There’s nothing I love more than testing a tech company’s claims. I used to host a CNET YouTube series called Breaking Point, where I got to do what most people try to avoid: torture their tech. That meant purposely dropping thousands of dollars’ worth of phones on the ground, dunking them in water and pushing them to their limits. 

So when Apple released its first water-resistant iPhone in 2016, I thought, “Game on!”

It’s hard to imagine now, but before the iPhone 7, a splash of water could make your Apple phone useless. Its IP67 rating means it can withstand 30 minutes in 1 meter of water (about 3 feet). To test it, I took the iPhone 7 for a “swim” in the pool at my old San Francisco apartment building, along with then CNET senior editor Sean Hollister, and filmed the whole thing.

We ended up testing it beyond its limits and didn’t exactly follow all the care instructions. Spoiler alert: The iPhone mostly survived, but it did come away with a battle scar or two.

At the time, I was blissfully unaware that just because a phone is fine after being underwater doesn’t mean it escaped damage. Chlorinated pool water and ocean saltwater can slowly corrode the materials inside a phone, even after it’s dried out. That wouldn’t kill a phone right away, but it would shorten its lifespan. Water damage (intentional or not) will void Apple’s warranty. 

Looking back at this 10-year-old video of Hollister and me throwing an iPhone 7 into a pool is ironic. Instead of taking my phone on more water-related adventures, safe in the knowledge that my tech should be fine, I’m careful with my iPhone around water. Less fun, yes, but more informed.

Watch my full Did the iPhone 7 Survive Our Water Test? video down below.

Vanessa Hand Orellana

Principal Writer

Advertisement

Vanessa is a lead writer at CNET, reviewing and writing about the latest smartwatches and fitness trackers. She joined the brand first as an on-camera reporter for CNET’s Spanish-language site, then moved on to the English side to host and produce some of CNET’s videos and YouTube series. When she’s not testing out the latest health tech, you can catch her on a hike or trail run with her family.

See full bio

Samsung’s software fix for the Galaxy S26 Ultra‘s red tint issue started rolling out last week, but it isn’t working for all affected owners and is reportedly causing new display quirks for some.

Some Galaxy S26 Ultra owners are still seeing red

Samsung originally announced the fix earlier this month, saying the red tint came from a calibration issue and not a hardware defect. At the time, the company urged users to visit a local service center for a display calibration or wait for an over-the-air update. That update landed last week, but it doesn’t seem to resolve the issue for all users.

According to Newsway, some owners say the red tint persists following the update, while others have reported a tint along the edges. A few users who took their phones in for calibration at a service center have highlighted different problems, including colored spots near the edges and reddening at the top and bottom.

What Samsung is telling affected users

Samsung’s advice for anyone still seeing the red tint is to turn off display features like Adaptive color tone, Color correction, Color filter, Extra dim, and Eye comfort shield to check whether that clears it. If the tint persists, the company has directed users to visit a service center for further evaluation and guidance.

The situation has left many early adopters in a difficult position. When complaints first surfaced, Samsung reassured customers that a software update would resolve the issue without requiring panel replacements. However, recent user reports suggest that software adjustments alone may not be enough for every affected device.

Samsung has not officially responded to the new batch of complaints that surfaced following the update, and the company has yet to confirm how many units are affected overall. For now, owners who are still dealing with the red tint after installing the patch have little recourse beyond visiting an authorized service center or holding out for a follow-up patch.

The Hubble telescope has an almost legendary status. Over the last three decades, Hubble gave us some of the most spectacular close-ups of the universe ever captured, and now, NASA’s newest space telescope wants to see the bigger picture. The Nancy Grace Roman Space Telescope successfully launched aboard a SpaceX Falcon Heavy from Kennedy Space Center in Florida at 7:26 a.m. EDT on August 30. Roman is now beginning a roughly three-month journey toward the Sun-Earth L2 point, nearly a million miles away, where it will spend its primary five-year mission surveying huge portions of the cosmos.

