An anonymous reader quotes a report from MIT Technology Review: In a paper published in Nature today, the researchers say it is the first time a large language model has been used to discover a solution to a long-standing scientific puzzle — producing verifiable and valuable new information that did not previously exist. “It’s not in the training data — it wasn’t even known,” says coauthor Pushmeet Kohli, vice president of research at Google DeepMind. Large language models have a reputation for making things up, not for providing new facts. Google DeepMind’s new tool, called FunSearch, could change that. It shows that they can indeed make discoveries — if they are coaxed just so, and if you throw out the majority of what they come up with.

FunSearch (so called because it searches for mathematical functions, not because it’s fun) continues a streak of discoveries in fundamental math and computer science that DeepMind has made using AI. First Alpha Tensor found a way to speed up a calculation at the heart of many different kinds of code, beating a 50-year record. Then AlphaDev found ways to make key algorithms used trillions of times a day run faster. Yet those tools did not use large language models. Built on top of DeepMind’s game-playing AI AlphaZero, both solved math problems by treating them as if they were puzzles in Go or chess. The trouble is that they are stuck in their lanes, says Bernardino Romera-Paredes, a researcher at the company who worked on both AlphaTensor and FunSearch: “AlphaTensor is great at matrix multiplication, but basically nothing else.” FunSearch takes a different tack. It combines a large language model called Codey, a version of Google’s PaLM 2 that isfine-tuned on computer code, with other systems that reject incorrect or nonsensical answers and plug good ones back in.

The researchers started by sketching out the problem they wanted to solve in Python, a popular programming language. But they left out the lines in the program that would specify how to solve it. That is where FunSearch comes in. It gets Codey to fill in the blanks — in effect, to suggest code that will solve the problem. A second algorithm then checks and scores what Codey comes up with. The best suggestions — even if not yet correct — are saved and given back to Codey, which tries to complete the program again. After a couple of million suggestions and a few dozen repetitions of the overall process — which took a few days — FunSearch was able to come up with code that produced a correct and previously unknown solution to the cap set problem, which involves finding the largest size of a certain type of set. Imagine plotting dots on graph paper. […] To test its versatility, the researchers used FunSearch to approach another hard problem in math: the bin packing problem, which involves trying to pack items into as few bins as possible. This is important for a range of applications in computer science, from data center management to e-commerce. FunSearch came up with a way to solve it that’s faster than human-devised ones.

Read more of this story at Slashdot.

Source: Slashdot – Google DeepMind Uses LLM To Solve Unsolvable Math Problem

Emma Roth reports via The Verge: Google is releasing a tool to help users diagnose problems with their Pixel phones. Users can launch the app by entering #*#7287#*# on the dial pad, allowing them to check whether their phone is working correctly before or after a repair. There are a bunch of diagnostic tools available within the app. While users can run a full diagnostic test to detect issues across the entire device, there are also options to run individual tests for physical defects and problems affecting the phone’s display, sensors, and connectivity. The new tool is available on all Pixel phones in English.

Aside from the diagnostics app, Google is introducing redesigned repair manuals the company says are “easier for technicians and DIYers to use.” Users can download repair manuals from Google’s website in English and French, but they’re only available for the Pixel Fold, Pixel 8, and Pixel 8 Pro for now. Google says it will add repair manuals for previous and future devices “in the coming months.” There’s also a new Repair Mode Google introduced earlier this month, which Pixel Phone owners can toggle on when their device is getting repaired. The feature is meant to protect users’ private information while their phone is in the hands of a technician […].

Read more of this story at Slashdot.

Source: Slashdot – Google Releases On-Device Diagnostics Tool, Repair Manuals For Pixel Phones

Natasha Singer reports via The New York Times: According to new research from Stanford University, the popularization of A.I. chatbots has not boosted overall cheating rates in schools (Warning: source may be paywalled; alternative source). In surveys this year of more than 40 U.S. high schools, some 60 to 70 percent of students said they had recently engaged in cheating — about the same percent as in previous years, Stanford education researchers said. “There was a panic that these A.I. models will allow a whole new way of doing something that could be construed as cheating,” said Denise Pope, a senior lecturer at Stanford Graduate School of Education who has surveyed high school students for more than a decade through an education nonprofit she co-founded. But “we’re just not seeing the change in the data.”

