An anonymous reader quotes a report from Ars Technica: Meta plans to test out X's algorithm for Community Notes to crowdsource fact-checks that will appear across Facebook, Instagram, and Threads. In a blog, Meta said the testing in the U.S. would begin March 18, with about 200,000 potential contributors already signed up. Anyone over 18 with a Meta account more than six months old can also join a waitlist of users who will "gradually" and "randomly" be admitted to write and rate cross-platform notes during initial beta testing.

Meta claimed that borrowing X's approach would result in "less biased" fact-checking than relying on experts alone. But the social media company will delay publicly posting any notes until it's confident that the system is working. For users of Meta platforms, notes could help flag misleading content overlooked by prior fact-checking efforts. However, Meta confirmed that users will not be allowed to add notes correcting misleading advertisements, which means notes won't help reduce scam ads that The Guardian reported last August have been spreading on Facebook for years. Meta confirmed that the company plans to tweak X's algorithm over time to develop its own version of community notes, which "may explore different or adjusted algorithms to support how Community Notes are ranked and rated."

Tuesday Meta released Llama 3.1, its largest open-source AI model to date. But just one day Mistral released Large 2, notes this report from TechCrunch, "which it claims to be on par with the latest cutting-edge models from OpenAI and Meta in terms of code generation, mathematics, and reasoning...

"Though Mistral is one of the newer entrants in the artificial intelligence space, it's quickly shipping AI models on or near the cutting edge." In a press release, Mistral says one of its key focus areas during training was to minimize the model's hallucination issues. The company says Large 2 was trained to be more discerning in its responses, acknowledging when it does not know something instead of making something up that seems plausible. The Paris-based AI startup recently raised $640 million in a Series B funding round, led by General Catalyst, at a $6 billion valuation...

However, it's important to note that Mistral's models are, like most others, not open source in the traditional sense — any commercial application of the model needs a paid license. And while it's more open than, say, GPT-4o, few in the world have the expertise and infrastructure to implement such a large model. (That goes double for Llama's 405 billion parameters, of course.)
Mistral only has 123 billion parameters, according to the article. But whichever system prevails, "Open Source AI Is the Path Forward," Mark Zuckerberg wrote this week, predicting that open-source AI will soar to the same popularity as Linux: This year, Llama 3 is competitive with the most advanced models and leading in some areas. Starting next year, we expect future Llama models to become the most advanced in the industry. But even before that, Llama is already leading on openness, modifiability, and cost efficiency... Beyond releasing these models, we're working with a range of companies to grow the broader ecosystem. Amazon, Databricks, and NVIDIA are launching full suites of services to support developers fine-tuning and distilling their own models. Innovators like Groq have built low-latency, low-cost inference serving for all the new models. The models will be available on all major clouds including AWS, Azure, Google, Oracle, and more. Companies like Scale.AI, Dell, Deloitte, and others are ready to help enterprises adopt Llama and train custom models with their own data.
"As the community grows and more companies develop new services, we can collectively make Llama the industry standard and bring the benefits of AI to everyone," Zuckerberg writes. He says that he's heard from developers, CEOs, and government officials that they want to "train, fine-tune, and distill" their own models, protecting their data with a cheap and efficient model — and without being locked into a closed vendor. But they also tell him that want to invest in an ecosystem "that's going to be the standard for the long term." Lots of people see that open source is advancing at a faster rate than closed models, and they want to build their systems on the architecture that will give them the greatest advantage long term...

One of my formative experiences has been building our services constrained by what Apple will let us build on their platforms. Between the way they tax developers, the arbitrary rules they apply, and all the product innovations they block from shipping, it's clear that Meta and many other companies would be freed up to build much better services for people if we could build the best versions of our products and competitors were not able to constrain what we could build. On a philosophical level, this is a major reason why I believe so strongly in building open ecosystems in AI and AR/VR for the next generation of computing...

I believe that open source is necessary for a positive AI future. AI has more potential than any other modern technology to increase human productivity, creativity, and quality of life — and to accelerate economic growth while unlocking progress in medical and scientific research. Open source will ensure that more people around the world have access to the benefits and opportunities of AI, that power isn't concentrated in the hands of a small number of companies, and that the technology can be deployed more evenly and safely across society. There is an ongoing debate about the safety of open source AI models, and my view is that open source AI will be safer than the alternatives. I think governments will conclude it's in their interest to support open source because it will make the world more prosperous and safer... [O]pen source should be significantly safer since the systems are more transparent and can be widely scrutinized...

