware, you just press a button, and you’re there. It almost feels like cheating. And honestly, the fact that the whole thing worked flawlessly during my demo might be my favourite part.

Why does my OLED suddenly look like a TN?

Alienware also throws in dedicated presets for Counter-Strike, Valorant, and Apex Legends, with more customization available through the Alienware Command Center. But then there’s TN Mode, which had me doing a double take. Why on Earth would I buy an OLED monitor and then make it look like a TN panel?

The TN mode looks pretty shitty. But that’s precisely the point!

The answer, surprisingly, makes sense. Turn TN Mode on, and the colours become noticeably duller. You see, competitive players have spent years using TN displays because of their speed, and jumping straight from that familiar, washed-out look to OLED can be a bigger adjustment than you’d expect. TN Mode essentially acts as a bridge. You can get comfortable with the new monitor, keep the visual character you’re used to, and gradually make the leap to proper OLED. It’s a weird feature, but one that shows Alienware has actually thought about the people it’s trying to win over.

There’s also DisplayPort 2.1 UHBR13.5, which provides enough bandwidth to run the AW2527HX at its native 1080p and 560Hz without Display Stream Compression. For most people, that’s just another acronym to scroll past. For the sort of competitive gamer who counts every millisecond and probably has opinions about which direction their Ethernet cable should face, it’s exactly the kind of detail that matters.

OLED has entered the arena

That’s what ultimately impressed me about the AW2527HX. Alienware isn’t just throwing a 560Hz refresh rate at an OLED panel and calling it a day; it’s clearly built around how competitive players actually play, right down to the ridiculous stand and that wonderfully convenient True Stretch button. It’s specific, obsessive, and very much unapologetic about who it’s for.

And after my first hands-on, I’m starting to think Alienware might be onto something. OLED has finally earned a seat at the esports table, and the AW2527HX makes a surprisingly strong case for it. Now I just want to see what happens when I get to really put those 560Hz pixels through their paces.

off-prem

No restoration time for platform used by more than 30,000 UK organizations, including NHS trusts and medical practices

UPDATED Docmail, a service used by more than 30,000 UK businesses and public sector organizations, has been offline for several days, prompting concern among users.

Operated by CFH Docmail Ltd, the cloud-based hybrid mail service allows users to send physical letters without running their own print rooms or making trips to the post office. In other words, it handles the printing and postage of business communications.

However, Docmail has been unavailable since the start of business on September 2, according to its status page. It initially warned that the service was “experiencing slower than usual processing times,” before saying soon afterward:

“Docmail is currently unavailable while we carry out essential maintenance. Our technical team is working to complete the maintenance and restore full service as quickly as possible. We apologise for any inconvenience caused and will provide a further update once the system is available again.” 

Throughout the rest of the day, techies continued to deploy “required updates,” and the service remained unavailable as work took place to “ensure the service can be brought back online safely and reliably.”

A further update on September 3 said Docmail was still “temporarily unavailable while our teams continue to work to restore the service.” The company thanked customers for their patience and apologized for the “disruption.” At the end of the day, it admitted: “We do not yet have a confirmed restoration time. Work remains ongoing as a priority.”

By Friday – day three of disruption – the service remained unavailable, with techies still toiling to restore it and promising further updates when more information became available.

Our attention was drawn to the Docmail debacle by Reg readers concerned about the length of the outage and the lack of a restoration time.

We contacted CFH Docmail for details of the disruption’s cause and when it expects the service to be restored.

CFH says many of its more than 30,000 business and public sector customers operate in UK healthcare. It claims that more than 75 percent of medical practices use Docmail for bulk mail and individual patient mailings, such as appointment reminders.

Customers also include NHS Trusts such as Croydon Health Services, while dental practices, pharmacies, opticians, and medical research networks also rely on the service, as do some local authorities and businesses.

While the service remains offline, important communications such as health appointment reminders and letters from local councils may go unsent. ®

Updated at 12.59 UTC on September 4, 2026, to add:

Following publication of this article, a spokesperson at CFH Docmail Ltd, told us: 

“We have been made aware of an issue impacting an isolated server, used by only one client. Out of an abundance of caution and in line with our security protocols we took the decision to take Docmail offline. This proactive step is why the system has been temporarily unavailable.

“Our internal team has been working closely to investigate what happened and to ensure that when all systems are switched back on, it will be entirely safe to do so.

“This process does take time, and we are acting on the advice of external consultants to ensure we are taking the right precautions. While Docmail is ready to be reinstated, we are waiting completion of the external review before doing so. We apologise for the disruption this has caused to our customers and are working to restore service in the quickest but safest way possible.”