Blogroll

SAVE $150: As of August 7, get the Apple iPad Air with M3 chip for $449 and save $150. That knocks 25% off its $599 price tag.

Opens in a new window Credit: Apple Apple iPad Air, 11-inch (M3 chip, 128 GB, WiFi) $449 at Amazon
$599 Save $150 Get Deal

The line between Apple's tablets and laptops is becoming strikingly thin. Now that the brand has put its M series of chips into both iPads and MacBooks, both devices offer similar processing speeds. So if you're looking for a new laptop or tablet as you head back to school, iPads are a great candidate, especially the iPad Air. And there are plenty of deals on Apple iPad models today.

As of August 7, get the iPad Air with M3 chip for just $449. That knocks $150 off its list price of $599 for 25% off. This isn't the first iPad deal we've spotted this week, as Amazon has dropped the price on the 10th generation model down to $349 and the iPad Mini is down to $399. But what makes the iPad Air so special?

SEE ALSO: Shop Apple's most affordable iPad for its lowest price yet

Among the iPads currently on sale, the iPad Air is the most expensive. It's just about the same size as the iPad 10th generation, but it has a much better core processor. The iPad 10th generation and iPad Mini both use Apple's A series chips, which are zippy and great but made for iPhones. Meanwhile, the iPad Air is using the M3 chip, which goes into the brand's MacBooks. That more advanced processor means you can get more work done. It's a better fit for creatives looking to edit photos and video or do graphic design work on the tablet.

Right now, get the iPad Air with M3 chip, a tablet built for creatives, for just $449 and save $150.

SAVE $150: As of August 7, get Apple's 10th generation iPad (A14 chip, 256 GB, WiFi) for just $349 and save $150. That saves 30% off its $499 price tag.

Opens in a new window Credit: Apple Apple iPad, 10th generation (A14 chip, 256 GB, WiFi) $349 at Amazon
$499 Save $150 Get Deal

As the summer winds down and we think about heading back to school, it might be time to think about a new device for the school year. If you're looking for a tablet for note-taking, or even to use as a laptop alternative, iPads are the everyman tablet, fit for both work and play. Even the most affordable model, the iPad 10th generation that was released in 2022, allows you to run multiple apps at once so you can multitask. And right now, it just dropped to its lowest price ever.

As of Aug. 7, the Apple iPad, 10th generation (A14 chip, 256 GB, WiFi) just dropped down to $349. That's the tablet's lowest price ever — even cheaper than it was during Prime Day. That shaves $150 off of its $499 list price for a 30% discount.

SEE ALSO: Need a new tablet? The Samsung Galaxy Tab A9+ is under $160.

The Apple iPad usually starts at 64 GB of storage, but this upgraded model comes with 256 GB, plenty of space to download music, shows, and photos. If you're planning on using the device as a laptop, with the Magic Keyboard Folio, it makes it easy to type up notes fast and on-the-go. Better yet? The Magic Keyboard Folio is 20% off at Amazon, bringing it down to $199.

Shop the Apple iPad, 10th generation with 256 GB of storage for $349 now and save $150.

President Donald Trump has signed an executive order to make it easier to include cryptocurrency in Americans' 401(k) retirement funds. It's a potentially risky move that has been criticised by financial experts.

SEE ALSO: Stablecoin bill advances in U.S. Senate as Trump critics call to end his crypto dealings

Announced on Thursday, Trump's new executive order instructs the U.S. Department of Labor (DOL) to reexamine its guidance regarding 401(k) plans' investment in alternative assets such as cryptocurrency, real estate, and private equity. Trump's order further directs the Securities and Exchange Commission (SEC) to revise regulations and guidance in order to facilitate such alternative asset investment.

In short, this means that the U.S. government may soon overhaul regulations in order to help 401(k) funds invest in crypto. Cryptocurrency prices surged in the wake of Trump's executive order, with investors anticipating it will lead to more widespread adoption of the currency.

While 401(k) plans weren't explicitly banned from investing in cryptocurrency previously, in 2022 the DOL cautioned that those considering doing so should "exercise extreme care." Expressing "serious concerns" about the wisdom of such investments, the DOL warned of a slew of hazards associated with cryptocurrency, including that such assets are highly speculative, extremely volatile, and face issues regarding recordkeeping and valuation. 

The Trump administration subsequently rescinded this guidance this May, claiming that the DOL's advice to be careful with highly risky financial investments was a case of "overreach" by the previous Biden administration. Instead, Trump's DOL stated that it had a neutral stance toward 401(k) plans investing in cryptocurrency.

Thursday's executive order to revise regulations now indicates a more supportive view of 401(k) cryptocurrency investment, with Trump pledging to transform the U.S. into the "crypto capital of the world."

Though the White House claims assets such as cryptocurrency "offer competitive returns and diversification benefits," financial experts have warned against relying on such assets to fund your retirement. Alicia H. Munnell, a senior advisor at Boston College's Center for Retirement Research, called such 401(k) investment a "terrible idea," likening it to gambling and noting that it is unlikely to improve returns.

