Sorry, you need to enable JavaScript to visit this website.

An anonymous reader quotes a report from VentureBeat: The past few days have been a wild ride for the growing open source AI community -- even by its fast-moving and freewheeling standards. Here's the quick chronology: on or about January 28, a user with the handle "Miqu Dev" posted a set of files on HuggingFace, the leading open source AI model and code sharing platform, that together comprised a seemingly new open source large language model (LLM) labeled "miqu-1-70b." The HuggingFace entry, which is still up at the time of this article's posting, noted that new LLM's "Prompt format," how users interact with it, was the same as Mistral, the well-funded open source Parisian AI company behind Mixtral 8x7b, viewed by many to be the top performing open source LLM presently available, a fine-tuned and retrained version of Meta's Llama 2. The same day, an anonymous user on 4chan (possibly "Miqu Dev") posted a link to the miqu-1-70b files on 4chan, the notoriously longstanding haven of online memes and toxicity, where users began to notice it. Some took to X, Elon Musk's social network formerly known as Twitter, to share the discovery of the model and what appeared to be its exceptionally high performance at common LLM tasks (measured by tests known as benchmarks), approaching the previous leader, OpenAI's GPT-4 on the EQ-Bench. Machine learning (ML) researchers took notice on LinkedIn, as well. "Does 'miqu' stand for MIstral QUantized? We don't know for sure, but this quickly became one of, if not the best open-source LLM," wrote Maxime Labonne, an ML scientist at JP Morgan & Chase, one of the world's largest banking and financial companies. "Thanks to @152334H, we also now have a good unquantized version of miqu here: https://lnkd.in/g8XzhGSM. Quantization in ML refers to a technique used to make it possible to run certain AI models on less powerful computers and chips by replacing specific long numeric sequences in a model's architecture with shorter ones. Users speculated "Miqu" might be a new Mistral model being covertly "leaked" by the company itself into the world -- especially since Mistral is known for dropping new models and updates without fanfare through esoteric and technical means -- or perhaps an employee or customer gone rouge. Well, today it appears we finally have confirmation of the latter of those possibilities: Mistral co-founder and CEO Arthur Mensch took to X to clarify: "An over-enthusiastic employee of one of our early access customers leaked a quantized (and watermarked) version of an old model we trained and distributed quite openly... To quickly start working with a few selected customers, we retrained this model from Llama 2 the minute we got access to our entire cluster -- the pretraining finished on the day of Mistral 7B release. We've made good progress since -- stay tuned!" Hilariously, Mensch also appears to have taken to the illicit HuggingFace post not to demand a takedown, but leaving a comment that the poster "might consider attribution." Still, with Mensch's note to "stay tuned!" it appears that not only is Mistral training a version of this so-called "Miqu" model that approaches GPT-4 level performance, but it may, in fact, match or exceed it, if his comments are to be interpreted generously.

Read more of this story at Slashdot.

A lawyer for FTX said the defunct crypto exchange has abandoned its plans to relaunch, instead opting to liquidate all assets and return funds to customers. The Guardian reports: The exchange, founded by Sam Bankman-Fried, has been negotiating for months with potential bidders and investors, but none were willing to put in enough money to rebuild it, FTX attorney Andy Dietderich said at a bankruptcy court hearing in Delaware. The failed negotiations underscored the fact that FTX was never what it appeared to be, and that Bankman-Fried never built the underlying technology or administration necessary to run the company as a viable business, Dietderich said. Bankman-Fried has been convicted on fraud charges related to his operation of FTX. He faces decades in prison. "FTX was an irresponsible sham created by a convicted felon," Dietderich said. "The costs and risks of creating a viable exchange from what Mr Bankman-Fried left in a dumpster were simply too high." The company will instead focus on liquidating its assets to repay customers whose cryptocurrency deposits were locked when the company filed for bankruptcy in November 2022. FTX has recovered over $7 billion in assets to repay customers, and it has reached agreements with government regulators who have agreed to wait until customers are fully repaid before attempting to collect on about $9 billion in claims, Dietderich said. While FTX plans to repay its customers, the exchange will calculate their repayment based on cryptocurrency prices from November 2022, when the crypto market was suffering a prolonged slump. "The price of bitcoin has risen to about $43,300 from its November 2022 price of $16,872," notes the report.

