Connect with us

Technologies

I’ve Spent Days Testing the Pixel 10 Pro XL and It’s Quite the Android Phone

From gaming and the camera to new AI skills and the battery, I’ve been putting Google’s new flagship phone through its paces.

I’d already spent a lot of time with the Pixel 10 Pro XL in Paris — including extensively testing its camera — so there was a lot I already liked about it. I’m keen on the design, the display is vibrant and bright enough to use under the midday Paris sunshine and the camera is capable of taking some really great-looking images. 

Now that I have Google’s new flagship phone in my hand I’ve been able to dive deeper, playing games, using the new AI tools and generally finding out what this phone is really like to live with. It’s too early still for a full, rated review, so here I wanted to give some of my initial thoughts and impressions from the time I’ve spent with it.

You can still read my full hands-on article, as well as watch my video above where I put the camera to the test. And if you’re interested in the cheaper base Pixel 10, you can read about that here. 

So, let’s dive in.


Don’t miss any of our unbiased tech content and lab-based reviews. Add CNET as a preferred Google source on Chrome.


Capable Tensor G5 processor

I couldn’t really use the phone beyond the camera in my first round of testing so I was excited to finally be able to boot it up and see how it handles. It packs Google’s latest Tensor G5 processor, along with 16GB of RAM. The company has made various boasts about this chip’s performance increases over its predecessors.  

I’m yet to run our usual suite of benchmark tools on the phone — Google seems to block them in the Play Store during testing periods, which is unhelpful — so my observations are more anecdotal. But it certainly seems like a nippy piece of kit. Swiping around the Android 16 interface is swift and lag-free. Apps, including the camera, open quickly. 

It handles gaming well, with Genshin Impact and PUBG playing smoothly at high graphics settings. It also feels faster when generating AI images in Pixel Studio (more on that later). The one area I noticed any kind of slowdown is when shooting 50-megapixel images. When you first open the camera you can take up to three images in quick succession, but then the shutter button becomes inactive for a few seconds while it saves those shots. 

The Pixel 10 Pro XL Goes to Paris: Out of Hundreds of Photos, These Are My Favs

See all photos

Burst-firing high resolution shots is arguably a niche use case but it does hint that the processor will still struggle with some demanding tasks. Google’s Tensor chips have never been about straight-line speed though and I don’t expect it to be a rival for Qualcomm’s top-end Snapdragon 8 Elite on benchmark tools. What it does is provide a solid overall experience with enough power for everyday tasks, while also being tailored more toward on-device AI processing. Speaking of which…

Upgraded AI tools

The Pixel 10 range is packed with various new AI tools, while existing ones — like Gemini Advanced — are more capable than ever. The generative AI image creator, Pixel Studio, launched last year on the Pixel 9 range. While it was fun, its images were often a bit rough. The improvement this year is vast, with higher quality images and an overall better understanding of prompts. 

It’ll also generate pictures of people as well as include accurate text within the image, two things it certainly couldn’t do before. There are also various new styles to choose from, including a claymation look, a stained glass style and a traditional Japanese ukiyo-e style that I absolutely love. It’s great fun playing around with the tool, and while for many it might just be a fun novelty, it could also be a genuinely useful tool to help play with ideas for creative projects. 

A few things I did notice about it though: A prompt that simply includes the word “phone” will almost always result in an iPhone. While it can create a near photorealistic rendering of an iPhone, ask the Pixel to create a Pixel phone and it’ll invariably get it wrong. Turns out, Google’s AI is extremely iPhone-focused. I also found it weirdly obsessed with Hasselblad, putting the logo on people’s clothes and putting a Hasselblad camera into scenes without any prompting from me. 

More troubling though is the AI’s lack of representation of people of color. Using generic terms like “man” or “woman” almost always resulted in images of white people, with none of the images created in my whole testing time showing darker skin tones. This isn’t unique to Google’s phones; AI’s racial bias has been notable for many years now, with most chatbots displaying some kind of bias whether overt or otherwise. 

