chat: AI news

"Codeberg is a democratic community-driven, non-profit software development platform operated by Codeberg e.V. and centered around Codeberg.org, a Forgejo-based software forge. "On Codeberg you can develop your own Free Software projects, contribute to other projects, browse through inspiring and useful free software, share your knowledge or build your projects a home on the web using Codeberg Pages, just to name a few. "Codeberg is not a for-profit corporation but an open community of free software enthusiasts providing a humane, non-commercial and privacy-friendly alternative to commercial services such as GitHub." 

The U.S. Transportation Security Administration (TSA) has moved forward with regulations establishing a smartphone-based national digital ID and tracking system, disregarding objections from organizations like the Identity Project. These new rules are based on the REAL-ID Act of 2005 and set the standards for how states will issue digital versions of driver’s licenses or ID cards, which will be accepted by the TSA and other federal agencies for certain official purposes. However, these rules do not extend to airline travel, where no ID is legally required, despite the TSA’s misleading statements to the contrary. Under the TSA’s new regulations, digital IDs can only be issued to individuals who already possess a physical driver’s license or state-issued ID card. Additionally, individuals are still obligated by state laws to carry their physical ID while driving, even if they have a digital ID on their smartphone.  This shift is not primarily about improving driver’s license technology,

"A Project of the Electronic Frontier Foundation "We’re the Electronic Frontier Foundation, an independent non-profit working to protect online privacy for over thirty years. This is Surveillance Self-Defense: our expert guide to protecting you and your friends from online spying. "Read the BASICS to find out how online surveillance works. Dive into our TOOL GUIDES for instructions to installing our pick of the best, most secure applications. We have more detailed information in our FURTHER LEARNING sections. If you’d like a guided tour, look for our list of common SECURITY SCENARIOS ." 

"The past 3 years of work in NLP [Natural Language Processing] have been characterized by the development and deployment of ever larger language models, especially for English.  "BERT, its variants, GPT-2/3, and others, most recently Switch-C, have pushed the boundaries of the possible both through architectural innovations and through sheer size.  "Using these pretrained models and the methodology of fine-tuning them for specific tasks, researchers have extended the state of the art on a wide array of tasks as measured by leaderboards on specific benchmarks for English.  "We provide recommendations including  Weighing the environmental and financial costs first,  Investing resources into curating and carefully documenting datasets rather than ingesting everything on the web,  Carrying out pre-development exercises evaluating how the planned approach fits into research and development goals and supports stakeholder values, and  Encouraging research directions beyond

"A Shoggoth is a perfect metaphor for AIs because no one knows what Shoggoths are.   "So, everyone can agree that 'AI is a Shoggoth ' means that AI is the thing that they already think AI is. (Apologies to the readers who propelled At the Mountains of Madness to the 776,464th place on Amazon’s best-selling list.)  In particular, Farrell and Shalizi call markets and democracies 'Shoggoths.'  "Wikipedia says that 'Cthulhu Mythos media ​​most commonly portray shoggoths as intelligent to some degree, but deal with problems using only their great size and strength': a description that almost seems lifted from the ' Stochastic Parrots ' paper. "Scott Alexander describes how the metaphors of Agent, Genie, or Oracle are not a good match for (non fine-tuned / RLHF’ed) language models like GPT-3. Rather they are best captured as what Janus called a Simulator and Scott  [Alexander] calls 'an unmasked Shoggoth.'  "In both the marke

 

Plans by Mark Zuckerberg's Meta to build an AI data centre in the US that runs on nuclear power were thwarted in part because a rare species of bee was discovered on land earmarked for the project, according to people familiar with the matter. Zuckerberg had planned to strike a deal with an existing nuclear power plant operator to provide emissions-free electricity for a new data centre supporting his artificial intelligence ambitions. However, the potential deal faced multiple complications including environmental and regulatory challenges , these people said. The discovery of the rare bee species on a location next to the plant where the data centre was to be built would have complicated the project, Zuckerberg told a Meta all-hands meeting last week, according to two people familiar with the meeting.