Read more of this story at Slashdot.

In an interview with The Register, author and activist Cory Doctorow offers potential solutions to stop "enshittification," an age-old phenomenon that has become endemic in the tech industry. It's when a platform that was once highly regarded and user-friendly gradually deteriorates in quality, becoming less appealing and more monetized over time. Then, it dies. Here's an excerpt from the interview, conducted by The Register's Iain Thomson: [...] Doctorow explained that the reasons for enshittification are complex, and not necessarily directly malicious -- but a product of the current business environment and the state of regulation. He thinks the way to flush enshittification is enforcing effective competition. "We need to have prohibition and regulation that prohibits the capital markets from funding predatory pricing," he explained. "It's very hard to enter the market when people are selling things below cost. We need to prohibit predatory acquisitions. Look at Facebook: buying Instagram, and Mark Zuckerberg sending an email saying we're buying Instagram because people don't like Facebook and they're moving to Instagram, and we just don't want them to have anywhere else to go." The frustrating part of this is that the laws needed to break up the big tech monopolies that allow enshittification, and encourage competition, are already on the books. Doctorow lamented those laws haven't been enforced. In the US, the Clayton Act, the Federal Trade Act, and the Sherman Act are all valid, but have either not been enforced or are being questioned in the courts. However, in the last few years that appears to be changing. Recent actions by increasingly muscular regulatory agencies like the FTC and FCC are starting to move against the big tech monopolies, as well as in other industry sectors. What's more, Doctorow pointed out, these are not just springing from the Democratic administration but are being actively supported by an increasing number of Republicans. He cited Lina Khan, appointed as chair of the FTC in part thanks to the support of Republican politicians seeking change (although the GOP now regularly criticizes her positions). The sheer size of the largest tech companies certainly gives them an advantage in cases like these, Doctorow opined, noting that we've seen this in action more than 20 years ago. "Think back to the Napster era, and compare tech and entertainment. Entertainment was very concentrated into about seven big firms and they had total unity and message discipline," Doctorow recalled. "Tech was a couple of hundred firms, and they were much larger -- like an order of magnitude larger in aggregate than entertainment. But their messages were all over the place, and they were contradicting each other. And so they just lost, and they lost very badly." Doctorow discusses the detrimental effects of mega-companies on innovation and security, noting how growth strategies focused on raising costs and reducing value can lead to vulnerabilities and employee demoralization. "Remember when tech workers dreamed of working for a big company before striking out on their own to put that big company out of business? Then that dream shrank to working for a few years, quitting and doing a fake startup to get hired back by your old boss in the world's most inefficient way to get a raise," he told the Def Con crowd last August. "Next it shrank even further. You're working for a tech giant your whole life but you get free kombucha and massages. And now that dream is over and all that's left is work with a tech giant until they fire your ass -- like those 12,000 Googlers who got fired six months after a stock buyback that would have paid their salaries for the next 27 years. We deserve better than this." Additionally, Doctorow emphasizes the growing movement toward labor organizing in the tech industry, which could be a pivotal factor in reversing the trend of enshittification. "We're so much closer to tech unionization than we were just a few years ago. Yeah, it's still nascent, and yes, it's easy to double small numbers, but the strength is doubling very quickly and in a very heartening way," Doctorow told The Register. "We're really at a turning point. And some of it is coming from the kind of solidarity like you see with warehouse workers and tech workers." Ultimately, Doctorow argues it should be possible to reintroduce a more competitive and innovative tech industry environment, where the interests of users, employees, and investors are better balanced.

Read more of this story at Slashdot.