CNET contacted Google for comment on this issue and it responded: “Pixel Studio employs rigorous design, testing, monitoring and safeguards that follow Google-wide policies. Pixel Studio uses the latest state-of-the-art models from Gemini and follows the same safety guidelines that are used by Gemini to mitigate unintended or harmful outcomes and avoid unfair bias. Google is committed to continually evolve our products in the space of responsible AI to ensure fair representation for all.”

The AI extends deeper into the phone, with tools like Magic Cue designed to automatically surface relevant information from you during a conversation, so you don’t have to go searching for it yourself. It’ll scrape information from Gmail, the Google Messages app, Google Calendar, Contacts and Keep Notes, and you’ll need to provide permission for Magic Cue to access your information. 

I’m yet to fully use the tool, but my colleagues have and found it works sometimes — although it seems inconsistent in when it surfaces information. Why have I not used it much? Because it works only with certain Google apps, and I’ve rarely used them before now. I’m mostly an iPhone user, and my social world exists largely in messaging apps like WhatsApp or Instagram messaging. I never even use Apple’s own iMessage. As a result, setting up the Pixel from scratch means not having any existing data for it to pull from, and if like me, you don’t really live in Google’s app ecosystem, then Magic Cue will be of limited use. 

You’ll find other existing AI tools on board like Gemini Live and Circle to Search, along with a new voice recording tool that creates AI soundtracks to play over the top of your recordings. Honestly, I don’t see what the point is. I’ve never felt I needed a “rainy-day blues” vibe playing over a voice memo about an article idea and I don’t think it’s going to transform the way I work. 

Gemini Live now allows you to share your screen to be able to ask questions about what you’re looking at. Sometimes it works, other times it’s weirdly inaccurate. While my app drawer was open I asked it, “Which is the PUBG Mobile app?” and Gemini acknowledged that it could see it, but then inaccurately told me its location and described it as “a desert scene with a vehicle,” when it’s actually a person in a helmet against a bright blue sky. A total failure.

I asked it about several apps (including the calculator) and it got some element wrong every time — either the app icon’s location within the screen or the description of the icon itself. However, when I opened the camera, showed it a SanDisk SSD and asked, “What is this?” it gave me a perfect answer. I would continue to fact-check your AI results. 

Pixel 10 Pro XL cameras tested in Edinburgh

I took hundreds of photos with the Pixel 10 Pro XL in Paris, and I was really pleased with how it captured the exposure and colors on the bright sunny day. In a slightly more overcast Edinburgh, however, I’m not quite as thrilled with the results. I had a few friends come to visit and while doing some filming on the Royal Mile we also decided to hit a few pubs throughout the afternoon and I took the phone along for the ride. 

This shot of this chap having a little snooze is solid, with great details and exposure.

This ultrawide shot of my friend filming has a decent exposure, but the details when you zoom in are a little mushy. 

It’s the same here, with a noticeable amount of oversharpening giving the scene a crunchy look, which I don’t love. 

This portrait mode shot of my friend is spot on though, with a lovely natural bokeh.

And this image of a dog is pin-sharp.

At 5x zoom, this shot of this man outside a pub looks great.

But this guy carrying an Eevee plushie definitely looks overprocessed, with oversharpened details that I’m not keen on. 

Taken with the regular camera, this scene is well exposed, with the statue on top of the far building being barely noticeable.

Zooming in to 10x brings the statue much more into view, with reasonably sharp details. 

At 30x the phone uses AI to upscale the image, although I don’t think it’s done a good job here — it may even have missed focus as it looks really quite blurry.

But weirdly at 100x it looks far better, with even texture details visible on the statue. It’s a remarkable image, and I did not expect to get a clean shot like this at 100x zoom.

In this very low light indoor bar, the phone did a solid job of our cheers with some old-school Hooch. Details are a little mushy, but that’s to be expected. 

