Meta

entities · 41 notes linked

Related: Large Language Models · Openai · Google · Microsoft · AI Safety · Yann LeCun · Open Source AI · Generative AI

Notes

• AI critic Gary Marcus: Meta's LeCun is finally coming around to the things I said years ago — Gary Marcus argues deep learning alone cannot achieve general intelligence
• AI for American-Produced Cement and Concrete — Meta's BOxCrete Bayesian optimization model for domestic concrete mix design
• AI is already linked to layoffs in the industry that created it | CNN Business — AI driving tech sector layoffs and workforce skill reshuffling
• Agreement on Transparency and Researcher Access, But Will It Matter? — Bipartisan congressional and industry consensus on PATA researcher access to social media data
• ChatGPT is 'not particularly innovative,' and 'nothing revolutionary', says Meta's chief AI scientist — Yann LeCun argues ChatGPT is solid engineering not scientific breakthrough
• Compared to Other Skilled Work - Engineers are Dramatically Underpaid — Tech generates far more revenue per employee than it pays to engineers
• Deepfakes are now trying to change the course of war | CNN Business — Deepfake videos deployed as wartime disinformation during Ukraine conflict
• Facebook owner reportedly paid Republican firm to push message TikTok is 'the real threat' — Meta hired GOP firm to run covert anti-TikTok media campaign
• From Google To Nvidia, Tech Giants Have Hired Hackers To Break AI Models — AI red teams at major tech companies probing model vulnerabilities
• George Carlin Estate Sues Creators of AI-Generated Comedy Special in Key Lawsuit Over Stars' Likenesses — First estate lawsuit over AI-generated deceased celebrity likeness and voice
• Getting Transparency Right — Legal analysis of privacy, Section 230, and First Amendment trade-offs in platform transparency legislation
• GitHub - tatsu-lab/stanford_alpaca: Code and documentation to train Stanford's Alpaca models, and generate the data. — Instruction-following LLaMA model fine-tuned via self-instruct data
• Google DeepMind boss hits back at Meta AI chief over 'fearmongering' claim — Hassabis vs LeCun debate on AI safety, regulation, and open-source control
• Here are the layoff severance packages Google, Microsoft, Amazon and other tech giants have promised — Comparative severance terms across major 2023 tech-sector mass layoffs
• How A.I. Is Changing the Way the World Builds Computers (Published 2025) — AI is fundamentally rebuilding computing hardware, data centers, and energy systems
• How machine learning powers Facebook's News Feed ranking algorithm — Multi-pass multitask neural ranking system for News Feed personalization at scale
• Inside the lucrative, disturbing world of humans training AI chatbots — Data annotators' pay, conditions, and experiences training AI models at Scale AI and others
• Law Requiring Social Media Transparency Would Break New Ground — Platform Accountability and Transparency Act requiring social media data disclosure to FTC and researchers
• Mark Zuckerberg says AI could soon do the work of Meta's midlevel engineers — Zuckerberg predicts AI replacing midlevel software engineers by 2025
• Meet TinyLlama: A Small AI Model that Aims to Pretrain a 1.1B Llama Model on 3 Trillion Tokens — Small 1.1B LLM pre-trained on 3 trillion tokens
• Meta Superintelligence Labs executives are pushing staff to ditch slow internal systems for faster engineering tools — Meta Superintelligence Labs abandons slow internal infra for Vercel and GitHub
• Meta is playing the AI game with house money — Meta Q2 2025 earnings signal all-in superintelligence bet
• Meta unveils HOT3D dataset for advanced computer vision training — Egocentric 3D hand-object interaction dataset from Meta Reality Labs
• Meta's Yann LeCun predicts 'new paradigm of AI architectures' within 5 years and 'decade of robotics' | TechCrunch — LeCun forecasts LLM paradigm obsolescence and rise of world models
• NeurIPS 2023: Key Takeaways From Invited Talks — NeurIPS 2023 invited talk summaries on LLM efficiency, generative AI, and responsible data
• One of the three 'godfathers of A.I.' feels 'lost' because of the direction the technology has taken | Fortune — AI pioneers Bengio and Hinton warn of existential risk; LeCun dissents
• Open-Source AI Is Uniquely Dangerous — Unsecured open-source AI uniquely risky because safety features cannot be re-patched
• Open-source AI matches top proprietary model in solving tough medical cases — Open-source Llama 3.1 405B matches GPT-4 on clinical diagnostic reasoning
• RAFT: Adapting Language Model to Domain Specific RAG — Fine-tuning recipe training LLMs to ignore distractor documents in domain RAG
• Sharing new breakthroughs and artifacts supporting molecular property prediction, language processing, and neuroscience — Meta FAIR releases OMol25 dataset, UMA model, Adjoint Sampling, and brain language study
• Tech companies like Google and Meta made cuts to DEI programs in 2023 after big promises in prior years — Big tech DEI program rollback after 2020 George Floyd commitments
• Tech workers react to UPS drivers landing a $170,000-a-year package with a mixture of anger and admiration — Tech worker sentiment on UPS Teamsters $170k total-compensation contract win
• Text classification · fastText — Tutorial for training supervised text classifiers with fastText
• The Four Horsemen of the Tech Recession — Four distinct causes behind tech-sector recession amid broader economic strength
• The Pentagon says AI is speeding up its 'kill chain' | TechCrunch — AI accelerating Pentagon kill chain planning amid usage policy tensions
• The Trick to Make LLaMa Fit into Your Pocket: Meet OmniQuant, an AI Method that Bridges the Efficiency and Performance of LLMs — OmniQuant learnable post-training quantization achieves low-bit LLM compression efficiently
• The complex math of counterfactuals could help Spotify pick your next favorite song — Spotify ML model applies twin-network counterfactual reasoning at scale
• The fight over return-to-office is getting dirty — Corporate RTO mandates lack data support and reflect managerial control instinct
• Why LLaMa Is A Big Deal — LLaMA enables GPT-3-class LLM inference on consumer hardware
• Why Tech Job Interviews Became Such a Nightmare — Tech hiring shift toward grueling interviews amid post-pandemic layoffs
• Zuckerberg credits Musk with kicking off tech trend of firing middle managers | Fortune — Zuckerberg credits Musk's Twitter layoffs for inspiring Meta's flattening