Google said while it was not clear whether Google Maps was involved in the incident, it had also made improvements so similar incidents did not occur in future. It said its teams were working on improvements to the timing of audio to provide enhanced guidance in situations such as those in this case. A spokesperson said: "This will involve an amended audio prompt as a driver approaches a junction where they would cross an overpass —'after the overpass, turn right'. "We hope that this will provide clearer guidance for drivers and help reduce the risk that they make an incorrect turn." TomTom recommended users updated their systems after it recently implemented additional safeguards to "limit drivers’ confusion where possible". It said this had been done by timing the verbal commands closer to the actual exit, and after passing the off-slip road. 

Mr. Benford [BoE’s Chief Data Officer] stressed the need for AI governance to ensure an ethical, safe, and effective use of AI.   Underscoring the significance of data to AI solutions, he advocated that the new wave of AI solutions is a reason to double down on strengthening data foundations.   In particular, the BoE is broadening the scope of its data governance framework to better address unstructured data. In addition to data governance, the BoE is working towards AI literacy for everyone at the BoE and AI fluency for its expert data professionals. The BoE is developing its use of AI through a series of targeted experiments which it is using to build its AI strategy. 

Eukaryotes invented organization, if we use the literal definition of “organize”: to be furnished with organs.   Inside a eukaryotic cell are self-contained, membrane-bound bundles that perform special functions, called organelles.   All eukaryotic cells —animal, plant, fungus or protist —have a nucleus that encloses and protects DNA.  Nearly all of them have mitochondria, which produce energy to fuel biochemical reactions. Any eukaryotic lineages that lack mitochondria used to have them and then lost them sometime in evolutionary history.  And across the evolutionary tree, different eukaryotes have evolved or procured additional organelles that  Assemble proteins,  Store water,  Turn sunlight into energy,  Digest biomolecules,  Get rid of waste, and more.  If prokaryotes are a loose pile of papers on the floor, eukaryotes are a sophisticated filing system that binds pages into packets and labels them.  “They’ve got the endoplasmic reticulum, Golgi apparatus, peroxisomes, lysosomes, va

 

 

"The AI industry should invest in frameworks that reward creators of all kinds for sharing valuable content.    "From YouTube to TikTok to X, tech platforms have proven they can administer novel rewards for distributed creators in complex content marketplaces. Indeed, fairer monetization of everyday content is a core objective of the web3 movement celebrated by venture capitalists.  "The same reasoning carries over to AI search. If queries yield lucrative engagement but users don’t click through to sources, commercial AI search platforms should find ways to attribute that value to creators and share it back at scale. "Of course, it’s possible that our digital economy was broken from the start. Subsistence on trickle-down ad revenue may be unsustainable, and the attention economy has inflicted real harm to privacy, integrity, and democracy online. Supporting quality news and fresh content may require other forms of investment or incentives.  "But we shouldn’t g

"China has extended its dominance at home and abroad over critical minerals that are essential to future high-tech and renewable-energy industries.   "Amid intensifying geopolitical competition, Western countries are increasing their efforts to claw back market share while countries in the Global South, where many of these minerals are mined, are attempting to capitalize on growing global demand.  "A recent article on the subject by The Economist stated that in 2023 Chinese companies invested roughly $16 billion in foreign mines, the highest figure in a decade, up from less than $5 billion the year before.  "This month, Chinese companies have announced plans to invest billions of dollars in mines in Afghanistan, Ghana, Zambia, and the Philippines.   "Keith Bradsher at The New York Times reported that over the past few weeks, the Chinese government has enacted measures to increase its grip over the mining and refining of rare minerals within China by making i

Llama 4 development is well underway, Zuckerberg told investors and analysts on an earnings call, with an initial launch expected early next year.   “We're training the Llama 4 models on a cluster that is bigger than 100,000 H100s, or bigger than anything that I've seen reported for what others are doing,” Zuckerberg said, referring to the Nvidia chips popular for training AI systems. “I expect that the smaller Llama 4 models will be ready first.” Increasing the scale of AI training with more computing power and data is widely believed to be key to developing significantly more capable AI models.  While Meta appears to have the lead now, most of the big players in the field are likely working toward using compute clusters with more than 100,000 advanced chips.  On Wednesday, Zuckerberg declined to offer details on Llama 4’s potential advanced capabilities but vaguely referred to “new modalities,” “stronger reasoning,” and “much faster .” 