"Participants don’t understand the product, it’s a speculative and volatile investment, straying from traditional investments is unlikely to enhance returns, and it’s probably not a prudent option for 401(k)s," Munnell wrote in response to the DOL's withdrawal of its 2022 guidance. "DOL should not be opening the door to this type of activity."

Numerous cautionary tales of nosediving values, scams, hacks, and heists have made it clear that cryptocurrency is far from a safe investment. Despite this, tales of swift, dramatic gains continue to tempt investors and build significant hype. Trump's executive order may encourage more people to venture into cryptocurrency, but whether it will be to their benefit remains to be seen.

TL;DR: Create anything, even NSFW art, with a lifetime subscription to Imagiyo for only £36.50.

Opens in a new window Credit: Imagiyo Imagiyo AI Image Generator: Lifetime Subscription (Standard Plan) £36.50
£368.69 Save £332.19 Get Deal

Digital creativity has never been more accessible, yet many of us remember the days when crafting a single image meant wrestling with layers and plugins for hours on end. Now there’s a way to generate stunning visuals in seconds simply by typing a description of what you have in mind.

Imagiyo uses Stable Diffusion AI alongside FLUX AI to turn text prompts into high-quality images ready for commercial use, and there aren’t many limits to what you can create. Here’s what that means.

What art can you make with Imagiyo

What do you want to make first? It only takes a brief description to put Imagiyo’s advanced algorithms to work, and unlike other image generators, Imagiyo actually lets you really follow your creativity. Craft stunning landscapes, visualise characters from books, or go for something a little more daring. Imagiyo supports NSFW content creation. Just set your prompts to private and let your mind run wild.

Imagiyo’s commercial-use license means you can take some of the images you generate and incorporate them into client projects, social media campaigns, or personal portfolios without fear of copyright issues.

Each month, you receive 500 image-generation credits and can submit up to two prompts at once. Unused credits roll over, so you never lose access to your creative potential. Best of all, Imagiyo delivers your purchased engine updates and feature improvements automatically, ensuring you always work with the latest AI models.

Get an Imagiyo AI Image Generator lifetime subscription for £36.50.

StackSocial prices subject to change

Yesterday we listed five research-backed strategies that can lead to a happier life—even for busy professionals. Here are five more:  Act nicely. Agreeableness is consistently found to be highly and positively correlated with happiness, and it can be increased relatively easily.  Be generous. Behaving altruistically toward others rewards the brain with happiness-enhancing boosts of dopamine, […]

257257

If you like playing daily word games like Wordle, then Hurdle is a great game to add to your routine.

There are five rounds to the game. The first round sees you trying to guess the word, with correct, misplaced, and incorrect letters shown in each guess. If you guess the correct answer, it'll take you to the next hurdle, providing the answer to the last hurdle as your first guess. This can give you several clues or none, depending on the words. For the final hurdle, every correct answer from previous hurdles is shown, with correct and misplaced letters clearly shown.

An important note is that the number of times a letter is highlighted from previous guesses does necessarily indicate the number of times that letter appears in the final hurdle.

If you find yourself stuck at any step of today's Hurdle, don't worry! We have you covered.

SEE ALSO: Hurdle: Everything you need to know to find the answers Hurdle Word 1 hint

A hauler.

SEE ALSO: Apple’s new M3 MacBook Air is $300 off at Amazon. And yes, I’m tempted. Hurdle Word 1 answer

MOVER

Hurdle Word 2 hint

To brag.

SEE ALSO: Wordle today: Answer, hints for August 4, 2025 Hurdle Word 2 Answer

BOAST

Hurdle Word 3 hint

Signaling.

SEE ALSO: NYT Connections Sports Edition today: Hints and answers for August 4 SEE ALSO: NYT Connections hints today: Clues, answers for August 4, 2025 Hurdle Word 3 answer

CUING

Hurdle Word 4 hint

Very good.

SEE ALSO: NYT Strands hints, answers for August 4 Hurdle Word 4 answer

DANDY

Final Hurdle hint

Not a nephew.

SEE ALSO: Mahjong, Sudoku, free crossword, and more: Games available on Mashable Hurdle Word 5 answer

NIECE

If you're looking for more puzzles, Mashable's got games now! Check out our games hub for Mahjong, Sudoku, free crossword, and more.

We're a day away from the full moon, so let's see what's happening in the current lunar cycle.

This is a series of eight unique phases of the moon's visibility. The whole cycle takes about 29.5 days, according to NASA, and these different phases happen as the Sun lights up different parts of the moon whilst it orbits Earth. 

So, what's happening with the moon tonight, Aug. 8?

What is today’s moon phase?

As of Thursday, Aug. 8, the moon phase is Waxing Gibbous. According to NASA's Daily Moon Observation, the moon will be 99% lit up tonight, we're so close!

So, what can you spot when you look up tonight? With no visual aids, you'll be able to see lots on the moon's surface, namely the Mare Imbrium, the Kepler Crater, and the Mare Vaporum.