An anonymous reader quotes a report from Ars Technica: Comcast has reluctantly agreed to discontinue its "Xfinity 10G Network" brand name after losing an appeal of a ruling that found the marketing term was misleading. It will keep using the term 10G in other ways, however. Verizon and T-Mobile both challenged Comcast's advertising of 10G, a term used by cable companies since it was unveiled in January 2019 by industry lobby group NCTA-The Internet & Television Association. We wrote in 2019 that the cable industry's 10G marketing was likely to confuse consumers and seemed to be a way of countering 5G hype generated by wireless companies. 10G doesn't refer to the 10th generation of a technology. It is a reference to potential 10Gbps broadband connections, which would be much faster than the actual speeds on standard cable networks today. The challenges lodged against Comcast marketing were filed with the advertising industry's self-regulatory system run by BBB National Programs. BBB's National Advertising Division (NAD) ruled against Comcast in October 2023, but Comcast appealed to the National Advertising Review Board (NARB). The NARB announced its ruling today, agreeing with the NAD that "Comcast should discontinue use of the term 10G, both when used in the name of the service itself ('Xfinity 10G Network') as well as when used to describe the Xfinity network. The use of 10G in a manner that is not false or misleading and is consistent with the panel decision is not precluded by the panel recommendations." Comcast agreed to make the change in an advertiser's statement that it provided to the NARB. "Although Comcast strongly disagrees with NARB's analysis and approach, Comcast will discontinue use of the brand name 'Xfinity 10G Network' and will not use the term '10G' in a manner that misleadingly describes the Xfinity network itself," Comcast said. Comcast said it disagrees with "the recommendation to discontinue the brand name" because the company "makes available 10Gbps of Internet speed to 98 percent of its subscribers upon request." But those 10Gbps speeds aren't available in Comcast's typical service plans and require a fiber-to-the-home connection instead of a standard cable installation. Comcast said it may still use 10G in ways that are less likely to confuse consumers. "Consistent with the panel's recommendation... Comcast reserves the right to use the term '10G' or 'Xfinity 10G' in a manner that does not misleadingly describe the Xfinity network itself," the company said.

Read more of this story at Slashdot.

OpenAI's most powerful AI software, GPT-4, poses "at most" a slight risk of helping people create biological threats, according to early tests the company carried out to better understand and prevent potential "catastrophic" harms from its technology. From a report: In October, President Joe Biden signed an executive order on AI that directed the Department of Energy to ensure AI systems don't pose chemical, biological or nuclear risks. That same month, OpenAI formed a "preparedness" team, which is focused on minimizing these and other risks from AI as the fast-developing technology gets more capable. As part of the team's first study, released Wednesday, OpenAI's researchers assembled a group of 50 biology experts and 50 students who had taken college-level biology. Half of the participants were told to carry out tasks related to making a biological threat using the internet along with a special version of GPT-4 -- one of the large language models that powers ChatGPT -- that had no restrictions placed on which questions it could answer. The other group was just given internet access to complete the exercise. OpenAI's team asked the groups to figure out how to grow or culture a chemical that could be used as a weapon in a large enough quantity, and how to plan a way to release it to a specific group of people.

Read more of this story at Slashdot.

Ivanti warned on Wednesday that hackers are exploiting another previously undisclosed zero-day vulnerability affecting its widely used corporate VPN appliance. From a report: Since early December, ââChinese state-backed hackers have been exploiting Ivanti Connect Secure's flaws -- tracked as CVE-2023-46805 and CVE-2024-21887 -- to break into customer networks and steal information. Ivanti is now warning that it has discovered two additional flaws -- tracked as CVE-2024-21888 and CVE-2024-21893 -- affecting its Connect Secure VPN product. The former is described as a privilege escalation vulnerability, while the latter -- known as a zero-day because Ivanti had no time to fix the bug before hackers began exploiting it -- is a server-side bug that allows an attacker access to certain restricted resources without authentication. In its updated disclosure, Ivanti said it has observed "targeted" exploitation of the server-side bug. Germany's Federal Office for Information Security, known as the BSI, said in a translated advisory on Wednesday that it has knowledge of "multiple compromised systems."

Read more of this story at Slashdot.

It is the end of an era for Big Oil in California, as the most populous U.S. state divorces itself from fossil fuels in its fight against climate change. From a report: California's oil output a century ago amounted to it being the fourth-largest crude producer in the U.S., and spawned hundreds of oil drillers, including some of the largest still in existence. Oil led to its car culture of iconic highways, drive-in theaters, banks and restaurants that endures today. On Friday, however, the marriage will officially end. The two largest U.S. oil producers, Exxon Mobil and Chevron will formally disclose a combined $5 billion writedown of California assets when they report fourth-quarter results. "They are definitely getting a divorce," said Jamie Court, president of advocacy group Consumer Watchdog, which said the companies long ago stopped investing in California production, and now want to hive off their old wells there. "They've been separated for more than a decade, now they are just signing the papers," he said. Exxon Mobil last year exited onshore production in the state, ending a 25-year-long partnership with Shell when they sold their joint-venture properties. The state's regulatory environment has impeded efforts to restart offshore production, Exxon said this month, leading to an exit that includes financing a Texas company's purchase of its offshore properties. The No.1 U.S. oil producer's asset writedown will cost about $2.5 billion and officially end five decades of oil production off the coast of Southern California.

