ICYMI: Perplexity AI Launches Comet Browser

Google plans to begin testing its recently announced verification scheme for Android developers in the coming weeks, but there's still precious little information on how the process will work. F-Droid, the free and open source app repository, isn't waiting for the full rollout to take a position. In a blog post, F-Droid staff say that Google's plan to force devs outside Google Play to register with the company threatens to kill alternative app stores like F-Droid, according to an article from ArsTechnica. F-Droid has been around for about 15 years and is the largest source of free and open source software (FOSS) for Android. Because the apps in F-Droid are not installed via the Play Store, you have to sideload each APK manually, and Google is targeting that process in the name of security. However, there’s a LOT of confusion surrounding this latest move by Google with the Android ecosystem. According to an article from Android Police, First and foremost, Google clarifies that it is not killing or doing away with sideloading on Android with this move. Instead, it says, “our new developer identity requirements are designed to protect users and developers from bad actors, not to limit choice. We want to make sure that if you download an app, it’s truly from the developer it claims to be published from, regardless of where you get the app.” Still, it does not address the fact that developers must pay Google for identity verification. Plus, they must accept the company's terms and conditions and upload personally identifiable documents. Meanwhile, you can read F-Droid’s full take on the situation here. To say that this issue is “resolved” would be overly optimistic and simplistic. We, the users of Android, will have to sit back and see how this all shakes out.

A viral app called Neon, which offers to record your phone calls and pay you for the audio so it can sell that data to AI companies, has rapidly risen to the ranks of the top-five free iPhone apps since its launch last week, according to an article from TechCrunch. The app already has thousands of users and was downloaded 75,000 times yesterday alone, according to app intelligence provider Appfigures. Neon pitches itself as a way for users to make money by providing call recordings that help train, improve, and test AI models. But Neon has gone offline, at least for now, after a security flaw allowed anyone to access the phone numbers, call recordings, and transcripts of any other user, TechCrunch can now report.

Security researchers have uncovered critical vulnerabilities in Tile’s location trackers that could allow stalkers to covertly monitor users by exploiting the devices’ lack of encryption, according to an article from eSecurity Planet. The flaws highlight longstanding privacy concerns surrounding Bluetooth-enabled trackers, which are marketed as tools to help people locate lost items but can be exploited for invasive surveillance. According to researchers from the Georgia Institute of Technology, the way Tile devices communicate leaves owners exposed to stalking. Unlike Apple AirTags and Samsung SmartTags, which rotate both unique identifiers and MAC addresses, Tile only rotates the unique identifier, allowing adversaries to fingerprint and track a tag indefinitely. There’s an additional informative article from Wired that goes into a deeper explanation.

Perplexity AI on October 2 announced that its artificial intelligence-powered web browser Comet is available worldwide, and will be free to users, according to an article from CNBC. The Comet browser is designed to serve as a personal assistant that can search the web, organize tabs, draft emails, shop and more, according to Perplexity. The startup initially launched Comet in July to Perplexity Max subscribers for $200 a month, and the waitlist has ballooned to “millions” of people, the company said. Perplexity’s decision to provide Comet for free could help it attract more users as it works to fend off rivals like Google, OpenAI and Anthropic that have their own AI browser offerings.

Web browsers collect a lot of data and share it with the sites we visit, so if you're concerned about your privacy, it's worth wondering which browsers are best for keeping our online habits to ourselves, according to an article from Lifehacker. Whether you're an activist concerned about surveillance, someone doing research in a country where your topic can get you in trouble, or simply a person who doesn't want spying eyes on their search history, using a more private browser can be one of the simplest steps you can take towards less worry. The author spoke to William Budington, a Senior Staff Technologist on The Electronic Frontier Foundation's (EFF) Public Interest Team, and Janet Vertesi, an Associate Professor of Sociology at Princeton University who publishes extensive work on human-computer interaction and online privacy. They had subtly different opinions on which browsers are best for your privacy, but they definitively agreed on one thing: It's not Chrome.