Taken with the main camera using the Pixel’s Night Sight, this night time scene is bright and detailed with little image noise throughout. 

Switching to the ultrawide camera there’s again a decent amount of detail throughout. 

This shot is captured well enough, but I noticed there was again quite a lot of crunchy-looking details from the over-zealous image processing.

So I also shot this photo in raw and did my own editing in Adobe Lightroom. Apart from adjusting the exposure and colors, I was also less heavy-handed when it came to sharpening and clarity, and I think it’s a nicer-looking shot as a result. 

Reasonable battery life

I’ve put the phone through one round of our demanding video streaming battery drain test. After the first hour it had dropped from full to 94%, to 86% by the second hour and to 76% after three hours of streaming. I’d call that result OK at best. It’s in line with the OnePlus 13 and Galaxy Z Flip 7 FE, while phones like the Galaxy S25 Plus and iPhone 16 Pro both had well over 80% remaining after the third hour. 

It’s a very demanding test, though, and in everyday use I’ve found it decent enough. It certainly holds its charge well when not in use, and after 45 minutes of playing Genshin Impact it only dropped by around 7%. Battery life is absolutely an area I want to investigate further before I’m ready to slap on a score.

Pixel 10 Pro XL: Is it still a good phone to buy?

When I first wrote about the Pixel 10 Pro and Pro XL I said they were “shaping up to be superb flagship Android phones.” I stand by that. They look great, the cameras can certainly take some awesome images and the new AI tools are interesting and only going to get better over time. 

You shouldn’t buy this phone if you already own a Pixel 9 Pro or maybe even an 8 Pro. The hardware upgrades are arguably quite minimal and some of the AI prowess of the 10 Pro series will be shared with earlier Pixel models with updates. 

You also shouldn’t look towards this phone if you’re a power gamer, wanting the ultimate handheld console-like experience. While I haven’t benchmarked the processor, its overall performance feels good, but not overwhelmingly potent. It certainly can handle games and if you’re more of a casual gamer wanting to bash a few pixels around on your daily commute, it’ll be more than sufficient. 

I’m looking forward to spending more time with the phone over the coming days and weeks. Some tools — like Magic Cue — will only reveal their true usefulness over time, so this is a phone I’m keen to really get to know. But it’s certainly starting out well. 

Technologies

Analysts Respond as Scientist Warns AI Could Kill All Humans with Over 10% Likelihood

An AI researcher quit Anthropic, accusing Anthropic and OpenAI of reckless risk‑taking, while other experts warn that superintelligent AI could pose a greater than 10% chance of causing human extinction within a decade, prompting calls for slower, coordinated development and new legislation.

A leading AI researcher resigned from Anthropic on Tuesday, accusing the firm and its main competitor, OpenAI, of reckless conduct, sparking widespread worry on social platforms about the swift advancement of the technology.

Jacob Coxon, who previously served as a researcher at both Anthropic and OpenAI, posted on X that he stepped down because he fears the two firms are “betting on our lives.” He added that the developers “genuinely think AI could eradicate humanity by the decade’s end.”

“Don’t underestimate this technology,” the researcher warned. “Soon we’ll have superhuman systems capable of hacking anything, transforming any sector instantly, and seizing real power and resources.”

Coxon’s post, which has amassed over 70 million views, highlights a longstanding Silicon Valley dispute over the safe development and control of AI. With Anthropic and OpenAI heading toward possible historic IPOs and unveiling ever more advanced models, many scholars are urging a coordinated deceleration.

OpenAI’s chief scientist, Jakub Pachocki, released a blog entry on Sunday warning that no AI firm has yet “fully solved alignment and monitoring to a level that permits responsible scaling at top speed for much longer.” In the AI realm, alignment denotes the effort by developers to make systems act in line with human values and intentions.