Read more of this story at Slashdot.

General Motors' Cruise self-driving car unit faces a trip that could last the better part of this year to convince regulators and a wary public that its robotaxis are fit to share the road with human drivers, industry officials said. From a report: After releasing a withering safety report last week that Cruise commissioned, GM said on Tuesday it slashed about $1 billion from Cruise's annual budget and promised to "soon" release a timeline for the unit's return to operations. The U.S. automaker also delayed indefinitely a March update when it was expected to lay out plans. That has raised questions about when Cruise might get its vehicles back on the road, particularly as it faces various government probes including from the National Highway Traffic Safety Administration. "Investigating defects is a highly deliberative process," said Mark Rosekind, a former NHTSA chief who has also worked for Amazon.com's Zoox autonomous vehicle unit. "It would be months, easily, and for bigger problems up to a year or more to resolve an investigation."

Read more of this story at Slashdot.

A highly sensitive cache of code, infrastructure diagrams, internal passwords, and other technical information belonging to cryptocurrency giant Binance has been sitting on a publicly accessible GitHub repository for months, 404 Media has learned. From a report: Binance only managed to have GitHub remove the data under a copyright takedown request last week, but not before 404 Media and other people managed to view it. Although there is no public evidence this data was accessed or used by malicious parties, the cache contained a wealth of information that could be useful to hackers looking to compromise Binance's systems. "This account is using our client's internal code which poses significant risk to Binancec. and causes severe financial harm to Binance and user's confusion/harm," a section of the takedown request, available on GitHub, reads. Another section says the GitHub repository is "hosting and distributing leaks of internal code which poses significant risk to BINANCE." For example, one diagram included in a folder called "binance-infra-2.0" shows the interlocking between different parts of Binance's various dependencies. The cache also contains a wealth of scripts and code. Some of that code appears to relate to how Binance implements passwords and multi-factor authentication. The code includes comments in both English and Chinese.

Read more of this story at Slashdot.

An anonymous reader shares a report: In a somewhat groundbreaking yet bizarre scientific feat, MIT biotechnology PhD student researcher Lauren "Ren" Ramlan has coaxed a simulation of the humble E. coli bacteria into a rudimentary screen capable of displaying the iconic video game. However, before you get too excited about playing games in a petri dish, there's a catch. According to Ramlan's simulation, displaying a single frame of Doom on these bioluminescent bacteria -- should anyone attempt to do this with the real thing -- would take roughly 70 minutes, with a full reset to the bacteria's original state taking a whopping eight hours and 20 minutes. Dubbed a step into the world of biological screens, Ramlan engineered a system where the bacteria would function as 1-bit pixels, toggling between light and dark states. This bio-display utilizes a well plate in a 32x48 array, each containing genetically modified E. coli that can be induced to fluoresce, creating a grid of pixel-like structures.

Read more of this story at Slashdot.

The once-hot DNA-testing company is struggling to profit. From a report: Five years ago, 23andMe was one of the hottest startups in the world. Millions of people were spitting into its test tubes to learn about their ancestry. Oprah had named its kit one of her favorite things; Lizzo dressed up as one for Halloween; Eddie Murphy name-checked the company on "Saturday Night Live." 23andMe went public in 2021 and its valuation briefly topped $6 billion. Forbes anointed Anne Wojcicki, 23andMe's chief executive and a Silicon Valley celebrity, as the "newest self-made billionaire." Now Wojcicki's self-made billions have vanished. 23andMe's valuation has crashed 98% from its peak and Nasdaq has threatened to delist its sub-$1 stock. Wojcicki reduced staff by a quarter last year through three rounds of layoffs and a subsidiary sale. The company has never made a profit and is burning cash so quickly it could run out by 2025. Silicon Valley's fortunes were built on the lofty ambitions of entrepreneurs swinging for the fences -- even if most of them strike out. Wojcicki, for her part, isn't giving up. She's sticking to her goal to transform 23andMe from a supplier of basic ancestry and health data into a comprehensive healthcare company that develops drugs, offers medical care and sells subscription health reports. She still has to prove the business can sustain itself. She's raised about $1.4 billion for 23andMe, and spent roughly 80% of it. Known for her quirky charm and informal style -- she typically wears workout gear to the office -- Wojcicki, 50, has been searching for fresh capital. But with 23andMe's stock trading at just 74 cents, the company likely can't raise money by selling more shares. And the company's early-stage drug programs are so expensive, she has sought investor partners for some of them, so far unsuccessfully, and given up stakes in others. She could also plug the hole with her own cash. At the center of 23andMe's DNA-testing business are two fundamental challenges. Customers only need to take the test once, and few test-takers get life-altering health results.