Roman carries a 2.4-meter primary mirror, the same diameter as Hubble’s, but its Wide Field Instrument can capture a field of view at least 100 times larger than Hubble’s. NASA expects it to use that combination of sharpness and scale to map billions of galaxies, study dark matter and dark energy, and discover enormous numbers of previously unknown worlds.

Imagine Hubble with panoramic vision

Roman won’t be replacing the Hubble or the James Webb Space Telescope. Webb is extraordinarily good at staring deeply into relatively small sections of the cosmos, while Roman will work like a survey machine. It can sweep across enormous areas while retaining the resolution needed to pick out distant objects.

NASA has previously illustrated the difference rather dramatically. In one simulated region containing millions of galaxies, Hubble would need roughly 85 years to map the area at the same depth. Roman could do it in about 63 days. This speed will help Roman construct an enormous 3D picture of the universe and track how galaxies have been distributed across cosmic history. Scientists can then use those patterns to investigate some of the universe’s biggest mysteries, like dark matter and dark energy.

According to the Reuters report, the roughly $4 billion observatory could study more than two billion galaxies over its mission. Aside from this, Roman will also be hunting exoplanets. NASA expects its surveys to uncover around 100,000 exoplanets, including worlds found through gravitational microlensing, where the gravity of an object passing in front of a distant star temporarily magnifies its light. Its experimental Coronagraph Instrument will also test technology intended to block overwhelming starlight so astronomers can directly study planets and planet-forming disks nearby.

From government optics to the dark universe

Key optics in the telescope, including its primary mirror, were originally made available to NASA by the National Reconnaissance Office. L3Harris subsequently reshaped the mirror and built upon the inherited hardware to turn it into the wide-field infrared observatory now heading into deep space.

Roman launched nine months ahead of schedule. Its baseline mission lasts five years, while the spacecraft has a 10-year mission goal. It was built to answer questions we’ve been struggling with for decades, like: Why is the universe expanding faster? Where is all the invisible matter shaping galaxies? How common are planetary systems like ours?

Singapore has committed S$220M, about $173M, to its fintech sector over the next three years. It made the announcement three days after KPMG reported that private investment in Singaporean fintech had fallen to its weakest first half in roughly a decade.

The timing is not a coincidence, and nobody is pretending otherwise. Gan Kim Yong, who is deputy prime minister, trade and industry minister and chairman of the Monetary Authority of Singapore, set out the commitment on Monday.

The money runs through the fourth iteration of the Financial Sector Technology and Innovation scheme, which has existed since 2015 and has backed more than 350 projects. FSTI 4.0 splits into six tracks covering institutional innovation, AI adoption, infrastructure and platforms, and talent.

The talent track has the most concrete commitment attached. At least 1,000 fintech internships will be created over the three years, with the stipends co-funded, which is a direct intervention in a labour market rather than a grant programme.

“These efforts will help our financial institutions, fintech firms, and workers innovate, scale, and build capabilities to seize emerging opportunities,” Gan said. He framed the AI opportunity in terms of capture rather than defence, adding that the financial industry is not a zero-sum game.

The private market has been telling a bleaker story. Singaporean fintech companies raised $499m across 53 deals in the first half of 2026, down from $1.45bn across 97 deals in the same period last year.

The composition is worse than the total. A single $320m cross-border payments round in June accounted for nearly two-thirds of the entire half, which means the rest of the ecosystem shared roughly $179m between it.

Deal count fell less steeply than value, from 97 to 53, which describes a market still doing business at smaller cheque sizes rather than one that has stopped. Early-stage activity holding up while growth rounds disappear is the usual shape of a downturn.

Split by sector, the pattern is familiar. AI and machine learning took $365.9m across 18 deals while digital assets managed 27 deals for $95.5m, so crypto is still doing volume, and AI is taking the money.

None of this makes Singapore a small market. It hosts around 1,800 fintech firms employing about 10,000 people, and the sector took roughly S$3bn during 2025 before the drop.

The decline is also not peculiar to Singapore, which is the more useful context. British fintech funding fell to its own decade low this year, and European rounds have been landing against a broader plunge rather than a recovery.