ChatGPT, developed by OpenAI in San Francisco, began to capture the public imagination late last year with its ability to fabricate human-sounding essays and emails. Almost immediately, classroom technology boosters started promising that A.I. tools like ChatGPT would revolutionize education. And critics began warning that such tools — which liberally make stuff up — would enable widespread cheating, and amplify misinformation, in schools. Now the Stanford research, along with a recent report from the Pew Research Center, are challenging the notion that A.I. chatbots are upending public schools.

Read more of this story at Slashdot.

Source: Slashdot – Cheating Fears Over Chatbots Were Overblown, New Research Suggests

An anonymous reader quotes a report from Ars Technica: Google’s parent company, Alphabet, is launching a new company called “Renew Home.” The new company will pull in some other projects from Nest and the rest of Alphabet to become a supposed one-stop shop for power savings and clean energy usage. The core concept is partnering with power companies to obtain data about the current condition of the power grid and using that data to change consumer habits. The new company is bankrolled by Sidewalk Infrastructure Partners (SIP), an Alphabet venture capital firm.

The first existing service getting pulled into Renew Home is Nest Renew. This service for Nest Thermostats uses power company data to tell consumers how their electricity is being generated and what it costs. That data lets your thermostat do things like automatically shift heating and cooling to times of day when energy is cheaper or cleaner, and shows various reports about the cleanness of the energy you’ve been using. (Nest’s feature that lets utility companies remotely take control of your thermostat, Rush Hour Rewards, does not seem to be part of Renew Home.)

Another Alphabet service being pulled into Renew Home is OhmConnect, which is the same basic idea as Nest’s grid data-power thermostat adjustments but for more than just your thermostat. OhmConnect is compatible with a very small list of smart devices, like Nest-rival Ecobee and Honeywell thermostats, TL-Link’s “Kasa” smart home system, and Tesla vehicles. The backbone of the service appears to be the in-house “OhmPlug” smart outlet, which can monitor the energy usage of anything that plugs into the wall. By seeing that you’ve turned these smart devices during peak usage times, OhmConnect offers people rewards like gift cards or cash for not using power when the grid is at capacity.

Read more of this story at Slashdot.

Source: Slashdot – Alphabet’s ‘Renew Home’ Company Brings Power Grid Data To Your Smart Home

Amrita Khalid reports via The Verge: Proton has released a desktop version of its Proton Mail app that will give users full access to both Proton Mail and Proton Calendar and (eventually) the ability to view your emails offline. The desktop app is available in beta is optimized for both Windows and macOS, and encrypts sent emails end-to-end just like with the browser version, according to the Swiss company, while offline access to emails will be available “soon.” […] It’s important to note that you’ll still need internet access to both send and encrypt your emails on Proton. But the offline feature will let you view and draft emails while traveling, during a power outage, or any other situation where you don’t have access to the internet.

Proton is also bringing encrypted auto-forwarding to paid users, both on its desktop and browser versions, though the encryption for forwards will only apply when the forwarded emails go to other Proton users. The company says it has made improvements to Proton Calendar, too, including a fully searchable web version. Not everyone will be able to access Proton’s desktop app right away. Proton is restricting access to its paid “Visionary” tier for legacy users at first (though the company is reopening subscriptions to that tier through January 3rd, 2024). The plan is to make the desktop app available to all users in early 2024.

Read more of this story at Slashdot.

Source: Slashdot – Proton Mail Finally Gets a Desktop App For Encrypted Email and Calendar

Meta CEO Mark Zuckerberg announced that Threads is now available to users in the European Union. “Today we’re opening Threads to more countries in Europe,” wrote Zuckerberg in a post on the platform. “Welcome everyone.” MacRumors reports: The move comes five months after the social media network launched in most markets around the world, but remained unavailable to EU-based users due to regulatory hurdles. […] In addition to creating a Threads profile for posting, users in the EU can also simply browse Threads without having an Instagram account, an option likely introduced to comply with legislation surrounding online services.

The expansion into a market of 448 million people should see Threads’ user numbers get a decent boost. Meta CEO Mark Zuckerberg said on a company earnings call in October that Threads now has “just under” 100 million monthly users. Since its launch earlier this year it has gained a web app, an ability to search for posts, and a post editing feature.

Read more of this story at Slashdot.