With binoculars, you'll also be able to see the Mare Humorum, the Endymion Crater, and the Posidonius Crater. Add a telescope to see the Descartes Highlands, the Schiller Crater, and the Rupes Altai.

When is the next full moon?

The next full moon will be on August 9. The last full moon was on July 10.

What are moon phases?

According to NASA, moon phases are caused by the 29.5-day cycle of the moon’s orbit, which changes the angles between the Sun, Moon, and Earth. Moon phases are how the moon looks from Earth as it goes around us. We always see the same side of the moon, but how much of it is lit up by the Sun changes depending on where it is in its orbit. This is how we get full moons, half moons, and moons that appear completely invisible. There are eight main moon phases, and they follow a repeating cycle:

New Moon - The moon is between Earth and the sun, so the side we see is dark (in other words, it's invisible to the eye).

Waxing Crescent - A small sliver of light appears on the right side (Northern Hemisphere).

First Quarter - Half of the moon is lit on the right side. It looks like a half-moon.

Waxing Gibbous - More than half is lit up, but it’s not quite full yet.

Full Moon - The whole face of the moon is illuminated and fully visible.

Waning Gibbous - The moon starts losing light on the right side.

Last Quarter (or Third Quarter) - Another half-moon, but now the left side is lit.

Waning Crescent - A thin sliver of light remains on the left side before going dark again.

Warning: This contains spoilers for the ending of Weapons.

Well, that was intense, wasn't it?

Zach Cregger's Weapons is every bit as twisty and horrific as the trailers have suggested, with an ending that answers many questions while leaving others a bit more open-ended.

SEE ALSO: 'Weapons' review: You're not ready for Zach Cregger's wild new horror film

From the missing children to the whole "weapons" motif, we've tried to unpack everything below.

What's Weapons about?

The Barbarian writer/director's new film has one hell of a premise: A town is thrown into chaos after an entire class of school children — minus one boy — goes missing on the same night at the exact same time. Doorbell camera footage shows the children leaving their homes seemingly of their own free will, running out into the night with their arms held out at the sides in a kind of creepy flying V. Nobody knows where they've gone, or why they've gone.

The aftermath of this incident leaves the townsfolk blaming each other as they look for answers. The film follows a few of these struggling figures, including the class teacher — and now town pariah — Justine (Julia Garner); a father of one of the missing boys, Archer (Josh Brolin); school principal Marcus (Benedict Wong); and Alex (Cary Christopher), the only member of Justine's class that didn't disappear.

What happens at the end of Weapons? Justine and Archer, on the trail of the missing kids. Credit: Warner Bros.

As the film progresses we learn that the figure responsible for the missing children is a witch named Gladys (Amy Madigan), Alex's aunt who has been using a mysterious potted tree — combined with personal effects and her own blood — to control the people around her. Gladys, who is terminally ill when she moves in with Alex's parents, is in some way able to restore herself back to full health by using the people she takes hold of. After Alex returns from school with name tags taken from his classmates' cubbyholes, she incorporates these into a ritual that draws all of the children to their house at once. Gladys keeps them down in the basement, then continues using the tree to cover up what she's done while Alex is forced to help her.

In the end, though, she underestimates those around her. Justine and Archer trace the missing children back to her house, and Alex is able to use her own tree against her: He escapes from his room and uses the stick that controls his classmates to turn them on her, with the children chasing her down in the street and literally ripping her to pieces. Archer rescues his son, Alex has his parents back, and the town is able to return to some version of normal.

Are the kidnapped children safe at the end? Poor little Alex doesn't have a great time in "Weapons." Credit: Warner Bros.

Yes and no. On the plus side, all of the children are found alive and are able to return home. We can tell from the blank looks on their faces after they've killed Gladys, however, that they haven't immediately reverted to their old selves. There is clearly residual trauma from the spell they've been placed under.

The movie's voiceover at the very end reenforces this. We hear that Alex is living with a new family and that his parents still eat soup (the food Gladys forced him to feed them while they were under her spell), and that the missing kids are back with their parents — and "some of them even started talking this year."

The implication seems to be that anyone placed under Gladys' control has been badly damaged, but there may still be some hope of recovery.

What's the deal with the rifle in the sky? And the title Weapons?

In one strange but memorable dream sequence in the movie, Archer sees a giant assault rifle hanging in the sky over his house while he's out looking for his son. It's a moment that reminds us the film is in fact called Weapons, even though the presence of literal weapons in the movie is somewhat limited.

So what's the deal with the title? Well, there are a couple of arguments you could make. The film's focus on the collective grief of a town after a class full of children is suddenly erased has an undeniable echo of the many school shootings plaguing America. There's also an exploration of the parental fear that comes with not being able to fully protect your children from the outside world, along with the terror of them being brainwashed into violence — a topic that was explored in grim detail in Netflix's brilliant miniseries Adolescence. This second point is emphasized by Archer saying the children seem like they've been "weaponized," and comparing them to heat-seeking missiles — a reminder that Gladys is quite literally using her victims as weapons.