Singapore also has a specific reason to act rather than wait. Its financial centre competes directly with Hong Kong, Dubai and London for the same firms, and a hub that looks becalmed for two years does not simply resume where it left off.

What differs is the response. Britain reached for a £1bn private growth fund to close a gap it had identified, while Singapore is deploying a comparatively modest sum of public money into infrastructure, adoption and people.

S$220m will not replace a billion dollars of vanished private capital, and it is not designed to. Co-funding an internship or an AI deployment lowers the cost of trying something, which is a different instrument from writing a cheque into a cap table.

Whether that works depends on what caused the drop. If investors have simply repriced fintech everywhere, cheaper experimentation keeps companies alive until sentiment turns, and if Singapore is losing ground to a rival hub, a talent subsidy will not fix it.

FSTI has at least the advantage of a record. Eleven years and 350 projects is long enough to know what the scheme does well, which is more than most industrial policy can claim at the point of renewal.

The government is betting on the first reading. Three years, six tracks, a thousand internships, and a stated intention to be positioned when the money comes back.

IBM never sold this as a catalog item, yet the logo sits on the case all the same. Parco Pacified Limited of Hong Kong built the HUB-001SR as a compact speaker dock that also works as a three-port USB 1.1 hub and an FM radio, then someone stamped IBM branding on a batch and handed units out as corporate gifts or show swag. The same plastic shell later turned up with a Bisty mark tied to Japanese pachinko prizes. You end up with a desk object that looks official until you notice the generic Chinese tooling and the motor that slides a cover open when you press it.

Parco describes the product as a Windows and Mac-friendly speaker with a three-port hub and radio that requires no additional drivers to run with Windows 98 through XP or Mac OS 8.6 through OS X. Three USB-A sockets are included; these operate at the comparatively sluggish speed of USB 1.1, 12 Mbps, and are fitted with over-current protection to ensure safety. An AUX jack can be utilized to connect a music player with a 3.5 mm cable; any music player from that era would suffice. Each speaker can output 500 mW. The gadget may be operated using either a USB connection or four basic AAA batteries. FM reception is standard, with frequencies ranging from 88 to 108 MHz.

Sale

UGREEN 7-in-1 Powered USB 3.0 Hub, 7 Ports, 5xUSB-A and 2xUSB-C, 5Gbps

• 7-Port Expansion with Individual Switches: Turn one USB-A port into 5xUSB-A and 2xUSB-C data ports for keyboards, mice, flash drives, printers, and…
• Up to 5Gbps Data Transfer: Move large files, photos, and videos quickly with transfer speeds up to 5Gbps when connected to compatible hosts and…
• 24W Power Adapter Included: Use the power input to help keep multiple peripherals connected reliably, especially power-demanding storage devices. The…

A light press on the device’s front activates a small motor, causing a panel to glide open, exposing the ports and controls. When we examined LGR’s duplicate of the item, we discovered that the motor glided smoothly, as one would expect from a quality piece of equipment. Inside the machine, there are USB ports, an AUX jack, volume controls, and a radio button, as well as two drivers that aren’t particularly noteworthy. The radio is a basic scan tuner. It quickly searches the band, locks onto a station as soon as it detects a clean signal, and stays with it, but no useful options such as manual tuning are available. The device failed to pick up a signal inside a metal building, as expected, but the signal increased when we stepped outside, as you’d expect from a low-cost radio. Plugging it into a modern PC running Windows results in a generic USB hub with no extra sound card or software, so don’t expect to connect it to your PC and get a nice USB sound card. The audio remains on the speakers unless you connect it to the AUX input instead.

Parco still lists the item as an iPod USB speaker with auto-sliding cover on their website, which would explain the bulky appearance, even though later units never included an iPod port. However, the company’s catalog is a little amusing, with items like power banks shaped like old cassette tapes and milk-carton boomboxes. The lineup says it all, as the HUB-001SR was never intended for the IBM ThinkPad accessory drawer. It was only a freebie that happened to carry the Big Blue brand for a short while. The scan radio and motorized lid are an upgrade above a standard four-port brick, but the USB port is still the same old 12 Mbps hardware from the late 1990s. When the PC is turned off, four AAA batteries power the radio and speakers. On the other hand, connecting it to a computer via USB cable powers the entire device.