AI has been used to paraphrase deadly proteins in ways that slipped past DNA security safeguards. A Microsoft-led team found that some AI-crafted ricin variants evaded detection entirely, according to an article from eWeek. In a study detailed in the journal Science, the researchers tested two major companies’ biosecurity screening techniques and found that up to 100% of the AI-generated ricin-like proteins evaded detection. The finding exposes how existing filters can fail when tested against AI-designed toxins. Around October 2023, Microsoft’s Eric Horvitz and Bruce Wittmann began a red-teaming study to probe weaknesses in DNA biosecurity safeguards. Borrowing a term from military strategy, the exercise was designed to mimic how a malicious actor might try to exploit artificial intelligence to bypass security controls. The team used open-source protein design models to digitally reformulate 72 proteins under legal control, including ricin, botulinum, and Shiga toxins. In total, they created more than 70,000 synthetic DNA sequences that could code for variant forms of these toxins. None of the sequences were manufactured in the lab, but they were run through the same biosecurity screening software used by DNA synthesis companies to flag dangerous orders.

Asahi Group Holdings has reported a cyberattack on its domestic operations in Japan, according to an article from Just Drinks. The Peroni beer and Nikka whisky owner said the incident happened earlier on September 29. In a short statement sent to Just Drinks just after midday, Asahi said its Japanese operations had seen a “system failure.” “On September 29, around 7:00 a.m. Japan time, Asahi Group experienced a system failure due to the impact of a cyberattack on operations in Japan. At this time, there is no estimated timeline for recovery. There has been no confirmed leakage of personal information or other data to external parties. The system failure is currently limited to our operations within Japan.” Reuters reported the company had stopped orders and shipments as a result of the incident.

Intel Corp., the embattled chipmaker now backed by the US government, introduced new products and manufacturing technology that are central to its turnaround bid, according to an article from Bloomberg. The company announced Thursday that its Panther Lake processor designs are in full production and will go on sale in laptops early next year. The new chips are made with 18A technology, which Intel says offers advantages that none of its competitors can match yet. The unveiling follows a furious six-month stretch for Chief Executive Officer Lip-Bu Tan. After taking the job in mid-March, he’s tried to shake up Intel while also seeking outside help. The US government has become the chipmaker’s biggest investor as part of an unconventional deal brokered by the White House, and Nvidia Corp. and SoftBank Group Corp. have acquired multibillion-dollar stakes.

On Friday, Oct. 3, Discord announced that a third-party service provider it uses for customer service efforts suffered a breach, according to an article from Lifehacker. It warned a “limited number of users” who had communications with certain Discord teams were affected, though the “unauthorized party” did not gain access to any Discord networks directly. In that initial announcement, Discord said a number of user data types might have been stolen. That included their names, usernames, email addresses, billing information, last four digits of credit cards, purchase histories, IP addresses, messages with Discord service agents, and “limited corporate data,” such as training materials and internal presentations. While all of this information is sensitive, it unfortunately isn't surprising to see it as part of a breach like this. However, Discord also revealed that the hackers may have also gained access to a “small number” of government ID images, including driver's licenses and passports. As it turns out, that “small number” turned out to be 70,000. Discord confirmed as much to The Verge on Wednesday. If you were among these affected users, Discord will have reached out to you via email.

Take note, Nvidia: OpenAI is tightening its grip on the AI economy. With a new partnership, the company is targeting the single biggest bottleneck in its supply chain: the chip, according to an article from eWeek. According to a press release published Monday, OpenAI has struck a deal with Broadcom to co-develop its first in-house AI processor. The move is part of an effort to reduce reliance on third parties, such as Nvidia, and exert greater control over the infrastructure that drives its models. “Partnering with Broadcom is a critical step in building the infrastructure needed to unlock AI’s potential,” OpenAI CEO Sam Altman said in a statement. “Developing our own accelerators adds to the broader ecosystem of partners, all building the capacity required to push the frontier of AI to provide benefits to all humanity.” This news comes just a week after OpenAI signed a multi-year partnership with chipmaker AMD to supply the processors powering its next-generation AI systems.