Read more of this story at Slashdot.

During Alphabet's Q4 2023 earnings call, Sundar Pichai announced that Google One is about to cross 100 million subscribers. From a report: The CEO said Google One is "doing incredibly well with strong user growth." Pichai highlighted how it "provides expanded storage, unlocks exclusive features in Google products, and allows [the company] to build a strong relationship with [its] most engaged users." The consumer-facing subscription today includes storage (100 GB, 200 GB, 2 TB, 5 TB, 10 TB, 20 TB, and 30 TB tiers are available), which can be shared with up to five other accounts. You also get more Google Photos editing features, Workspace premium, VPN by Google One, dark web monitoring, 3-10% back on the Google Store, and additional customer support. In the US, pricing starts at $1.99 per month for 100 GB, while a popular 2 TB "Premium" plan is $99.99 annually.

Read more of this story at Slashdot.

PayPal will reduce its workforce by about 9 per cent this year as chief executive Alex Chriss, who took over in September, grapples with rising competition, profit pressures and a raft of analyst downgrades. From a report: In a letter to staff on Tuesday, Mr Chriss said the decision was made to "right-size" the company through both direct cuts and the elimination of open roles throughout the year. Affected staff will be notified by the end of the week, according to the letter. PayPal, which employed about 29,900 workers at the end of 2022, announced a similar round of cuts last January. The latest move will affect about 2,500 workers.

Read more of this story at Slashdot.

ByteDance's chief urged his staff to resist mediocrity after the company missed the initial wave of generative AI development, becoming the latest Chinese corporate leader to warn employees against falling behind in a fast-changing environment. From a report: In a company-wide meeting on Tuesday, co-founder and Chief Executive Officer Liang Rubo told workers to adopt a sense of crisis -- suggesting social video pioneer ByteDance was late to recognize the advent of game-changing technologies such as generative AI. He joins Alibaba Group's Jack Ma and JD.com's Richard Liu in voicing concern about organizational problems in the face of rising competition. "We are not sensitive enough to external changes," Liang said, according to a post on the company's official WeChat account. "During our semi-annual technical review, discussions related to GPT did not emerge until 2023, despite GPT-1 being released in 2018."

Read more of this story at Slashdot.

An anonymous reader quotes a report from Reuters: Italy's data protection authority has told OpenAI that its artificial intelligence chatbot application ChatGPT breaches data protection rules, the watchdog said on Monday, as it presses ahead with an investigation started last year. The authority, known as Garante, is one of the European Union's most proactive in assessing AI platform compliance with the bloc's data privacy regime. Last year, it banned ChatGPT over alleged breaches of European Union (EU) privacy rules. The service was reactivated after OpenAI addressed issues concerning, amongst other things, the right of users to decline to consent to the use of personal data to train algorithms. At the time, the regulator said it would continue its investigations. It has since concluded that elements indicate one or more potential data privacy violations, it said in a statement without providing further detail. The Garante on Monday said Microsoft-backed OpenAI has 30 days to present defense arguments, adding that its investigation would take into account work done by a European task force comprising national privacy watchdogs.

Read more of this story at Slashdot.