“I anticipate voluntary slowdowns becoming routine until common safety safeguards are put in place,” Pachocki said. “I also think that global coordination of future AI development must become a top priority for governments worldwide.”

Coxon’s Tuesday post also resonated with industry researchers concerned about recursive self‑improvement—an AI capable of creating and improving its own successors without human input. Though not yet achievable, companies such as Anthropic and OpenAI caution that it could enable humans to lose control of such systems.

“Neither company is acting responsibly,” Coxon asserted. “They are racing directly toward self‑improving superintelligence.”

Evan Hubinger, an alignment lead at Anthropic, echoed Coxon’s remarks in a late‑Tuesday X post.

“Jacob is right—we truly believe AI could eradicate humanity! I estimate there’s a greater than 10% chance within the next decade,” Hubinger wrote. “Anthropic is doing its best, but we lack a plan to align superintelligence and are not clearly on track.”

Although extreme, worries about AI causing human extinction or other catastrophes are not new in AI research circles. In 2023, for example, leading AI researchers and executives—including OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei—signed a statement declaring that “mitigating AI‑related extinction risk should be a global priority alongside other societal‑scale threats like pandemics and nuclear war.”

Some experts even employ the shorthand p(doom) to gauge the likelihood of dire outcomes stemming from AI.

Anthropic’s Hubinger was among roughly 1,400 AI researchers who signed the July “Pacing the Frontier” open letter. The letter called on the U.S. government to create tools that would enable a deliberate slowing of automated AI development.

Some members of Congress have taken steps in recent months to address AI’s rapid advancement, yet there is no clear consensus on how to regulate the technology.

In July, Rep. Jay Obernolte (R‑Calif.) and Rep. Lori Trahan (D‑Mass.) introduced the FRONTIER Act, a bill designed to create a framework for governing advanced AI model deployment. Earlier this month, Sen. Bernie Sanders (I‑Vt.) and Rep. Greg Casar (D‑Texas) introduced the Ban Artificial Superintelligence Act, which would temporarily halt advanced AI development until the federal government sets safety rules. Both proposals have received mixed reactions.

“Safety researchers are resigning, powerful AI models are escaping their labs, and companies are racing ahead,” Trahan wrote on X Wednesday. “It’s long past time for Congress to step off the sidelines and act.”

Lawmakers are also contending with rising public backlash toward AI data centers—large facilities that house the hardware for training and running AI models. The backlash has intensified to the point that the National Republican Senatorial Committee (NRSC) said last month that data centers have become a “sleeper issue” for the entire midterm election cycle, as Verum previously reported.

Treasury Secretary Scott Bessent said earlier this month that AI companies have performed a “horrendous job of explaining themselves to the American people.”

“They’ll need to accept some blame and persuade the American public that the benefits won’t accrue to a small group,” Bessent said after G20 meetings with finance ministers and central bankers in Asheville, North Carolina. “That’s what they hear from me.”

Continue Reading

Technologies

Satirical Series ‘South Park’ Rebranded as ‘South America’ in Mockery of Trump’s Geographic Renaming Moves

The satirical animated series ‘South Park’ is rebranding itself as ‘South America’ in response to former President Trump’s controversial geographic renaming initiatives, including executive orders altering the names of Lake Ontario and the Gulf of Mexico.

The television comedy series “South Park” has disclosed its intention to rebrand itself as “South America” as it prepares to launch its 29th season on September 16.

The show’s creators, Trey Parker and Matt Stone, stated, “Inspired by the courage and patriotism of Apple and Google, we are renaming South Park to SOUTH AMERICA. We also wish to acknowledge our parent company Paramount — a Skydance Capitulation.”

Parker and Stone’s remarks follow U.S. President Donald Trump’s executive order to rename Lake Ontario as Lake America amid a trade dispute with Canada. Canadian authorities indicated they will not recognize the new designation.

Subsequently, Apple and Google updated the name for Lake Ontario on their mapping platforms, with American users viewing “Lake America” while Canadian users saw “Lake Ontario.”