SimonMed Imaging has revealed that a data breach resulting from a ransomware attack has impacted more than 1.2 million individuals, according to an article from Security Week. According to its website, SimonMed Imaging is one of the largest medical imaging providers and physician radiology practices in the US, with more than 170 facilities across 10 states. The Arizona-based healthcare organization learned in late January 2025 that one of its vendors had been breached, and an investigation conducted by SimonMed showed that its own network had also been hacked. The probe revealed that hackers had access to SimonMed systems between January 21 and February 5, and they managed to steal information such as name, address, date of birth, health insurance information, driver’s license number, government-issued ID, SSN, financial account number, authentication credentials, and a wide range of medical information.

In August 2025, cybersecurity researchers uncovered a sophisticated hacking campaign exploiting Microsoft Edge’s Internet Explorer (I.E.,) mode to compromise users’ devices, according to an article from eSecurity Planet. By leveraging social engineering and zero-day vulnerabilities within IE’s outdated Chakra JavaScript engine, threat actors successfully bypassed modern browser protections. This discovery highlights the persistent risks of maintaining legacy compatibility features in today’s rapidly evolving digital landscape. Microsoft Edge’s IE mode was originally developed to provide compatibility for older web applications and technologies that relied on outdated frameworks such as ActiveX, Silverlight, or Flash. Many enterprises, government portals, and industrial systems still depend on these legacy components, making full deprecation impractical. However, attackers have now weaponized this compatibility feature to bypass modern browser security protections. In this campaign, adversaries combined social engineering and zero-day exploits to manipulate unsuspecting users into reloading web pages in IE mode.

U.S. cybersecurity company F5 disclosed that nation-state hackers breached its systems and stole undisclosed BIG-IP security vulnerabilities and source code, according to an article from Bleeping Computer. The company states that it first became aware of the breach on August 9, 2025, with its investigations revealing that the attackers had gained long-term access to its system, including the company's BIG-IP product development environment and engineering knowledge management platform. F5 is a Fortune 500 tech giant specializing in cybersecurity, cloud management, and application delivery networking (ADN) applications. The company has 23,000 customers in 170 countries, and 48 of the Fortune 50 entities use its products. BIG-IP is the firm's flagship product used for application delivery and traffic management by many large enterprises worldwide. F5 has released a Security Incident “report” on its website.

Microsoft ended official support for Windows 10 on October 14, 2025, but the operating system will continue to function. While major software vendors and game studios will eventually leave the platform behind, users who are concerned can begin planning alternatives immediately, and Commodore is now offering an escape route, according to an article from TechSpot. The official Commodore Twitter account recently promoted Vision OS as a potential alternative to Microsoft's ecosystem. Commodore OS Vision 3.0 is a Linux-based platform specifically designed to protect users from what the company describes as Big Tech's monopolistic practices, constant digital noise, and pervasive surveillance. The operating system can be downloaded for free, with Commodore providing detailed installation instructions. It is also included with the Commodore 64X PC, a modern reinterpretation of the classic computing brand featuring recent x86 CPUs and contemporary hardware components. I don’t know about you, but I wasn’t aware that this company was even still around!!!

Google has released an urgent security update for its Chrome browser, addressing a serious vulnerability that could allow attackers to take control of users’ systems simply by visiting a malicious website, according to an article from eSecurityPlanet. Chrome versions prior to 141.0.7390.107/.108 for Windows and Mac and 141.0.7390.107 for Linux are impacted. The Hong Kong CERT team stated, “A remote attacker could exploit this vulnerability to trigger remote code execution on the targeted system.” The vulnerability (CVE-2025-11756) affects Chrome’s Safe Browsing feature, a key layer of defense designed to protect users from phishing sites and malware downloads. It was discovered by a researcher in September 2025, and reported to Google. At the time of publication, Google has not released detailed technical details of this vulnerability. Because Safe Browsing runs with elevated privileges, a successful exploit could bypass Chrome’s sandbox protections—potentially granting full access to the underlying operating system.

A major worldwide outage hit several of the biggest websites and apps on the morning of October 20, 2025, with major names such as Snapchat, Roblox, Canva, and Duolingo all suffering downtime, according to an article from TechRepublic. Amazon Web Services (AWS), the cloud provider behind many of these platforms, is at fault for the disruption. It began reporting faults across multiple AWS services in the US-EAST-1 region at midnight Pacific Time, with the underlying DNS issue fully mitigated by 3AM PDT. However, three hours later, Amazon reported new API and connectivity issues, with further investigation underway to resolve them. Even after mitigation efforts, services running on AWS are expected to remain unstable for a few hours due to traffic spikes and other technical factors. Beyond the thousands of apps, games, and websites affected, some of Amazon’s own services also went down during the outage. Amazon’s retail site was unavailable for several hours, as was its Ring doorbell system. Delivery drivers in the UK reported that Amazon warehouses were unable to sign off on packages due to technical faults.