Weapons is in theatres now.

Additional reporting by Belen Edwards.

Whether you're applying for a role that requires Excel expertise, want to add more strings to your data analysis bow, or simply want to do more with your data in Microsoft's spreadsheet program, add these five Excel skills to your toolkit to unlock the next level.

Apple's most recent iPad Air, launched earlier this year, packs some insane hardware, but its $600 price point might be too steep for some. Thankfully, $450 might be much more palatable.

Visual Studio Code, the popular development text editor and IDE, has rolled out its big July 2025 release. The v1.103 update adds new Git and terminal integrations, better language support, and more.

Seasoned social media users will remember when Snapchat launched a controversial feature that showed where your friends were at all times. Well, Instagram has something similar, and you may want to adjust some settings.

The biggest vacuum drop so far this year has just happened — over half a year after CES. Dreame unveiled a total of 16 new vacuums at a press event on Aug. 6, doubling iRobot's March drop of eight new Roombas.

Dreame's summer 2025 drop officially consists of the following: Seven robot vacuums (the Matrix10 Ultra, Aqua10 Roller, Aqua10 Ultra Roller, L50s Ultra, L40s Ultra, L40s Ultra AE, and L40s Ultra CE), four wet/dry vacuums (the H15 Pro CarpetFlex, H15 Pro Heat, H12 Pro FlexReach, and G10 Pro), and five stick vacuums (the R10 Pure, R20 Pure, Z30, Z30 Station, and Z30 Essential). Shipping will begin in September for some and October for others, with pricing info on the way. I'll update this guide with prices once they're announced.

SEE ALSO: As an anxious cat mom, I love my robot vacuum with a livestream camera

Several of the vacuums fall under the same "series" and have small differences, like docking stations. But amid all of these enigmatic names are some pretty cool features that could make several tricky aspects of floor cleaning more practical on a daily basis.

While I wait to get my hands on these new vacuums for at-home testing, let’s take a quick look at my four standouts from this dreamy drop:

Dreame Aqua 10 Roller and Aqua 10 Ultra Roller Credit: Timothy Beck Werth / Mashable Credit: Timothy Beck Werth / Mashable

The most unique addition to the Dreame robotic vacuum cleaner family is the Aqua10 Roller series. Here, Dreame ditched the spinning mopping pads (that just about every hybrid vacuum and their mother have nowadays) for a paint roller-esque roller mop that rinses itself as it cleans. This is meant to mitigate any possible spreading of a spill while the vacuum goes back and forth over the whole floor — exciting news for someone (me) who has watched a robot vacuum drag ranch dressing across the entire kitchen. This feels like an even more sanitary approach than the robotic mops that go back to the dock mid-cleaning session to rinse their mops, and should cut mopping time down significantly. To me, this is a much more practical fancy flex than the robotic arm on the Roborock Saros Z70.

SEE ALSO: Robot vacuum brands are losing the plot with the latest 'innovations' Dreame Matrix10 Ultra Credit: Timothy Beck Werth / Mashable

The Matrix10 Ultra is another new flagship in the lineup. Its claim to fame has to do with mopping pad maintenance: The Multi-Mop Switching Dock can store three sets of double spinning mopping pads of varying textures, which the robot vacuum can switch between depending on the floor it's about to clean. The two rotating mops can also be lifted and pressed down at different heights simultaneously, ensuring closer scrubbing of uneven flooring (like the thresholds in doorways where hardwood changes to tile or carpet).

Both of these new flagship robot vacuums clean with 30,000 Pa of suction power, beating the 20,000 and 22,000 Pa of my beloved Roborock Saros 10 and Saros 10R. And, speaking of thresholds between floor types, several of the new robot vacuums are outfitted with Dreame's ProLeap system: A set of retractable legs on either side of the vacuum that hoist it over obstacles a few inches tall (up to 3.15 inches with the Aqua10 Roller series). I first experienced ProLeap when testing the Dreame X50 Ultra that was announced at CES 2025, and to be fully transparent, the use case felt quite limited. Actual stair climbing is still out of the question, but it could be helpful if your floors have any height changes, or if you need a robot vacuum to cross sliding door tracks.

Dreame H15 Pro CarpetFlex Credit: Dreame

If you're strictly in the "I'll do it myself" camp when it comes to cleaning, the H15 Pro CarpetFlex is a... well, dream. It's Dreame's first hard floor and carpet cleaning combo vacuum, which adds heated wet roller mops to the traditional vacuum cleaning head for carpet stain liquefaction. Inside the vacuum head also lies the TangleCut system to prevent hair from wrapping around the roller. The H15 Pro CarpetFlex even washes and dries its wet cleaning system on the charging dock.

Dreame Z30 Credit: Dreame

While Dreame gives Roborock a run for its money in the premium robot vacuum department, it's coming for Dyson with these new cordless stick vacuums. The Dreame Z30, in particular, has strong potential to make the next cut of my best cordless vacuums list. Why? Because, on paper at least, it's more powerful than the most powerful Dyson — it shells out 310AW of suction power compared to the Dyson Gen5outsize's 250AW.