The bottom line is that open source AI represents the world's best shot at harnessing this technology to create the greatest economic opportunity and security for everyone... I believe the Llama 3.1 release will be an inflection point in the industry where most developers begin to primarily use open source, and I expect that approach to only grow from here. I hope you'll join us on this journey to bring the benefits of AI to everyone in the world.

An anonymous reader quotes a report from TechCrunch: Facebook parent Meta is teaming up with Hugging Face and European cloud infrastructure company Scaleway to launch a new AI-focused startup program at the Station F startup megacampus in Paris. The underlying goal of the program is to promote a more "open and collaborative" approach to AI development across the French technology world. The timing of the announcement is notable, coming amid a growing push for regulation and a marked conflict between the "open" and "closed" AI realms. [...]

While Meta itself has been open sourcing its own generative AI models, Hugging Face -- a billion-dollar VC-backed startup in its own right -- has set out its stall as a sort of open source alternative to OpenAI, replete with open alternatives to the likes of ChatGPT and spearheading community projects such as BigScience. So in many ways, Meta and Hugging Face's tie-up today makes a great deal of sense, given their respective stances on the whole "open" versus "closed" AI discussion. "For me, open source AI is the most important topic of the decade as it is the cornerstone toward democratizing ethical AI," Hugging Face CEO Clement Delangue said in a statement.

From today through December 1 (2023), startups can apply to join the new "AI Startup Program" at Station F, with five winners proceeding to the accelerator program that will run from January to June. The chosen startups, selected by a panel of judges from Meta, Hugging Face and French cloud company Scaleway, will have at least one thing in common -- they will be working on projects substantively built on open foundation models, or at the very least can demonstrate a "willingness to integrate these models into their products and services," according to the announcement issued by Meta today. "With the proliferation of foundation models and generative artificial intelligence models, the aim is to bring the economic and technological benefits of open, state-of-the-art models to the French ecosystem," the announcement noted. Indeed, the winning startups will receive mentoring from researchers and engineers at Meta, gain access to Hugging Face's various platforms and tools, and compute resources from Scaleway.

Facebook is blaming "technical issues" for its broken promise to the U.S. Congress to stop recommending political groups to its users, reports The Markup: Facebook made the pledge once in October, in the run-up to the presidential election, and then falsely reiterated it had taken the step after rioters overtook the U.S. Capitol on Jan. 6, a deadly event partially coordinated by users on the platform.

The Markup first revealed that Facebook was still recommending groups...in an investigation published on Jan. 19. Examining the top 100 groups recommended to roughly 1,900 users on our Citizen Browser panel, we identified 12 as political — including groups with posts calling for violence against lawmakers, spreading election-related conspiracy theories, and coordinating logistics for attending the rally that led to the Capitol riot.

Citizen Browser is a data-driven project examining the choices Facebook makes about what content to amplify.

A week after our report, U.S. senator Ed Markey sent a letter to the company, demanding an explanation, and on Feb. 10, Facebook replied in a letter to Markey. "The issue stemmed from technical issues in the designation and filtering process that allowed some Groups to remain in the recommendation pool when they should not have been," Facebook said in its response. "Since becoming aware of this issue, we have worked quickly to update our processes, and we continue this work to improve our designation and filtering processes to make them as accurate and effective as possible...." Following publication of our story, recommendations for political groups dropped precipitously, as our Citizen Browser panelist data shows...

The "technical issues" meant that, from Election Day on Nov. 3 to the Jan. 6 Capitol Hill riots to President Biden's inauguration on Jan. 20, Facebook was still recommending political groups to its users. Our analysis found that Facebook particularly pushed political groups to more conservative users... Facebook's own internal research has consistently pointed to the danger posed by political groups on its platform. Researchers warned Facebook in a 2016 internal report that 64 percent of new members of extremist groups joined because of the social network's recommendations, according to The Wall Street Journal.