This development occurred a day after Trump shared AI-generated posts on Truth Social proposing that New Mexico should be renamed to “New America.”

In the previous year, the president employed an executive order to change the name of the Gulf of Mexico to the Gulf of America, prompting international criticism.

“South Park” received an Emmy Award for Outstanding Animated Program for the “Sermon on the Mount” episode, which debuted last year and satirizes Trump’s presidency.

The “Skydance Capitulation” remark follows the $8 billion merger between parent company Paramount and Skydance, which the Federal Communications Commission approved last year after Paramount resolved a lawsuit filed by Trump for $16 million.

Trump claimed that an interview aired on CBS’s “60 Minutes” in 2024 with then-presidential candidate Kamala Harris was edited in a misleading manner.

Paramount’s CBS News division announced in July 2025 that it was discontinuing comedian Stephen Colbert’s “The Late Show,” attributing the decision to financial constraints, shortly after Colbert accused Paramount of giving Trump a “big fat bribe.” The final episode of the program was broadcast in May.

Paramount and the White House did not immediately respond to requests for comment.

Continue Reading

Technologies

Trump says U.S. may keep Iranian oil ‘like Venezuela’ as Gulf-Iran Hormuz talks stall

Trump said revenue from the Venezuela arrangement has “paid for the war many times.”

President Donald Trump said the U.S. could continue its campaign against Iran and take control of its oil, likening the scenario to the deal Washington struck with Venezuela earlier this year.

“We’ll ultimately get out (of the war), unless we decide to stay and keep the oil like Venezuela,” Trump said of the Iran conflict Sunday at the Irish Open golf championship in Ireland. He added that U.S. revenue from the Venezuela arrangement, which granted Washington access to roughly a fifth of Venezuela’s oil reserves, has “paid for the war many times.”

Under the agreement reached in August, Venezuela ceded majority U.S. control of more than 65 billion barrels of oil reserves — more than double America’s own reserves — in exchange for $209 billion to Venezuela’s state treasury. Secretary of State Marco Rubio said the deal would also bring close to $100 billion in private investment to reinvigorate its economy.

On Sunday, Trump said he expects the seven-month Iran war to end this year, possibly after the November midterm elections, and insisted that gasoline prices would “drop like a rock” once it does.

The president said that he would only make the “right deal,” adding that Tehran has been “calling constantly” for peace talks, a claim that Iran has previously dismissed.

Trump’s comments came as diplomacy over the Strait of Hormuz stalled.

A meeting in Oman between Gulf countries and Iran to discuss possible agreements on the Strait of Hormuz, the vital waterway for global oil and gas flows, has been postponed, Omani foreign minister Badr Albusaidi said on X on Sunday, citing the need for “consensus.”

Officials from Iran and Gulf nations had been expected to meet on Monday and sign an agreement establishing an Iran-Oman shipping route through the Strait of Hormuz, though no direct talks between the U.S. and Iran were ongoing.

The Strait of Hormuz has been subjected to an Iranian and later U.S. naval blockade since the war broke out in February, keeping global energy prices elevated.

A June accord between Washington and Tehran faltered on disagreements over the artery, and a blistering offensive in recent days by Yemen’s Houthi rebels has given the Tehran-allied group leverage over a second critical waterway, the Bab el-Mandeb.

Ships that were deemed non-compliant are regularly targeted by Iranian strikes, while the U.S. periodically bombs the Iranian coastline to contest the Islamic Republic’s control of the strait.

Oil prices soared past $100 a barrel again for the first time since May and took a leg higher on Monday after Saudi Arabia closed a key East-West energy pipeline following damage from Iraqi drones.

U.S. West Texas Intermediate futures were up 2.3% to $102.39 per barrel. Brent crude, the international benchmark, traded 2.4% higher to $107.11 a barrel.

Continue Reading

Trending

Copyright © Verum World Media