Dreame's new CelesTect Dust Reveal Technology also poses another inevitable Dyson comparison. Dreame describes the illumination on the Z30's cleaning head as LEDs on the multi-surface brush that "effectively detect and expose tiny dust particles in hard-to-reach corners or dimly lit areas." I'll be pumped if another vacuum brand finally gets on the level of Dyson's green dust-highlighting laser, but the mention of LEDs is just giving lightbulbs. Time (and testing in my apartment) will tell.

OpenAI’s best-ever large language model is here. GPT-5 is set to breathe new life into ChatGPT, making it faster, more helpful, and easier to use. Along with GPT-5, OpenAI launched a ton of new features for ChatGPT — more customization options, a better voice mode, a new user interface, and more. But all of those features pale in comparison to one specific improvement to ChatGPT — its vibe coding abilities.

Vibe coding is certainly the buzzword du jour. It's a type of software development that doesn't require any traditional coding skills. Rather, you simply describe the type of software, app, or tool you want to create and let AI do the rest.

In general, OpenAI promises that GPT-5 makes big improvements in coding. We'll let developers speak to how well GPT-5 handles hardcore coding assignments. But based on the GPT-5 livestream event and our own initial tests, GPT-5's coolest feature by far is its ability to make custom, interactive applications based on simple, natural-language prompts. It's a vibe coder's dream come true.

For education, for work, or just for fun, ChatGPT can now make fully realized apps and software in minutes — and it's all thanks to GPT-5.

GPT-5 is your own personal coder Credit: Screenshot: OpenAI

Vibe coding has become much more popular as AI services have improved. Vibe coding is essentially a term given to “coding” using natural language. You ask AI assistants to handle the actual code. If you need any fixes, you ask for changes in plain old English (or whatever language you speak). 

Vibe coding is very much in its infancy, but GPT-5 has the potential to change this. According to the company, GPT-5 is your own personal coder. Can’t find the perfect app for your specific needs? Or have an idea for a new website that you need help creating? GPT-5 can do it.

At the GPT-5 launch event, OpenAI leaders demoed this feature in real time. An OpenAI team member described a French language learning app to GPT-5, including a “Snake” like game that pronounces words in French and uses a mouse and cheese (in place of a snake). The prompt also requested a way to track progress, practice quizzes, and more. GPT-5 wrote hundreds of lines of code in a matter of minutes, and the end result was exactly what the prompt described.

Credit: Screenshot: OpenAI

OpenAI says that GPT-5 is better at “agentic coding.” That basically means you’ll be able to ask it to create something, and it’ll go off and do so for “many minutes” at a time – all in the background, and without you watching it work.

We tried vibe coding with GPT-5 ourselves

To put GPT-5's vibe coding abilities to the test, Mashable's tech editor asked it to create a visualization of the sinking of the Titanic, using this prompt:

Please create a custom visualization that shows the sinking of the Titanic, with a clock showing the timeline of events. The visualization should show inside the ship, including the waterproof compartments and the leaks that ultimately caused the ship to sink. You should be able to pause the visualization and click on the ship to reveal information about the timeline of events.

First, we tried this prompt with GPT-4o. The results weren't great. Its first attempt didn't even work; ChatGPT had to de-bug its own program. When it was finished, the visualizations were also extremely minimal, and it was barely interactive. However, GPT-5 made quick work of this assignment. (See the output for yourself at ChatGPT.)

GPT-5 created a realistic visualization, just like we asked, showing how water filled each compartment of the Titanic. The application takes you through the full timeline of events, and you can pause, speed up the visualization, and jump ahead to specific times of interest.

Left: Credit: OpenAI Right: Credit: OpenAI

Now, it's not perfect. With further prompt engineering, you could give ChatGPT clearer instructions on the types of visuals. And the graphic of the waterproof compartments blocks other elements in the app as the timeline progresses. However, this feels like a massive leap forward for vibe coding.

I don't expect GPT-5 to get everything right all of the time, especially when it comes to coding, and especially just after launch. When you’re vibe-coding, you’re trying to communicate what you want in natural language. And if you’ve ever been misunderstood by a friend or spouse, then you know that it’s not always easy to communicate ideas. You’ll likely have to tweak and refine any software that you code with GPT-5.

GPT-5 also likely won’t create code that’s completely bug-free, as our experience demonstrates. However, we can say that it's remarkably better than GPT-4o.

Natural language is the next coding language

Even if GPT-5 isn’t perfect, it’s clear that the way we create software is changing. I've often heard it said that the next big coding language is actually just natural language. Now, I'm starting to really believe it.

In the future, you won’t need to learn actual programming languages to be able to code — you’ll just need to be able to communicate your ideas.