"Sophos analysts identified a series of hacking campaigns that had started with indiscriminate mass exploitation of its products but eventually became more stealthy and targeted, hitting   Nuclear energy suppliers and regulators,  Military targets including a military hospital,  Telecoms,  Government and  Intelligence agencies, and the  Airport of one national capital.  "While most of the targets —which Sophos declined to identify in greater detail —were in South and Southeast Asia, a smaller number were in Europe, the Middle East, and the United States. "Sophos' report ties those multiple hacking campaigns —with varying levels of confidence —to Chinese state-sponsored hacking groups including those known as APT41, APT31, and Volt Typhoon, the latter of which is a particularly aggressive team that has sought the ability to disrupt critical infrastructure in the US, including power grids.  "But the common thread throughout those efforts to hack Sophos' devic

"We find the reference to artificial intelligence (AI) in many documents and debates of the policy realm, assigning it a strong potential to contribute to all these domains.   "AI for the public interest, and its close relatives, AI for (common or social) good, have become a common theme not only for tech companies, but also for political actors in the EU, including for instance international NGO networks.  "However, most often the definition of the public interest in the best case is limited to references to AI ethics. Yet, the practical meaning of what a good use of AI and a purpose 'for good' entails in its development and implementation is unclear.   "What is often missing is an understanding that spells out in practice what it means for the process of development and deployment of AI systems to serve the public interest, let alone a holistic view on the conditions for AI to best serve the collective well-being."   

"Meta is 'working with the public sector to adopt Llama across the US government,' according to CEO Mark Zuckerberg. "The comment, made during his opening remarks for Meta’s Q3 earnings call on Wednesday, raises a lot of important questions:  Exactly which parts of the government will use Meta’s AI models?  What will the AI be used for?  Will there be any kind of military-specific applications of Llama?  Is Meta getting paid for any of this? "When I asked Meta to elaborate , spokesperson Faith Eischen told me via email that ' we’ve partnered with the US State Department to see how Llama could help address different challenges —from expanding access to safe water and reliable electricity, to helping support small businesses.'  "She also said the company has 'been in touch with the Department of Education to learn how Llama could help make the financial aid process more user friendly for students and are in discussions with others about how Llama

Google is building a bunch of AI products, and it’s using AI quite a bit as part of building those products, too.   “More than a quarter of all new code at Google is generated by AI, then reviewed and accepted by engineers,” CEO Sundar Pichai said on the company’s third quarter 2024 earnings call. It’s a big milestone that marks just how important AI is to the company. AI is helping Google make money as well. Alphabet reported $88.3 billion in revenue for the quarter, with Google Services (which includes Search) revenue of $76.5 billion, up 13 percent year-over-year, and Google Cloud (which includes its AI infrastructure products for other companies) revenue of $11.4 billion, up 35 percent year-over-year. Operating incomes were also strong. Google Services hit $30.9 billion, up from $23.9 billion last year, and Google Cloud hit $1.95 billion, significantly up from last year’s $270 million. The results indicate that, while many people feel Google isn’t as reliable as it once was, the c

"This assignment invites undergraduate students to create spells , a poem based on a wish, and prompt an LLM to write a spell of its own, then reflect on their input and LLM output compared to their original compositions.   "Students analyzed the function of language and intent in manual and LLM composition, drawing together meaning and expression, and how LLM technology replicates or revises that expression.   "This assignment can be adapted to most rhetoric, composition, and cultural studies courses."