In a very public, very awkward lesson on the difference between “solving” a problem and simply “finding” the answer, OpenAI’s latest model, GPT-5, has stirred up controversy, according to an article from eWeek. Top researchers at the company took a premature victory lap on social media, claiming the AI had cracked a set of notoriously difficult mathematical riddles. The celebration, however, was short-lived, earning the AI giant a stern rebuke and a fresh wave of industry ridicule. Mathematician Thomas Bloom, who runs the Erdős Problems website, quickly poured cold water on the celebrations. Responding on X, Bloom wrote, “Hi, as the owner/maintainer of http://erdosproblems.com, this is a dramatic misrepresentation. GPT-5 found references, which solved these problems, that I personally was unaware of.” He clarified that when a problem is marked as “open” on his website, it simply means he hasn’t seen a paper that solves it, not that the problem is still unsolved. In other words, GPT-5 didn’t actually produce new proofs; it merely discovered papers that already contained the solutions. Even Google DeepMind’s CEO, Demis Hassabis, chimed in, calling the situation “embarrassing.” And Meta’s AI chief Yann LeCun took a sharper jab, quipping, “Hoisted by their own GPTards.”

For all the discoveries we’re making of faraway galaxies, we’re still struggling to fully understand our own galaxy, the Milky Way. For example, researchers have known for decades of an odd concentration of gamma rays near the center of the Milky Way, although they weren’t sure where the high-energy light was coming from. A new study proposes an entirely new perspective—that the light may actually be coming from neutron stars, as astronomers have suspected, according to an article from Gizmodo. If not, however, this could be the “first proof” of dark matter, according to the paper, published recently in Physical Review Letters. Given the evolution of the Milky Way, the researchers argue that the gamma ray excess most likely emerged from the collision of dark matter particles, the researchers claim.

NASA has officially confirmed that Earth has gained a new, albeit temporary, cosmic companion — a “quasi-moon” named 2025 PN7, according to an article from Southern Digest (and widely reported on by multiple media outlets). First discovered by the University of Hawaii during a telescope survey earlier this year, the asteroid travels in near synchronization with Earth’s orbit around the Sun. While it doesn’t qualify as a true moon, its unique movement pattern means it acts like a shadow companion, orbiting the Sun alongside us for decades. According to NASA’s calculations, this celestial partner will likely remain in Earth’s neighborhood until 2083, making it one of the longest-staying quasi-moons ever recorded. Measuring an estimated 18 to 36 meters wide — roughly the height of a four- to eight-story building — 2025 PN7 is small but significant, adding a new chapter to humanity’s understanding of near-Earth objects. Unlike the Moon that we see in the night sky, which is gravitationally bound to Earth, a quasi-moon follows an orbit around the Sun, not Earth itself. However, its orbit is synchronized in such a way that it appears to loop around Earth as both move through space. NASA compares the behavior of a quasi-moon to that of “a runner on a nearby lane”, keeping pace without ever colliding or falling behind. 2025 PN7’s orbit has been in step with Earth’s for nearly six decades, and scientists predict that it will maintain this synchronized dance for another half-century.

Running out of storage space forces impossible choices: delete old files, pay monthly cloud subscriptions that add up fast, or buy expensive SSDs that cost ten times more per gigabyte than traditional hard drives. The Seagate 20TB external hard drive just dropped to $229 on Amazon (down from $499) and puts it at barely over one cent per gigabyte, according to an article from Gizmodo. This is an all-time low for this much capacity, and for context, cloud storage typically runs around $10 per month for just 2TB, meaning you’d spend $100 annually for a tenth of what this drive offers as a one-time purchase. If you’re backing up years of photos, managing video projects, archiving work files, or hoarding a media library, this drive delivers more practical storage than any cloud service or SSD at this price point.