We’ll also have to wait and see exactly how much better at programming GPT-5 really is, or if the improvements are more iterative. After all, Anthropic’s Claude is often considered to be better at coding than ChatGPT. And while the benchmarks reported in the GPT-5 system card put GPT-5 ahead of Claude, more independent testing will be needed to confirm this.

Disclosure: Ziff Davis, Mashable’s parent company, in April filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.

Despite dabbling with Minecraft back when it was still just a Java applet, I’ve never really seen the appeal of survival crafting games. Managing a hunger meter and gathering resources always felt like work, which I was always adamant was not the reason I played video games.

OpenAI has officially launched GPT-5, promising a faster and more capable AI model to power ChatGPT.

The AI company boasts state-of-the-art performance across math, coding, writing, and health advice. OpenAI proudly shared that GPT-5's hallucination rates have decreased compared to earlier models.

Specifically, GPT makes incorrect claims 9.6 percent of the time, compared to 12.9 percent for GPT-4o. And according to the GPT-5 system card, the new model’s hallucination rate is 26 percent lower than GPT-4o. In addition, GPT-5 had 44 percent fewer responses with “at least one major factual error.”

While that's definite progress, that also means roughly one in 10 responses from GPT-5 could contain hallucinations. That's concerning, especially since OpenAI touted healthcare as a promising use case for the new model.

SEE ALSO: How to try OpenAI's GPT-5 for yourself today How GPT-5 reduces hallucinations

Hallucinations are a pesky problem for AI researchers. Large language models (LLMs) are trained to generate the next probable word, guided by the massive amounts of data they're trained on. This means LLMs can sometimes confidently generate a sentence that is inaccurate or pure gibberish. One might assume that as models improve through factors like better data, training, and computing power, the hallucination rate decreases. But OpenAI's launch of its reasoning models o3 and o4-mini showed a troubling trend that couldn't be entirely explained even by its researchers: they hallucinated more than previous models, o1, GPT-4o, and GPT-4.5. Some researchers argue that hallucinations are an inherent feature of LLMs, instead of a bug that can resolved.

That said, GPT-5 hallucinates less than previous models according to its system card. OpenAI evaluated GPT-5 and a version of GPT-5 with additional reasoning power, called GPT-5-thinking against its reasoning model o3 and more traditional model GPT-4o. A significant part of evaluating hallucination rates is giving models access to the web. Generally speaking, models are more accurate when they're able to source their answers from accurate data online as opposed to relying solely on its training data (more on that below). Here are the hallucination rates when the models are given web-browsing access:

• GPT-5: 9.6 percent

• GPT-5-thinking: 4.5 percent

• o3: 12.7 percent

• GPT-4o: 12.9 percent

In the system card, OpenAI also evaluated various versions of GPT-5 with more open-ended and complex prompts. Here, GPT-5 with reasoning power hallucinated significantly less than previous reasoning model o3 and o4-mini. Reasoning models are said be more accurate and less hallucinatory because they apply more computing power to solving a question, which is why o3 and o4-mini's hallucination rates were somewhat baffling.

Overall, GPT-5 does pretty well when it's connected to the web. But the results from another evaluation tell a different story. OpenAI tested GPT-5 on its in-house benchmark, Simple QA. This test is a collection of "fact-seeking questions with short answers that measures model accuracy for attempted answers," per the system card's description. For this evaluation, GPT-5 didn't have web access, and it shows. In this test, the hallucination rates were way higher.

• GPT-5 main: 47 percent

• GPT-5-thinking: 40 percent

• o3: 46 percent

• GPT-4o: 52 percent

GPT-5 with thinking was marginally better than o3, while the normal GPT-5 hallucinated one percent higher than o3 and a few percentage points below GPT-4o. To be fair, hallucination rates with the Simple QA evaluation are high across all models. But that's not a great consolation. Users without web search will encounter much higher risks of hallucination and inaccuracies. So if you're using ChatGPT for something really important, make sure it's searching the web. Or you could just search the web yourself.

It didn't take long for users to find GPT-5 hallucinations

But despite reported overall lower rates of inaccuracies, one of the demos revealed an embarrassing blunder. Beth Barnes, founder and CEO of AI research nonprofit METR, spotted an inaccuracy in the demo of GPT-5 explaining how planes work. GPT-5 cited a common misconception related to the Bernoulli Effect, Barnes said, which explains how air flows around airplane wings. Without getting into the technicalities of aerodynamics, GPT-5's interpretation is wrong.

This Tweet is currently unavailable. It might be loading or has been removed.

Dreame is determined not to let Dyson and Shark be the only vacuum companies dominating the hair care space.

During a press event on Aug. 6, the brand unveiled the three latest additions to its already sizable beauty tech lineup: the Dreame Miracle Pro, the Dreame Dazzle, and the Dreame Pocket Ultra. Between the three releases, Dreame targets the same markets as the Dyson Supersonic Nural, the Shark FlexStyle, and the Dyson Airwrap, respectively. So far, there are no available release dates, but the Dreame Miracle Pro has a listing on the brand's website, with a "coming soon" tag. I've reached out to Dreame about prices and release dates, and will update this post once I hear back.