Source: Slashdot – Threads Launches In the European Union

An anonymous reader quotes a report from TechCrunch: Hackers compromised the code behind a crypto protocol used by multiple web3 applications and services, the software maker Ledger said on Thursday. Ledger, a company that makes a widely used and popular crypto hardware and software wallet, among other products, announced on X (previously Twitter) that someone had pushed out a “malicious version” of its Ledger Connect Kit, a library that decentralized apps (dApps) made by other companies and projects use to connect to the Ledger wallet service.

“A genuine version is being pushed to replace the malicious file now. Do not interact with any dApps for the moment. We will keep you informed as the situation evolves,” Ledger wrote. Soon after, Ledger posted an update saying that the hackers had replaced the genuine version of its software some six hours earlier, and that the company was investigating the incident and would “provide a comprehensive report as soon as it’s ready.” After this story was published, Ledger spokesperson Phillip Costigan shared more details about the hack with TechCrunch and on X.

Costigan said that a former Ledger employee was victim of a phishing attack on Thursday, which gave the hackers access to their former employee’s NPMJS account, which is a software registry that was acquired by GitHub. From there, the hackers published a malicious version of the Ledger Connect Kit. “The malicious code used a rogue WalletConnect project to reroute funds to a hacker wallet,” Costigan said. Then, Ledger deployed a fix within 40 minutes of the company becoming aware of the hack. The malicious file, however, was live for round 5 hours, but “the window where funds were drained was limited to a period of less than two hours,” according to Costigan. Ledger also “coordinated” with WalletConnect which “quickly disabled the the rogue project,” essentially stopping the attack, according to Costigan. Costigan also said Ledger pushed out a genuine software update that is “safe to use.” “We are actively talking with customers whose funds might have been affected, and working proactively to help those individuals at this time,” the Ledger spokeperson said, adding that the company believes it has identified the hackers’ wallet.

Read more of this story at Slashdot.

Source: Slashdot – Supply Chain Attack Targeting Ledger Crypto Wallet Leaves Users Hacked

The Federal Communications Commission has taken a step toward prohibiting early termination fees charged by cable and satellite TV providers. From a report: If given final approval, the FCC action would also require cable and satellite providers to provide a prorated credit or rebate to customers who cancel before a billing period ends. The new rules are being floated in a Notice of Proposed Rulemaking (NPRM) that the FCC voted to approve this week in a 3-2 vote, with both Republicans dissenting. The NPRM seeks public comment on the proposed rules and could lead to a final vote in a few months or so.

“Today’s action proposes to adopt customer service protections that prohibit cable operators and DBS (Direct Broadcast Satellite) providers from imposing a fee for the early termination of a cable or DBS video service contract,” the FCC said. “Additionally, the NPRM recommends the adoption of customer service protections to require cable and DBS providers to grant subscribers a prorated credit or rebate for the remaining whole days in a monthly or periodic billing cycle after the subscriber cancels service.”

FCC Chairwoman Jessica Rosenworcel said, “Consumers are tired of these junk fees. They now have more choices when it comes to video content. But these friction-filled tactics to keep us subscribing to our current providers are aggravating and unfair. So today we kick off a rulemaking to put an end to these practices.” Cable lobby group NCTA-The Internet & Television Association opposes the plan and said it will submit comments to support “consumer choice and competitive parity.”

Read more of this story at Slashdot.

Source: Slashdot – FCC Floats Ban on Cable TV ‘Junk Fees’ That Make It Hard To Ditch Contracts

Amazon, which launched its first internet satellites in October, says it will use space lasers to ensure reliable broadband coverage even in the middle of the ocean. From a report: The technology will enable a satellite to deliver data to a cruise ship, say, even if the spacecraft isn’t in range of an Amazon ground station connected to the internet. Instead, the laser will pass the request on to another satellite that has a clear view of a ground station.

The two prototype satellites successfully tested the technology, which is officially called optical inter-satellite links, Amazon said on Thursday. “If you’re going to serve maritime customers, air customers, you have got to be able to get data up to your satellites,” said Rajeev Badyal, who leads Amazon’s Project Kuiper internet-from-space initiative. “And in the middle of the ocean, it’s difficult if not impossible” to install ground stations.

Read more of this story at Slashdot.