There are usually other things worth trying first.

Does Bluetooth make your Wi-Fi slower?

Blackzheep/Getty Images

What blocks the Wi-Fi signal the most?

The European Commission has classified ChatGPT as a search engine. Monday’s designation brings the chatbot under the Digital Services Act as a Very Large Online Search Engine, alongside Reddit and Roblox, which have been designated as Very Large Online Platforms.

ChatGPT is the first AI chatbot to be designated under the regime. The Commission announced the decision after signalling in July that it intended to act as soon as August.

The designation itself follows a fairly simple calculation. Services with more than 45 million monthly users in the EU fall under the DSA’s threshold, and ChatGPT reported 159 million monthly active users across the six months to March 2026.

Reddit and Roblox also cleared the threshold, although by a smaller margin. Reddit reported 57.2 million monthly users, while Roblox had roughly 48 million, putting both above the limit but well below ChatGPT’s reported reach.

The more unusual part is the decision to treat ChatGPT as a search engine. A chatbot that responds to questions in prose is now being regulated in the same category as services that organise and direct users to information on the web.

The distinction matters because the DSA imposes different requirements depending on what kind of service is being designated.

For ChatGPT, that means the Commission is applying search-related obligations to a system that does not simply retrieve and rank web pages, but generates answers.

That raises questions about how those answers are produced, what risks they create, and how those risks can be assessed.

The designation itself does not introduce a new set of rules. Instead, it brings the three companies into a compliance regime that requires annual systemic risk assessments covering areas including illegal content, the protection of minors, physical and mental wellbeing, fundamental rights, electoral processes and public security.

The companies will also have to submit to independent audits and provide access to data for regulators and vetted researchers.

Data access has been one of the more contentious parts of the DSA for large online platforms, particularly when researchers seek information that companies consider commercially sensitive.

“These new designations mean that ChatGPT, Reddit and Roblox will now be held to a higher standard of scrutiny and accountability,” said Henna Virkkunen, the Commission’s executive vice-president for tech sovereignty, security and democracy.

The companies have four months to comply, putting the deadline at the end of November. The DSA allows the Commission to impose fines of up to 6% of a company’s global annual turnover for serious violations.

For Roblox, the requirements around minors are particularly significant. Much of the platform’s audience is made up of children and teenagers, so the company will now have to formally assess and mitigate risks to their safety and wellbeing, with those measures subject to independent auditing.

OpenAI faces a different problem. Its assessment will have to address systemic risks around electoral processes and public security, among other areas.

For a general-purpose AI system used for everything from everyday questions to political information, that means translating broad concerns about how AI can influence people at scale into risks that can actually be identified, measured, and audited.

Reddit’s designation raises another set of questions. It is the smallest of the three services by user numbers, but it hosts a huge amount of public discussion and has become an increasingly valuable source of data for AI companies.

Reddit has spent the past two years turning access to that content into a licensing business, making the platform’s relationship with AI developers an increasingly important part of its commercial model.

The designations also have a specific legal anchor. They apply to named entities rather than simply to the consumer-facing brands: OpenAI Ireland Limited, Reddit Netherlands B.V. and Roblox Corporation are listed in the Commission’s register, with national regulators in Ireland and the Netherlands sharing enforcement responsibilities with the Commission.

The decision comes as European technology enforcement is becoming more active across several fronts.

Poland has asked the Commission to fine Meta €250 million over scam advertising, while the Commission is separately preparing measures that would require Google to open Android to rival AI assistants under the Digital Markets Act.

Neither OpenAI, Reddit nor Roblox has publicly challenged the designations. They can appeal the decisions, although doing so would put the companies in the unusual position of disputing a classification based largely on user numbers they have themselves reported.