SEE ALSO: I tested the best Dyson Airwrap dupes under $300: The Shark FlexStyle is no longer your only option

These drops follow the recent release of the Dreame AirStyle Pro, which looks even more remarkably similar to the Dyson Airwrap, and comes with six styling attachments (seven if you count the different direction curling barrels). It retails for $399.99.

In my ongoing journey to test out the best Dyson dupes on the market, I've tested (and been impressed by) the Dreame Gleam and Dreame Pocket hair dryers. I haven't been able to get my hands on these newest launches just yet, but below, you'll find details on what each tool aims to bring to the beauty tech table.

Dreame Pocket Ultra release info The Dreame Pocket Ultra sticks to the same design as the regular Pocket, plus a few new attachments. Credit: Timothy Beck Werth / Mashable

The Dreame Pocket Ultra builds on the Dreame Pocket, a foldable hair dryer that can also be used like the Airwrap. I found the original Pocket to be a great option for anyone who wants both the Supersonic and Airwrap experience. This tiny tool provided a bit of both, with one curling barrel and one "unfrizz" nozzle, which grabbed hair to provide a smoothing attack. The result was a versatile tool in surprisingly compact packaging.

The Pocket Ultra ups the ante with three new attachments: another curl barrel (for the opposite direction), a diffuser, and a styling concentrator. While the new attachments will add to the storage space required, they also add a good degree of flexibility to the styling experience, especially for folks with textured hair. Beyond the new attachments, the Pocket Ultra looks identical to the Pocket, offering the same base design, as well as five styling settings and two air speeds.

Dreame Miracle Pro release info The Dreame Miracle Pro comes with a digital display like the T3 Aire IQ. Credit: Timothy Beck Werth / Mashable Credit: Timothy Beck Werth / Mashable

The past couple of years have brought the release of smart hair dryers, from the Dyson Supersonic Nural to T3's Aire IQ. The Dreame Miracle Pro combines the tech of these two dryers: it has the same digital display as the T3 Aire IQ, plus the scalp detect mode of the Supersonic Nural, which automatically adjusts the dryer's speed and heat based on how far it is from your scalp. It also features a red light therapy component, similar to the (seemingly discontinued) Zuvi Halo and (still available) L'Oreal AirLight Pro.

While the Zuvi and L'Oreal dryers' infrared light was aimed at drying hair with less heat, Dreame's press materials for the Miracle Pro call their tech "red light scalp therapy," suggesting that the light is for promoting hair growth — a point I've asked Dreame to clarify. Dreame takes the extra features of this dryer a step further than existing premium options, adding in an "essence mode" that releases a hair mist during styling. The dryer also comes with four attachments — a styling nozzle, smoothing nozzle, unfrizz nozzle, and diffuser — and a bottle of hair care essence.

Dreame Dazzle release info The Dreame Dazzle can be used vertically or folded in an L-shape (pictured here). Credit: Timothy Beck Werth / Mashable

A lot of brands dupe the Dyson Airwrap, but the less common move is to dupe the Shark FlexStyle. Maybe it seems redundant to make a dupe of a dupe, but the FlexStyle does have a twisting base, which allows you to use the multi-styler vertically, or in an L-shape like a traditional dryer.

The Dreame Dazzle matches this design choice, offering five attachments, including a diffuser, directional curling barrels, a round brush, a styling concentrator, and what looks like a gentle air attachment — I've reached out to Dreame to confirm. Dreame's press release also states that the Dazzle comes with smart nozzle recognition and a digital display, both of which cannot be found on the Shark FlexStyle.

ChatGPT is about to get a whole lot better. OpenAI has finally taken the wraps off of GPT-5, its latest and greatest large language model. GPT-5 has been rumored and speculated about for as long as GPT-4 has been around, and for good reason. The new foundational OpenAI model packs a ton of improvements that will make ChatGPT more accurate, more customizable, and more helpful.

For starters, GPT-5 is smarter, faster, and more accurate — which, yes, means fewer “hallucinations,” which occur when ChatGPT makes up information. As one OpenAI leader described it during the announcement livestream, it's like having a subject matter expert on anything in your pocket. GPT-5 is also being positioned as a model that can do everything. It can write, code, create interactive apps, check your email, and conduct deep research.

Curious about what's on offer in GPT-5 and whether or not it will actually improve the ChatGPT experience? Like the rest of the world, we're still investigating GPT-5 for ourselves. In the meantime, here are all the best new features to try with GPT-5, which rolls out today for free to all ChatGPT users.

The integrated ChatGPT experience

GPT-5 integrates all of the various models on offer by ChatGPT. That means instead of picking the right model for the task at hand, ChatGPT will pick the best model(s) for your prompt.

Simply ask GPT-5, and it will decide how to approach a problem on its own. 