Source: Slashdot – Amazon’s Internet Satellites Will Use Space Lasers To Transmit Data

Google is about to launch its grand plan to block third-party cookies in Chrome that many websites use to track your activity across the web for profit. From a report: Starting on January 4th, Google will start testing its new Tracking Protection feature that will eventually restrict website access to third-party cookies by default. It will come to a very small subset of Chrome users at the start, specifically to one percent of users globally. Afterward, Google plans to phase out the use of third-party cookies for all users in the second half of 2024.

If you’re randomly selected to try Tracking Protection, Google will notify you when opening Chrome on desktop or Android. If there are issues detected by Chrome while you’re browsing, a prompt will appear asking if you’d like to temporarily re-enable third-party cookies for the site.

Read more of this story at Slashdot.

Source: Slashdot – Google Will Turn Off Third-Party Tracking for Some Chrome Users Soon

Intel unveiled new computer chips on Thursday, including Gaudi3, an AI chip for generative AI software. Gaudi3 will launch next year and will compete with rival chips from Nvidia and AMD that power big and power-hungry AI models. From a report: The most prominent AI models, like OpenAI’s ChatGPT, run on Nvidia GPUs in the cloud. It’s one reason Nvidia stock has been up nearly 230% year-to-date while Intel shares are up 68%. And it’s why companies like AMD and, now Intel, have announced chips that they hope will attract AI companies away from Nvidia’s dominant position in the market.

While the company was light on details, Gaudi3 will compete with Nvidia’s H100, the main choice among companies that build huge farms of the chips to power AI applications, and AMD’s forthcoming MI300X, when it starts shipping to customers in 2024. Intel has been building Gaudi chips since 2019, when it bought a chip developer called Habana Labs.

Read more of this story at Slashdot.

Source: Slashdot – Intel Unveils New AI Chip To Compete With Nvidia and AMD

General Motors’ Cruise robotaxi unit announced today that it will lay off 900 employees, or 24% of its workforce. The news follows the departure of nine executives amid an ongoing safety investigation following an inccident in which a pedestrian was dragged by one of the company’s self-driving cars. CNBC reports: The company had 3,800 employees before Thursday’s cuts, which also follow a round of contractor layoffs at Cruise last month. Affected employees will receive paychecks until Feb. 12 and at least an additional eight weeks of pay, plus severance based on tenure. A Cruise representative also told CNBC that the company’s goal is now to work on a fully driverless L4 service, as well as relaunching ride-hailing in one city to start. In a statement, a Cruise spokesperson said, “We shared the difficult news that we are reducing our workforce, primarily in commercial operations and related corporate functions. These changes reflect our decision to focus on more deliberate commercialization plans with safety as our north star. We are supporting impacted Cruisers with strong severance and benefits packages and are grateful to the departing employees who played important roles in building Cruise and supporting our mission.”

GM added, “GM supports the difficult employment decisions made by Cruise as it reflects their more deliberate path forward, with safety as the north star. We are confident in the team and committed to supporting Cruise as they set the company up for long-term success with a focus on trust, accountability and transparency.”

Read more of this story at Slashdot.

Source: Slashdot – GM’s Cruise Cuts 24% of Its Workforce

Adobe said US regulators are probing the company’s cancellation rules for software subscriptions, an issue that has long been a source of ire for customers. From a report: The company has been cooperating with the Federal Trade Commission on a civil investigation of the issue since June 2022, Adobe said Wednesday in a filing. A settlement could involve “significant monetary costs or penalties,” the company said.

Users of Adobe programs including Photoshop and Premiere have long complained about the expense of canceling a subscription, which can cost more than $700 annually for individuals. Subscribers must cancel within two weeks of buying a subscription to receive a full refund; otherwise, they incur a prorated penalty. Some other digital services such as Spotify and Netflix don’t charge a cancellation fee. Digital subscriptions have been a recent focus for the FTC. It proposed a rule in March that consumers must be able to cancel subscriptions as easily as they sign up for them.

“Too often, companies make it difficult to unsubscribe from a service, wasting Americans’ time and money on things they may not want or need,” President Joe Biden said in a social media post at the time. Adobe said the FTC alerted the company in November that commission staff say “they had the authority to enter into consent negotiations to determine if a settlement regarding their investigation of these issues could be reached. We believe our practices comply with the law and are currently engaging in discussion with FTC staff.”

Read more of this story at Slashdot.