The broader significance is that the DSA is now being applied well beyond the social networks and content platforms it was originally associated with.

In a single announcement, the Commission has brought a chatbot, an online forum and a platform heavily used by children under the same high-level regulatory framework, even though the services create very different kinds of risks.

The practical test will come in November, when the first risk assessments are due. For OpenAI in particular, the interesting question will be whether a framework designed around traditional online platforms can meaningfully account for a system that generates information rather than simply hosting or distributing it.

What just happened? A federal appeals court ruled on Friday that Nevada can enforce its gambling laws against Kalshi’s sports prediction markets, dealing the company a major setback in its fight to keep state regulators out of the business. The decision could give regulators across the West more leverage to challenge sports contracts offered on federally regulated prediction exchanges, while moving the industry closer to a Supreme Court fight over whether those products are financial derivatives or gambling.

The Ninth Circuit Court of Appeals found that Kalshi’s sports contracts amount to sports betting, even though the company offers them through a federally regulated exchange. The decision could have implications well beyond Nevada, particularly in states that have been trying to limit or block sports-related prediction markets.

Kalshi lets users trade contracts based on event outcomes. The platform lists markets tied to sports, elections, economic data, weather, entertainment, and other subjects. Its sports contracts have become a major part of its activity as prediction markets have grown quickly this year.

The company has argued that its contracts are federally regulated financial products, not gambling. Kalshi is regulated by the Commodity Futures Trading Commission and has said that federal oversight should prevent states from applying their own gaming laws to its operations.

The Ninth Circuit rejected that argument.

“The substance of the sports event contracts offered on Kalshi’s (exchange) is sports gambling, regardless of whether Kalshi calls them swaps,” the three-judge panel wrote.

The court also said Kalshi had not shown a meaningful difference between its sports markets and traditional sportsbook wagers. “Kalshi’s attempts to distinguish its sports event contracts from sportsbooks betting are unpersuasive,” the ruling said.

The panel took particular issue with Kalshi’s position that its products were not sports betting. The judges said it was “disingenuous” for the company to make that argument after using sports-betting language in earlier marketing materials.

The case began after Nevada regulators moved against Kalshi and similar prediction-market offerings from Crypto.com and Robinhood. Nevada has long treated sports wagering as a tightly regulated activity, and state officials argued that the platforms should be subject to the same rules and taxes as other betting operators.

Gov. Joe Lombardo said the ruling would help “safeguard the integrity of our gaming industry.” Mike Dreitzer, chair of the Nevada Gaming Control Board, said it “completely vindicates what we have been saying all along.”

Kalshi said it intends to challenge the ruling. Spokeswoman Dani Lever said the company believes CFTC regulations do not bar sports contracts and noted that the agency is considering further clarification of its rules.

“Despite the Ninth Circuit’s opinion, we still believe the CFTC regulations as written do not prohibit sports contracts, and in any event, the CFTC is working to clarify those regulations,” Lever said in a statement. “We will be seeking further review.”

The decision conflicts with a ruling earlier this year from the Third Circuit Court of Appeals. In that case, the court blocked New Jersey from enforcing its gaming laws against Kalshi and other prediction-market companies.

That split raises the odds that the Supreme Court will eventually decide whether federally regulated prediction exchanges can offer sports contracts without state gaming licenses.

“It’s the first ruling against Kalshi at the appellate level, and the opinion seemed to be pretty brutal for the company,” Dustin Gouker, an independent journalist who covers the prediction industry, told CNN. “This gets us one step closer to an almost inevitable Supreme Court case on the legality of sports event contracts.”

The Ninth Circuit’s ruling now applies to federal courts in California, Arizona and six other states in the circuit. It could give regulators in those states more leverage as they challenge prediction-market companies.

The ruling comes as the industry is gaining users and trades. Kalshi has reported billions of dollars in weekly trading volume, with sports markets driving much of that activity. Polymarket, the second-largest prediction platform in the United States, was not part of the Nevada case, but the legal issues raised by the ruling could affect its business as well.

ai and ml

Company’s three year pursuit of debt from non-existent address ended by Oxford judge

British energy company SSE Energy Supply refused to believe that there was no unit 8b at the property of Lyle Hopkins, a doctoral student at the University of Oxford.