Credit: OpenAI

OpenAI is introducing additional models along with GPT-5, which is the do-it-all model that will be OpenAI’s flagship model going forward. There’s also GPT-5 Mini and GPT-5 Nano, which are more cost-effective than GPT-5, and will be accessible to developers who want to pay less. Users will automatically transition to using these smaller models when they hit their GPT-5 usage limits; however, OpenAI says they’re still very capable models — and even more capable than GPT-o3 in many cases.

Your own personal coder

OpenAI is positioning GPT-5 as being your own personal coder, and the company says GPT-5 is significantly better at writing code. If you're a programmer, then OpenAI wants GPT-5 to be your new coding partner. But what about the rest of us? The so-called vibe coders of the world?

During a live demo, company leaders showed how to use GPT-5 to create custom, interactive applications on demand using a simple prompt. So, you could think of an app or a piece of software that you wish you had, and simply ask ChatGPT to make it.

Of course, in the real world GPT-5 probably won’t be quite as seamless as that — it’s still likely to make errors, and won’t get everything right the first time. But if you’re interested in so-called “vibe-coding,” GPT-5 looks like a big leap forward.

Try deep research and reasoning

Because GPT-5 integrates all of OpenAI's models into one AI chatbot — and because it's free to all ChatGPT users — free users can finally gain access to OpenAI's most advanced reasoning models. Reasoning models can conduct multi-part searches based on your prompt, then synthesize vast amounts of data into a response. And according to OpenAI CEO Sam Altman, GPT-5 can do this at a PhD level.

We haven't tested this out for ourselves yet, and we're always nervous about hallucinations. However, if you only use the free ChatGPT tier, then this kind of deep research is definitely worth trying. But GPT-5 doesn't just do research. During the livestream, OpenAI leaders showed how GPT-5 could create custom visualizations, interactive apps, games, quizzes, and other tools to help you learn about the subject at hand.

Personalities

OpenAI is also debuting four new preset personalities in ChatGPT. These can inform how ChatGPT will interact with you as you chat and are set on a chat-by-chat basis. Initially, there will be four options:

• Cynic

• Robot

• Listener

• Nerd

You don't actually have to use any of these personalities if you don't want to, but you can use them for specific use cases as needed. You can also adjust them in ChatGPT settings. OpenAI says the new personalities were specifically designed to reduce the sycophancy issue with GPT-4o.

Better voice mode

Voice Mode makes it easy to get information from ChatGPT without having to type every single word. With GPT-5, Voice Mode is getting better, OpenAI says. The company says Voice Mode will now adapt and better understand your instructions so that it can tailor how it speaks depending on the moment. All users will have access to this improved ChatGPT Voice Mode, including free users, though of course paid users will have higher usage limits.

OpenAI is also making Voice Mode more accessible, and it will now be available in custom GPTs. The lack of advanced voice mode in custom GPTs has been sorely missing, so this update will be super helpful for those who use custom GPTs for different tasks.

Connect your Gmail and Google Calendar

The ability to connect ChatGPT to external apps like Google Drive has been helpful for those who want their AI assistant to be able to reference information in their Google account. Now, those connections are getting much more helpful.

You'll be able to enable Gmail, Google Calendar, and Google Contacts connections with ChatGPT, and going forward, ChatGPT will automatically pull information from them and reference them as necessary. That means that you won't have to explicitly select the connections in a chat each time. For some, of course, this might raise privacy concerns, but rest assured, you don't have to connect these apps or services if you don't want to.

At launch, you might not be able to connect them anyway. Initially, this feature will only be available to Pro users and will start rolling out next week. Other tiers will get it too, but we don't know exactly when.

Customize your chats

Apart from simply being underpinned by a better model, GPT-5 also heralds a few new features for ChatGPT that can even further enhance the experience of using it. An example of that is the fact that you can now choose colors for different chats. That may sound like a small update (and to be clear, it is), but this feature could help you much better organize your chats and more easily jump between them.

What features did we not get?

OpenAI says GPT-5 can write better, code better, and even better inform users on issues around health. (To be clear, you should still consult with a medical professional before you make any health-related decisions. ChatGPT does not replace your doctor.)

But there are some ways in which GPT-5 is not better than existing OpenAI large language models. For example, it has a wider context window than GPT-4o at 256,000 tokens, but that's still nowhere near as impressive as GPT-4.1, which has a one-million-token context window. Thankfully, OpenAI says it will continue to support GPT-4.1, which might be a better model for those times in which you need ChatGPT to process massive amounts of information at a time. 

Disclosure: Ziff Davis, Mashable’s parent company, in April filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.

The Samsung Galaxy Buds 3 are good, but they have a big flaw. They use Bixby out of the box as their assistant (some people like it, even), and while you can change it to Google Assistant, it's not that huge of an upgrade—and Google Assistant is getting phased out anyway. Now, it just got an upgrade to bring it to current times.

Do you want to try Arch Linux, but feel scared because of the recent malware incident? Are you a current Arch user wondering how to keep your system safe and secure? Well, I have been using the Linux distro for five years now, and here’s my complete guide on staying safe on Arch Linux.