Source: Slashdot – FTC is Investigating Adobe Over Its Rules for Canceling Software Subscriptions

The Economist reports: “Shake, shake, shake, shake,” Taylor Swift sings from the stage of Lumen Field in Seattle at 10.35 in the evening on July 22nd. The fans respond, enthusiastically; the stadium duly shakes; a nearby seismometer takes note. To pop aficionados “Shake it off” is an empowering up-tempo anthem played at 160 beats per minute. To the Pacific Northwest Seismic Network, which is designed to monitor earthquakes, it is a 2.6 hertz signal in which the amplitude of the acceleration was as large as one centimetre per second, per second.

The well-situated seismometer first came to public attention in January 2011, when it recorded the response of fans of the Seattle Seahawks, an American football team, to a magnificent touchdown by Marshawn Lynch, a running back known as “Beast Mode.” The “Beast Quake” went down in local sporting history. When Ms Swift came to town for two nights of her Eras tour, Jacqueline Caplan-Auerbach, a geology professor at Western Washington University, used the opportunity to learn more about how events in the stadium shake its surroundings. On December 11th she presented some of her conclusions at the American Geophysical Union’s autumn meeting in San Francisco.

[…] Dr Caplan-Auerbach wanted to see whether such resonant amplification might also be at play elsewhere, and to distinguish between the effect of the music itself and the audience’s response. Her concert-night data showed two distinct sets of signals, one in higher frequencies (30-80hz), one in lower frequencies (1-8hz). The higher-frequency signals were present during the sound check, when the band were on stage but the stadium empty, and absent during the concerts’ “surprise songs,” played without the band by Ms Swift alone. The lower frequencies were absent when the audience had yet to arrive. Clearly those higher frequencies were from the music itself.

Read more of this story at Slashdot.

Source: Slashdot – The Excitement of 70,000 Swifties Can Shake the Earth

The Information: In the future, [Google] plans to use Gemini across nearly its entire line of products, from its search engine to its productivity applications and an AI assistant called Pixie that will be exclusive to its Pixel devices, two people familiar with the matter said. […] The rollout of Pixie, an AI assistant exclusively for Pixel devices, could boost Google’s hardware business at a time when tech companies are racing to integrate their hardware with new AI capabilities. Pixie will use the information on a customer’s phone — including data from Google products like Maps and Gmail — to evolve into a far more personalized version of the Google Assistant, according to one of the people with knowledge of the project. The feature could launch as soon as next year with the Pixel 9 and the 9 Pro, this person said.

Read more of this story at Slashdot.

Source: Slashdot – Google Working on Upgraded Voice Assistant That Uses Gemini, Report Says

Intel launched its Intel Core Ultra processors for AI-powered PCs at its AI Everywhere event today. From a report: The big chip maker said these processors spearhead a new era in computing, offering unparalleled power efficiency, superior compute and graphics performance, and an unprecedented AI PC experience to mobile platforms and edge devices. Available immediately, these processors will be used in over 230 AI PCs coming from renowned partners like Acer, ASUS, Dell, Gigabyte, and more.

The Intel Core Ultra processors represent an architectural shift for Intel, marking its largest design change in 40 years. These processors harness the Intel 4 process technology and Foveros 3D advanced packaging, leveraging leading-edge processes for optimal performance and capabilities. The processors boast a performance-core (P-core) architecture enhancing instructions per cycle (IPC). Efficient-cores (E-cores) and low-power Efficient-cores (LP E-cores). They deliver up to 11% more compute power compared to competitors, ensuring superior CPU performance for ultrathin PCs.

Features of Intel Core Ultra
Intel Arc GPU: Featuring up to eight Xe-cores, this GPU incorporates AI-based Xe Super Sampling, offering double the graphics performance compared to prior generations. It includes support for modern graphics features like ray tracing, mesh shading, AV1 encode and decode, HDMI 2.1, and DisplayPort 2.1 20G.
AI Boost NPU: Intel’s latest NPU, Intel AI Boost, focuses on low-power, long-running AI tasks, augmenting AI processing on the CPU and GPU, offering 2.5x better power efficiency compared to its predecessors.
Advanced Performance Capabilities: With up to 16 cores, 22 threads, and Intel Thread Director for optimized workload scheduling, these processors boast a maximum turbo frequency of 5.1 GHz and support for up to 96 GB DDR5 memory capacity.
Cutting-edge Connectivity: Integrated Intel Wi-Fi 6E and support for discrete Intel Wi-Fi 7 deliver blazing wireless speeds, while Thunderbolt 4 ensures connectivity to multiple 4K monitors and fast storage with speeds of 40 Gbps.
Enhanced AI Performance: OpenVINO toolkits, ONNX, and ONNX Runtime offer streamlined workflow, automatic device detection, and enhanced AI performance.