For more than 20 months, the company billed Hopkins at business rates, a total of £1,091.01, for an unused electricity meter on the property that had been associated with a previous owner’s unsuccessful effort to convert the property into flats.

A company representative even sent an email on June 15, 2024, stating that Hopkins was not liable. Nonetheless, SSE then sent in debt collectors to recover debt attributed to a disconnected meter for a non-existent address.

He never paid the bill. But he wanted the harassment to end.

So Hopkins, a software engineer working on his PhD at Oxford Interdisciplinary Bioscience DTP, turned to GPT-5.5 and Claude Fable for legal guidance to challenge the unwarranted and annoying collection effort. Representing himself in court, he prevailed on July 17, when a judge at the County Court at Oxford (St Aldate) awarded him £1,087.88, including expenses and interest.

Citing The Register’s coverage of how courts have required lawyers to declare and verify AI-generated documents, Hopkins said his experience has been that it’s the verification that matters.

“I used AI to test an energy company’s court claims against its own records, then ran the hearing myself and won,” he said.

Hopkins said he wouldn’t have been able to quote the case law without the help of AI.

“I had to review everything and give it steering, and make sure the case law actually existed,” he said. “I mostly just gave it the facts and it came up with the arguments, checked the laws, court rules, case law, and regulations. Really I just knew what they had done couldn’t be right, and relied on the AI to find the legal justifications as to why.”

Hopkins said that when he first submitted his claim back on October 7, 2025, GPT models hallucinated more and weren’t as good at checking case law.

“If Fable had been around at the time, I probably would have claimed for a lot more, probably around £5,000 and included harassment, which would have been more in line with the case law it found when we were preparing the court documents about three weeks before the hearing.”

Hopkins didn’t consider a solicitor because the cost would have been prohibitive. His spending on AI was more modest but not trivial – three active AI subscriptions and extra API fees.

“I burned a lot of tokens on this, on my Github Copilot subscription, OpenAI subscription, and Anthropic subscription,” he said. “Due to the deadlines I ended up paying API rates for some of it, which was painful.” (~£175).

His award could have been higher. “Before the hearing, SSE offered me more than the court ultimately awarded, conditional on confidentiality and non-disparagement clauses,” Hopkins said. “I refused because it was more important to be able to tell people what they had done.”

A rollercoaster ride

The judge hearing the case – shared with The Register – said Hopkins, who sought the intercession of an Energy Ombudsman after repeated communication with SSE, could not have done more to make clear that he was not liable for claimed debt.

“The defendant has subjected Mr Hopkins to a rollercoaster ride and not a good one,” the judge said, adding, “pursuing Mr Hopkins for that liability, if continued, would amount to harassment.”

Yet even after Hopkins won, the billing contniued. Following the judgement on July 17, another bill  datedJuly 23 went out.

“Since the judgment…in direct contradiction to the ruling which stated any further attempts to claim the invalid debt from me would constitute harassment, they have sent me another bill with payment demand,” Hopkins told The Register.

He’s unsure whether he will pursue a harassment claim.

“This dragged on for nearly three years and cost hundreds of hours, lost sleep, and time that should have gone into my medical-research doctorate at Oxford,” he said. “My exhibits alone were over 100 pages. Given how difficult this was, I struggle to see how people less academic than me wouldn’t end up with CCJs and ruined credit ratings over debts they never owed.”

Hopkin’s doctoral funding ended before he could finish his degree, a delay he attributes to time spent on this case. So he has launched a funding campaign to support the completion of his doctorate. 

The Register emailed SSE seeking comment and received an autoreply indicating that that SSE media team was unavailable outside of business hours in the UK. The urgent inquiry number yielded a representative who indicated that he only handled questions about outages. If company representatives reply after returning to the office, we’ll update this story. ®