Facebook has dropped the "likes" tally from public pages used by artists, public officials and brands as part of a new redesign. The New York Post reports: Those pages will now only show follower count, which Facebook says is intended to make things simpler. "Unlike Likes, Followers of a Page represent the people who can receive updates from Pages, which helps give public figures a stronger indication of their fan base," the company said in a blog post. Facebook has also redesigned pages to have their own dedicated News Feeds, where they will be able to "discover and join conversations, follow trends, interact with peers and engage with fans."

Facebook is challenging Zoom with the introduction of "Messenger Rooms," a group messaging service that allows Facebook users to host group calls of up to 50 people that anyone can join. Engadget reports: Instead of inviting people individually, Facebook users can post links in their News Feed or in Groups or event pages. And unlike Messenger's existing video chat features, participants don't need to have the Messenger app or even a Facebook account to join a room. When a room is created, anyone can join via their browser, though hosts can opt to "lock" rooms to new guests to prevent party crashers.

There are other Zoom-like features, too. Messenger Rooms will have "immersive 360-degree backgrounds that transport you to beautiful and iconic spaces, from the beach to a luxe apartment on the water," along with "14 new camera filters that offer ambient lighting to brighten your space and your face." Facebook also notes that there are no time limits for video chat sessions, which can be scheduled in advance. Messenger Rooms is starting to roll out now, and will be live in the U.S. "in the coming weeks." Facebook also announced that it's integrating Messenger's video calling into Facebook Dating, so users can participate in "virtual dates" while they can't plan IRL get-togethers. The update will be available "in the coming months."

Facebook has finally revealed the details of its cryptocurrency Libra. From a blog post: Today we're sharing plans for Calibra, a newly formed Facebook subsidiary whose goal is to provide financial services that will let people access and participate in the Libra network. The first product Calibra will introduce is a digital wallet for Libra, a new global currency powered by blockchain technology. The wallet will be available in Messenger, WhatsApp and as a standalone app -- and we expect to launch in 2020. [...] For many people around the world, even basic financial services are still out of reach: almost half of the adults in the world don't have an active bank account and those numbers are worse in developing countries and even worse for women. The cost of that exclusion is high -- approximately 70% of small businesses in developing countries lack access to credit and $25 billion is lost by migrants every year through remittance fees. This is the challenge we're hoping to address with Calibra, a new digital wallet that you'll be able to use to save, send and spend Libra.

From the beginning, Calibra will let you send Libra to almost anyone with a smartphone, as easily and instantly as you might send a text message and at low to no cost. And, in time, we hope to offer additional services for people and businesses, like paying bills with the push of a button, buying a cup of coffee with the scan of a code or riding your local public transit without needing to carry cash or a metro pass. When it launches, Calibra will have strong protections in place to keep your money and your information safe. We'll be using all the same verification and anti-fraud processes that banks and credit cards use, and we'll have automated systems that will proactively monitor activity to detect and prevent fraudulent behavior. We'll also offer dedicated live support to help if you lose your phone or your password -- and if someone fraudulently gains access to your account and you lose some Libra as a result, we'll offer you a refund. Facebook's currency is backed by more than two dozen companies ranging from Visa and Mastercard to Lyft and Spotify. Apple, Google, Amazon and Microsoft have not yet signed up. Banks decided not to join the starting roster because of uncertainties about regulation and concerns over logistical issues that could hamper take-up, Financial Times reported citing several industry executives. From the report: If successful, the project could dramatically reshape some corners of the finance industry, disintermediating payments platforms and stealing business from retail banks and fintech groups, particularly those that specialise in sending payments across borders. Jorn Lambert, executive vice-president for digital solutions at Mastercard, said he was not worried that fee-free transactions would threaten the payment card business. "It's an addition to what we do, not instead of what we do. It is not a zero-sum game. Today, 85 per cent of transactions are made in cash." It is unclear whether Libra will clear the steep hurdles needed to get off the ground, win over regulators, such as the US Securities and Exchange Commission, and be embraced, or strongly resisted, by the financial services industry. Central banks have already questioned the impact of company-created cryptocurrencies on financial stability. "We see hurdles to scale, we see hurdles to adoption, we see enough of this to decide that we would not participate in a scheme like this," said a senior payments executive at a large global bank.