mRNA-based Covid vaccines from Pfizer-BioNTech or Moderna may have an unexpected benefit for cancer patients who undergo immunotherapy. A new study suggests that these vaccines might boost the effects of immunotherapy drugs, perhaps by alerting the immune system and helping direct immune cells to attack tumors, according to an article from STAT News. That’s in addition to helping protect against Covid, which can be particularly important for cancer patients, who can sometimes have weakened immune systems. The study found that advanced cancer patients who received a Covid vaccine within 100 days before taking an immunotherapy drug during the pandemic lived longer than patients who did not, in a retrospective analysis. Researchers from MD Anderson Cancer Center presented the study at the European Society for Medical Oncology conference in Berlin on Sunday.

A new review suggests that vitamin D supplements may help protect the ends of our chromosomes, known as telomeres, which play a vital role in slowing the aging process, according to an article from Science Daily. This finding has raised hopes that the “sunshine vitamin” could support longer-lasting health. Researchers found that taking 2,000 IU (international units, a standard measure for vitamins) of vitamin D daily helped preserve telomeres -- the tiny protective caps on our DNA that function like the plastic tips on shoelaces, preventing damage each time a cell divides. Each of our 46 chromosomes is capped with a telomere that becomes shorter every time a cell replicates. When these structures get too short, cells stop dividing and eventually die. Shortened telomeres have been linked to major age-related diseases such as cancer, heart disease, and osteoarthritis. Factors like smoking, chronic stress, and depression can speed up this shortening process, while inflammation in the body also contributes to it.

OpenAI announced on October 20, 2025, that it’s rolling out a new internet browser called Atlas that integrates directly with ChatGPT, according to an article from Wired (and widely reported on by multiple media outlets). Atlas includes features like a sidebar window people can use to ask ChatGPT questions about the web pages they visit. There’s also an AI agent that can click around and complete tasks on a user’s behalf. “We think that AI represents a rare, once-a-decade opportunity to rethink what a browser can be about,” OpenAI CEO Sam Altman said during a livestream announcing Atlas. “Tabs were great, but we haven’t seen a lot of browser innovation since then.” Atlas debuts as Silicon Valley races to use generative AI to reshape how people experience the internet. Google has also announced a plethora of AI features for its popular Chrome browser, including a “sparkle” button that launches its Gemini chatbot. Chrome remains the most used browser worldwide. OpenAI says the Atlas browser will be available starting today for ChatGPT users globally on macOS. Windows and mobile options are currently in the works. Atlas is free to use, though its agent features are reserved for subscribers to OpenAI’s ChatGPT Plus or ChatGPT Pro plans.

Google announced research that shows — for the first time in history — that a quantum computer can successfully run a verifiable algorithm on hardware, surpassing even the fastest classical supercomputers (13,000x faster), according to an article from Google’s Technology blog. It can compute the structure of a molecule, and paves a path towards real-world applications. Today’s advance builds on decades of work, and six years of major breakthroughs. Back in 2019, Google demonstrated that a quantum computer could solve a problem that would take the fastest classical supercomputer thousands of years. Then, late last year (2024), Google’s new Willow quantum chip showed how to dramatically suppress errors, solving a major issue that challenged scientists for nearly 30 years. Today’s breakthrough moves Google much closer to quantum computers that can drive major discoveries in areas like medicine and materials science.

Researchers at the University of Maryland have identified the gene responsible for a rare wheat variety that develops three ovaries in each flower instead of just one, according to an article from SciTechDaily. Because every ovary can grow into a grain, this finding could greatly increase the amount of wheat produced per acre. The discovery was detailed in the Proceedings of the National Academy of Sciences on October 14, 2025. The unusual three-ovary trait was first found in a naturally occurring mutant of common bread wheat, but scientists did not initially know what caused it. To uncover the genetic difference, the Maryland team created a precise map of the mutant wheat’s DNA and compared it with that of ordinary wheat. Their analysis revealed that a normally inactive gene, known as WUSCHEL-D1 (WUS-D1), had been activated. When WUS-D1 turns on early during flower formation, it enlarges the floral tissue and allows the plant to produce additional female organs, such as pistils or ovaries.

