Hacker News with Generative AI: Meta

Mark Zuckerberg says social media is over (newyorker.com)
During testimony at Meta’s antitrust trial, the Facebook founder’s argument was, in so many words, that platforms like his are not what they used to be.
Llama 4 Smells Bad (fastml.com)
Meta has distinguished itself positively by releasing three generations of Llama, a semi-open LLM with weights available if you ask nicely (and provide your full legal name, date of birth, and full organization name with all corporate identifiers). So no, it’s not open source. Anyway, on Saturday (!) May the 5th, Cinco de Mayo, Meta released Llama 4.
Commission Finds Apple and Meta in Breach of the Digital Markets Act (europa.eu)
EU fines Apple €500M and Meta €200M (politico.eu)
Apple and Meta fined millions for breaching EU law (yahoo.com)
Apple was fined 500 million euros ($570 million) on Wednesday and Meta 200 million euros, as European Union antitrust regulators handed out the first sanctions under landmark legislation aimed at curbing the power of Big Tech.
At Trial, Instagram Co-Founder Says Meta Denied His Company Resources (nytimes.com)
Kevin Systrom, the co-founder of Instagram, testified on Tuesday in a landmark federal antitrust trial that his startup was starved of resources after Meta bought it because Mark Zuckerberg, the chief executive, was afraid of the success of the photo-sharing app.
Uncovered emails showed how Meta struggled to keep Facebook culturally relevant (techcrunch.com)
With the first week of Meta’s antitrust trial behind us, documents shared by the U.S Federal Trade Commission (FTC) offered more insight into Meta’s internal struggles to keep Facebook relevant.
Meta's Monopoly Made It a Fair-Weather Friend (wired.com)
This week, Mark Zuckerberg took the stand in an antitrust trial that could result in the breakup of Meta’s social networking empire.
Meta MCP: Chaining Tools via Prompt-Driven Arguments (cefboud.com)
This post explores the concept of an MCP tool that can chain multiple tools within a single request, where the arguments for each tool can be dynamically generated using prompts based on the outputs of previous tools.
Meta whistleblower alleges company worked with China on censorship (bbc.com)
A Meta whistleblower told US senators on Wednesday that the company undermined national security in order to build a $18 billion business in China.
How Meta AI Staff Deemed More Than 7M Books to Have No "Economic Value" (vanityfair.com)
As more than a dozen lawsuits churn ahead, newly unsealed case files reveal the company’s stance: The pirated books Meta used to train its AI, including ones by Beverly Cleary, Jacqueline Woodson, and Andrew Sean Greer, are individually worthless.
Zuckerberg defends Meta under oath in federal court (npr.org)
Meta whistleblower alleges company worked with China on censorship (bbc.com)
A Meta whistleblower told US senators on Wednesday that the company undermined national security in order to build a $18 billion business in China.
Show HN: Open-Source Meta Raybans with 6x of their battery (github.com/BasedHardware)
<h1>omiGlass - Open Source Meta Raybans with 6x of their battery</h1>
U.S. Prepares to Challenge Meta's Social Media Dominance (nytimes.com)
On Monday, Meta will face off against the federal government in a landmark antitrust trial over claims that it illegally squashed competition by buying Instagram and WhatsApp.
The biggest trial in Meta's history starts Monday. Here's what to know (npr.org)
The Federal Trade Commission's blockbuster antitrust case against Meta kicks off on Monday in a courtroom in Washington. It's the culmination of a nearly six-year investigation into whether the social media giant broke competition laws in acquiring Instagram and WhatsApp.
How Meta has blown $45B failing at the metaverse (yahoo.com)
Nearly $50 billion.
Meta whistleblower alleges company worked with China on censorship (bbc.com)
A Meta whistleblower told US senators on Wednesday that the company undermined national security in order to build a $18 billion business in China.
Whistleblower tells senators that Meta undermined U.S. security, interests (thehill.com)
Sarah Wynn-Williams, the former director of global public policy at Facebook, which is now called Meta, testified at a Senate hearing Wednesday that she saw Meta executives “repeatedly undermine U.S. national security and betray American values” during her seven-year stint in that job.
Whistleblower testifies to Congress regarding Meta AI aid to China (arstechnica.com)
Later today, a former Facebook employee, Sarah Wynn-Williams, will testify to Congress that Meta executives "repeatedly" sought to "undermine US national security and betray American values" in "secret" efforts to "win favor with Beijing and build an $18 billion dollar business in China."
Meta Whistleblower to Tell Congress That Company Aided China in AI Race (bloomberg.com)
Sarah Wynn-Williams, a Meta Platforms Inc. executive turned whistleblower, plans to testify before Congress that the social media giant threatened US interests while cozying up to China — claims the company said are false.
Meta got caught gaming AI benchmarks (theverge.com)
With Llama 4, Meta fudged benchmarks to appear as though its new AI model is better than the competition.
Meta got caught gaming LMArena (theverge.com)
With Llama 4, Meta fudged benchmarks to appear as though its new AI model is better than the competition.
Meta's surprise Llama 4 drop exposes the gap between AI ambition and reality (arstechnica.com)
On Saturday, Meta released its newest Llama 4 multimodal AI models in a surprise weekend move that caught some AI experts off guard.
Serious issues in Llama 4 training. VP of AI at Meta resigned (reddit.com)
Despite repeated training efforts, the internal model's performance still falls short of open-source SOTA benchmarks, lagging significantly behind.
Llama 4: Did Meta just push the panic button? (interconnects.ai)
Where Llama 2’s and Llama 3’s releases were arguably some of the top few events in AI for their respective release years, Llama 4 feels entirely lost.
Meta debuts Friends tab, Mark Zuckerberg pushes 'throwback to OG Facebook' (cnbc.com)
Meta pirated at least 101 of my books, and others (garymarcus.substack.com)
Meta pirated at least 101 of my books and articles, and tens of millions of others
Ex-Meta executive: 'People deserve to know what this company is like' (cnn.com)
Meta pirated books to train its AI (theatlantic.com)
Meta pirated millions of books to train its AI. Search through them here.