Read more of this story at Slashdot.

Source: Slashdot – Intel Core Ultra Processors Debut for AI-powered PCs

Pope Francis has called for a legally binding international treaty to regulate AI, saying algorithms must not be allowed to replace human values and warning of a “technological dictatorship” threatening human existence. From a report: The pope made his call on Thursday in a message for the Roman Catholic Church’s World Day of Peace, which is celebrated on Jan. 1. The title of the message, which is traditionally sent to world leaders and heads of institutions such as the United Nations, is “Artificial Intelligence and Peace.”

[…] “The global scale of artificial intelligence makes it clear that, alongside the responsibility of sovereign states to regulate its use internally, international organizations can play a decisive role in reaching multilateral agreements and coordinating their application and enforcement,” Francis wrote in the message. “I urge the global community of nations to work together in order to adopt a binding international treaty that regulates the development and use of artificial intelligence in its many forms,” the pope said.

Read more of this story at Slashdot.

Source: Slashdot – Pope Francis Calls for Binding Global Treaty To Regulate AI

After investigating reports that some users aren’t getting iMessages on Beeper Mini and Beeper Cloud, Beeper says that Apple seems to be “deliberately blocking” iMessages from being delivered to about five percent of Beeper Mini users. From a report: The company says that uninstalling and reinstalling the app fixes the issue and that it’s working on a broader fix.

Apple didn’t immediately reply to a request for comment about Beeper’s new claim, and it hasn’t replied to my original request for comment, either. But given that the company has already blocked Beeper Mini before, it’s not too surprising that it seems to be taking action against the app again.

Read more of this story at Slashdot.

Source: Slashdot – Beeper Says Apple is Blocking Some iMessages

LinkedIn has set aside an effort to relocate its data center technology out of its physical facilities and into Microsoft’s Azure cloud, CNBC reported Thursday, citing people familiar with the matter. From a report: The decision not to proceed with the project, code-named “Blueshift,” marks a major reversal for LinkedIn, which announced its plan to move to Azure in 2019, three years after Microsoft acquired the company for $27 billion. LinkedIn had been using Azure for specific tasks.

The U-turn represents a setback for Microsoft, which is chasing Amazon
Web Services in the lucrative cloud infrastructure market and has been counting on cloud technology and services to fuel much of its growth. Microsoft CEO Satya Nadella ran the cloud business before elevation to his current job in 2014. Mohak Shroff, LinkedIn’s vice president of engineering, wrote in a 2019 blog post announcing Blueshift that “moving to Azure will give us access to a wide array of hardware and software innovations, and unprecedented global scale.” Staffers started to learn of the decision not to follow through with the Azure migration last year, said the sources, who asked not to be named because of confidentiality. Executives stressed that the project was being put on hold, rather than getting canceled altogether, they said.

Read more of this story at Slashdot.

Source: Slashdot – LinkedIn Shelved Planned Move To Microsoft Azure, Opting To Keep Physical Data Centers

Dropbox has enabled a default setting to share user data with OpenAI for an AI search feature, but says it’s only accessed when the feature is used, not to train models, and deleted in 30 days. From a report: Even with assurances of data privacy laid out by Dropbox on an AI privacy FAQ page, the discovery that the setting had been enabled by default upset some Dropbox users. The setting was first noticed by writer Winifred Burton, who shared information about the Third-party AI setting through Bluesky on Tuesday, and frequent AI critic Karla Ortiz shared more information about it on X.

Wednesday afternoon, Drew Houston, the CEO of Dropbox, apologized for customer confusion in a post on X and wrote, “The third-party AI toggle in the settings menu enables or disables access to DBX AI features and functionality. Neither this nor any other setting automatically or passively sends any Dropbox customer data to a third-party AI service.”

Read more of this story at Slashdot.

Source: Slashdot – Dropbox Spooks Users With New AI Features That Send Data To OpenAI When Used