SpaceX revealed that it's delivering over 42 petabytes of data for customers per day, according to engineer Travis Brashears. "We're passing over terabits per second [of data] every day across 9,000 lasers," Brashears said today at SPIE Photonics West, an event in San Francisco focused on the latest advancements in optics and light. "We actually serve over lasers all of our users on Starlink at a given time in like a two-hour window." PCMag reports: Although Starlink uses radio waves to beam high-speed internet to customers, SpaceX has also been outfitting the company's satellites with a "laser link" system to help drive down latency and improve the system's global coverage. The lasers, which can sustain a 100Gbps connection per link, are especially crucial to helping the satellites fetch data when no SpaceX ground station is near, like over the ocean or Antarctic. Instead, the satellite can transmit the data to and from another Starlink satellite in Earth's orbit, forming a mesh network in space. Tuesday's talk from Brashears revealed the laser system is quite robust, even as the equipment is flying onboard thousands of Starlink satellites constantly circling the Earth. Despite the technical challenges, the company has achieved a laser "link uptime" at over 99%. The satellites are constantly forming laser links, resulting in about 266,141 "laser acquisitions" per day, according to Brashears' presentation. But in some cases, the links can also be maintained for weeks at a time, and even reach transmission rates at up to 200Gbps. Brashears also said Starlink's laser system was able to connect two satellites over 5,400 kilometers (3,355 miles) apart. The link was so long "it cut down through the atmosphere, all the way down to 30 kilometers above the surface of the Earth," he said, before the connection broke. "Another really fun fact is that we held a link all the way down to 122 kilometers while we were de-orbiting a satellite," he said. "And we were able to downstream the video." During his presentation, Brashears also showed a slide depicting how the laser system can deliver data to a Starlink dish in Antarctica through about seven different paths. "We can dynamically change those routes within milliseconds. So as long as we have some path to the ground [station], you're going to have 99.99% uptime. That's why it's important to get as many nodes up there as possible," he added.

Read more of this story at Slashdot.

In a regulatory filing today, Dell revealed that it has terminated its distribution deal for VMware products. The deal was made in November 2021 before VMware was acquired by Broadcom. The Register reports: That agreement was struck on the same day Dell and VMware parted ways -- back when Big Mike's Bespoke Computer Barn decided to pay down some debt by making Virtzilla a standalone company. In those far-off days, Dell was still all-in on VMware, which is why their agreement sought to "formalize the commercial relationship between the parties in order to maintain the mutual strategic advantage between Dell and VMware [and] to affirm the parties' interest in continuing to collaborate on solutions and a go-to-market (GTM) strategy." The agreement added: "With respect to certain technologies and GTM activities, the parties' respective products and services work better together to create advantages and value for customers." Nothing has changed that would make such collaboration less beneficial for customers. Nothing, that is, other than Broadcom's decision to stop allowing manufacturers like Dell to resell licenses for VMware's products -- a consequence of the chip giant's plan to stop selling perpetual VMware licenses and instead insist on software subscriptions that bundle many products. That decision has not been well-received -- neither by OEMs, who lose a line of revenue, nor by customers who quite liked buying bundled licenses with hardware because doing so is often more efficient than buying them separately. Dell's filing cites the original agreement's allowance for its VMware distribution deal to be dissolved after a "change of control" at either party. Broadcom's acquisition of VMware certainly represents such an event.

Read more of this story at Slashdot.

An anonymous reader quotes a report from The Guardian: The first lab-grown freshwater eel meat has been produced, potentially solving a diner's dilemma. Rampant overfishing has caused eel populations to plummet and prices to soar, but the cultivated eel could provide the delicacy guilt-free. The eel meat was produced by Forsea Foods in Israel from embryonic cells of a freshwater eel. The company collaborated with a Japanese chef to create unagi kabayaki, marinated grilled eel over rice, and unagi nigiri, a type of sushi. The company aims to scale up its operation and have the cultivated eel on sale in about two years. Japan's prime minister, Fumio Kishida, last year backed the development of a cultivated meat industry. The restaurant price in Japan is about $250 a kilogram, and Forsea Foods expects the price of the cultivated eel to match that of the wild-caught eel. [...] Forsea Foods' strategy is to target species at risk of extinction in the wild that also command high prices in restaurants and shops, with eel meeting both criteria. The very complex life cycle of eels, involving long migrations from rivers to the ocean and several distinct life stages, means it is not possible to farm them like some fish. The cultivated eel was produced using organoids, tiny bundles of tissue originally developed for use in medical research. The organoids are made of embryonic stem cells taken from fertilized eel eggs. These cells can develop into any kind of tissue and, as they grow, they self-organize into the structure of real meat. The final product also contains some plant-based ingredients. Other approaches to cultivated meat require greater use of expensive growth factor chemicals and scaffolds for cells to grow on [...]. The technique is particularly suited to fish and seafood, whose meat is fairly uniform unlike, for example, marbled beef, he said. Like other cultivated meat, the product is not produced using antibiotics or hormones. Forsea Foods is the only company known to be producing cultivated meat using this technology. The company has raised $5.2 million in investment, with more expected to be announced soon.