Facebook announced today in a blog post that group administrators can start charging $4.99 to $29.99 a month for exclusive membership in certain groups. "Parenting, cooking, and home cleaning groups will be the first ones to get the new feature as part of an early test," reports The Verge. From the report: As it stands now, free groups will remain intact, but they will soon have the option to launch premium sub-groups. For instance, lifestyle blogger Sarah Mueller's Declutter My Home group is starting an Organize My Home group that costs $14.99 a month to join. And the Grown and Flown Parents group is making a College Admissions group that charges $29.99 for access to college counselors. Facebook says the new feature is so that group admins, who put a lot of time and dedication to growing their communities, can also earn money at the same time. The company also says admins could take the money they earn to create higher-quality content for the group as well, whether that be more posts, videos, or offline meet-ups and events. Facebook reportedly won't be getting a cut of the subscription fees.

An anonymous reader quotes a report from The Verge: Today, Syria announced that it would sign the Paris climate agreement -- a landmark deal that commits almost 200 countries to reducing greenhouse gas emissions to fight global warming. With Nicaragua also joining the deal last month, the United States is now the only country in the world that opposes it. In June, President Donald Trump announced that the U.S. will withdraw from the Paris climate accord, unless it is renegotiated to be "fair" to the United States. But other countries in the deal, such as France, Germany, and Italy, said that's not possible. The Trump administration is also taking steps to roll back regulations passed under former President Barack Obama to achieve the emissions reduction goals set under the Paris deal. The U.S. is the second largest emitter of heat-trapping greenhouse gases in the world after China. "With Syria's decision, the relentless commitment of the global community to deliver on Paris is more evident than ever," Paula Caballero, director of the climate change program at the World Resources Institute, told the Times. "The U.S.'s stark isolation should give Trump reason to reconsider his ill-advised announcement and join the rest of the world in tackling climate change."

The Free Software Foundation will be celebrating its 30th anniversary on Oct. 3rd. Recently, you had a chance to ask its founder Richard Stallman about GNU/Linux, free software, and other issues of public concern. Below you'll find his answers to your questions. Learn more about how you can join the FSF here, and help fight the good fight.

An anonymous reader writes Facebook posted a career application which, in their own words is 'seeking a Linux Kernel Software Engineer to join our Kernel team, with a primary focus on the networking subsystem. Our goal over the next few years is for the Linux kernel network stack to rival or exceed that of FreeBSD.' Two interesting bullet points listing "responsibilities": Improve IPv6 support in the kernel, and eliminate perf and stability issues. FB is one of the worlds largest IPv6 deployments; Investigate and participate in emerging protocols (MPTCP, QUIC, etc) discussions,implementation, experimentation, tooling, etc.

msm1267 writes "Google, Yahoo and Facebook filed amended requests today with the U.S. Foreign Intelligence Surveillance Court reiterating their desire to publish numbers on requests for user data related to national security. Google, meanwhile, went a step further asking for an open, public hearing with the court so that the issue could be publicly debated." Statements from Yahoo's general counsel (filed motion [PDF]) and Facebook's general counsel (filed motion [PDF]). According to Facebook, "In recent weeks, it has become clear that the dialogue with the U.S. government that produced some additional transparency at the outset is at this point unlikely to result in more progress. As a result, today we are joining others in the industry in petitioning the Foreign Intelligence Surveillance Court to require the government to permit companies to disclose more information about the volume and types of national security-related orders they receive."

technirvana sends us to ReadWriteWeb for the scoop on the announcement this morning that representatives from Google and Facebook are joining the DataPortability Workgroup. Quoting: "The group is working on a variety of projects to foster an era in which users can take their data from the websites they use to reuse elsewhere... Good bye customer lock-in, hello to new privacy challenges. If things go right, today could be a very important day in the history of the internet. The non-participation of Google and Facebook, two companies that hold more user data and do more with it than almost any other consumer service on the market, was the biggest stumbling block to the viability of the project. These are two of the most important companies in recent history — what's being decided now is whether they will be walled-garden, data-horders or truly open platforms tied into a larger ecosystem of innovation with respect for user rights and sensible policies about data."