The Nobel Prize in Physics has been awarded to John Clarke, Michel H. Devoret and John M. Martinis for their work on quantum mechanics that is paving the way for a new generation of very powerful computers, according to an article from the BBC. “There is no advanced technology used today that does not rely on quantum mechanics, including mobile phones, cameras... and fibre optic cables,” said the Nobel Committee. The announcement was made by the Royal Swedish Academy of Sciences at a news conference in Stockholm, Sweden. “To put it mildly, it was a surprise of my life,” said Professor John Clarke, who was born in Cambridge, UK and now works at the University of California in Berkeley. Michel H. Devoret was born in Paris, France and is a professor at Yale University while John M. Martinis is a professor at University of California, Santa Barbara. The three winners will share prize money of 11 million Swedish kronor (£ 872,000).

The genes we inherit may have a subtle influence over whether we experiment with and make a habit of using cannabis, according to an article from Science Alert. A team led by researchers from Western University in Canada, the University of California, San Diego (UC San Diego), and personal genomics and biotechnology company 23andMe compared the full genomes of 131,895 individuals with self-reported frequency of cannabis use. In addition to identifying variations in two key genetic sequences associated with cannabis use, the study linked the same genetic patterns to more than a hundred other physical and mental health traits, providing clues on how drug use relates to our wellbeing. Numerous factors can affect whether an individual uses drugs for self-medication or recreational purposes, from where they live to how much money they have. Genetics often has a more indirect impact on our habits, making some prone to experimental or frequent use despite risks of ongoing harm.

A seemingly outlandish method to deliver oxygen rectally – one of the winners of the 2024 “Ig Nobel” satirical science prize – may one day actually help lung disease patients, after achieving a key step in clinical trials, according to an article from The Independent. The technique was first demonstrated in 2021 by Japanese researchers, who showed in experimental pig models that oxygen could be delivered to the body via the rectum in gas form. An enema-like process delivers super-oxygenated liquid to the large intestine, where the life-supporting gas is absorbed into the bloodstream. While this method of rescuing people with blocked airways led the research team to win a parody award, it might not be a joke after all.

Scientists are reporting the first compelling evidence in people that cognitive training can boost levels of a brain chemical that typically declines with age, according to an article from NPR. A 10-week study of people 65 or older found that doing rigorous mental exercises for 30 minutes a day increased levels of the chemical messenger acetylcholine by 2.3% in a brain area involved in attention and memory. The increase “is not huge,” says Étienne de Villers-Sidani, a neurologist at McGill University in Montreal. “But it's significant, considering that you get a 2.5% decrease per decade normally just with aging.” So, at least in this brain area, cognitive training appeared to turn back the clock by about 10 years.

AI “slop” has come for herbalism, a study published by a leading AI-detection company has found, according to an article from The Guardian. Originality.ai, which offers its tools to universities and businesses, says it scanned 558 titles published in Amazon’s herbal remedies subcategory between January and September this year, and found 82% of the books “were likely written” by AI. “This is a damning revelation of the sheer scope of unlabelled, unverified, unchecked, likely AI content that has completely invaded [Amazon’s] platform,” wrote Michael Fraiman, author of the study. “There’s a huge amount of herbal research out there right now that’s absolutely rubbish,” said Sue Sprung, a medical herbalist in Liverpool. “AI won’t know how to sift through all the dross, all the rubbish, that’s of absolutely no consequence. It would lead people astray.” One of the apparently AI-written books, Natural Healing Handbook, is a No 1 bestseller in Amazon’s skincare, aroma therapies and herbal remedies, subcategories. Its introduction touts the book as “a toolkit for self-trust”, urging readers to “look inward” for solutions. Natural Healing Handbook’s author is named as Luna Filby, whose Amazon page describes her as a “35-year-old herbalist from the coastal town of Byron Bay, Australia” and founder of the brand My Harmony Herb. Sarah Wynn, the founder of Wildcraft Journal, calls the book a “resource and an inspiration”. However, neither Luna Filby, My Harmony Herb, Wildcraft Journal or Sarah Wynn appear to have any online presence beyond the Amazon page for the book – an indication, said Fraiman, that they may not exist. The Guardian could find no evidence of the pair. Originality.ai’s tool flagged available samples of the text as AI-generated with “100 % confidence”.