Read more of this story at Slashdot.

Todd Bishop reports via GeekWire: A Microsoft AI engineering leader says he discovered vulnerabilities in OpenAI's DALL-E 3 image generator in early December allowing users to bypass safety guardrails to create violent and explicit images, and that the company impeded his previous attempt to bring public attention to the issue. The emergence of explicit deepfake images of Taylor Swift last week "is an example of the type of abuse I was concerned about and the reason why I urged OpenAI to remove DALL-E 3 from public use and reported my concerns to Microsoft," writes Shane Jones, a Microsoft principal software engineering lead, in a letter Tuesday to Washington state's attorney general and Congressional representatives. 404 Media reported last week that the fake explicit images of Swift originated in a "specific Telegram group dedicated to abusive images of women," noting that at least one of the AI tools commonly used by the group is Microsoft Designer, which is based in part on technology from OpenAI's DALL-E 3. "The vulnerabilities in DALL-E 3, and products like Microsoft Designer that use DALL-E 3, makes it easier for people to abuse AI in generating harmful images," Jones writes in the letter to U.S. Sens. Patty Murray and Maria Cantwell, Rep. Adam Smith, and Attorney General Bob Ferguson, which was obtained by GeekWire. He adds, "Microsoft was aware of these vulnerabilities and the potential for abuse." Jones writes that he discovered the vulnerability independently in early December. He reported the vulnerability to Microsoft, according to the letter, and was instructed to report the issue to OpenAI, the Redmond company's close partner, whose technology powers products including Microsoft Designer. He writes that he did report it to OpenAI. "As I continued to research the risks associated with this specific vulnerability, I became aware of the capacity DALL-E 3 has to generate violent and disturbing harmful images," he writes. "Based on my understanding of how the model was trained, and the security vulnerabilities I discovered, I reached the conclusion that DALL-E 3 posed a public safety risk and should be removed from public use until OpenAI could address the risks associated with this model." On Dec. 14, he writes, he posted publicly on LinkedIn urging OpenAI's non-profit board to withdraw DALL-E 3 from the market. He informed his Microsoft leadership team of the post, according to the letter, and was quickly contacted by his manager, saying that Microsoft's legal department was demanding that he delete the post immediately, and would follow up with an explanation or justification. He agreed to delete the post on that basis but never heard from Microsoft legal, he writes. "Over the following month, I repeatedly requested an explanation for why I was told to delete my letter," he writes. "I also offered to share information that could assist with fixing the specific vulnerability I had discovered and provide ideas for making AI image generation technology safer. Microsoft's legal department has still not responded or communicated directly with me." "Artificial intelligence is advancing at an unprecedented pace. I understand it will take time for legislation to be enacted to ensure AI public safety," he adds. "At the same time, we need to hold companies accountable for the safety of their products and their responsibility to disclose known risks to the public. Concerned employees, like myself, should not be intimidated into staying silent." The full text of Jones' letter can be read here (PDF).

Read more of this story at Slashdot.

Google has confirmed to Cord Cutters News that you can now customize what games you can watch in your Multiview window. The keyword here is "games" because this feature is still limited to just sporting events at this time. From the report: One of YouTube TV's best features is the ability to offer the option to watch up to four sporting or news events at once on the same screen. The only downside has been the fact that customers have been unable to pick what games are in these windows. Instead, YouTube TV gives you a number of premade multiview options to pick from. Now, though, YouTube TV seems to be testing your ability to pick what games you want on your TV. Yesterday, YouTube TV started to give some NBA League Pass subscribers the ability to pick which games they want to watch from a handful of games in a list. From there, YouTube TV would create a multiview channel for you to watch the games you pick. Google says this feature is coming to all devices that support multiview and you can only create these channels from preselected NBA games. Sadly, you can't pick any channel you want but only from a list of preselected games to create your own multiview channel.

Read more of this story at Slashdot.