Hacker News
Hacker News
On activation, Pyrefly silently writes disableLanguageServices = true to the user's global VS Code settings for three named third-party extensions, without prompting the user or notifying them in a...
Praveen Neppalli Naga, Uber's chief technology officer, revealed the plan in an interview at TechCrunch's StrictlyVC event in San Francisco on Thursday night, describing it as a natural extension of a nascent program the company announced in late January called AV Labs.
Zugzwang is a situation found in chess and other turn-based games wherein one player is put at a disadvantage because of their obligation to make a move; a player is said to be "in zugzwang" when any legal move will worsen their position.
A required part of this site couldn’t load. This may be due to a browser extension, network issues, or browser settings. Please check your connection, disable any ad blockers, or try using a different browser.
As artificial intelligence (AI) tools become widely adopted, large language models (LLMs) are increasingly involved on both sides of decision-making processes, ranging from hiring to content moderation. This dual adoption raises a critical question: do LLMs systematically favor content that resembles their own outputs? Prior research in computer science has identified self-preference bias -- the tendency of LLMs to favor their own generated content -- but its real-world implications have not been empirically evaluated. We focus on the hiring context, where job applicants often rely on LLMs to refine resumes, while employers deploy them to screen those same resumes. Using a large-scale controlled resume correspondence experiment, we find that LLMs consistently prefer resumes generated by themselves over those written by humans or produced by alternative models, even when content quality is controlled. The bias against human-written resumes is particularly substantial, with self-preference bias ranging from 67% to 82% across major...
Conversational large language models are fine-tuned for both instruction-following and safety, resulting in models that obey benign requests but refuse harmful ones. While this refusal behavior is widespread across chat models, its underlying mechanisms remain poorly understood. In this work, we show that refusal is mediated by a one-dimensional subspace, across 13 popular open-source chat models up to 72B parameters in size. Specifically, for each model, we find a single direction such that erasing this direction from the model's residual stream activations prevents it from refusing harmful instructions, while adding this direction elicits refusal on even harmless instructions. Leveraging this insight, we propose a novel white-box jailbreak method that surgically disables refusal with minimal effect on other capabilities. Finally, we mechanistically analyze how adversarial suffixes suppress propagation of the refusal-mediating direction. Our findings underscore the brittleness of current safety fine-tuning methods. More broadly, our work showcases how an understanding of...
An unknown Sega Saturn project has come to light after 29 years
🎨 Local-first, open-source alternative to Anthropic's Claude Design. ⚡ 19 Skills · ✨ 71 brand-grade Design Systems 🖼 Generate web · desktop · mobile prototypes · slides · images · videos · Hype...
MLJAR Studio is a private AI data lab for exploring data, running machine learning experiments, and building analysis tools. Runs locally with optional AI providers.
Latest Geekbench performance figures for macOS VMs, and testing of how few cores and how little memory is really needed: could you run a macOS VM usefully on a MacBook Neo?
AI that helps users fill PDF forms step by step.
Search 250 years of American newspapers with AI-powered semantic search. Explore millions of historical articles spanning the 1730s to the 1960s.
Parallel evolution.
About Us CollectWise is a fast growing and well funded Y Combinator-backed startup. We’re using generative AI to automate debt collection, a $35B market in the US alone. Our AI agents are already outperforming human collectors by 2X, and we’re doing so at a fraction of the cost. With a team of five, we scaled to a $2 million annualized run rate in just a few months, and we are now hiring a Senior Forward Deployed Engineer to help us reach $10 million within the next year. Role We are hiring a Senior Forward Deployed Engineer to lead customer implementations end-to-end. You’ll work at the intersection of engineering and operations, building the integrations and technical systems that bring clients into production. This role blends software engineering with hands-on customer work. You’ll thrive here if you enjoy ambiguity, love shipping quickly, and can translate business requirements into robust, testable systems. This...
As IAC continues to sharpen its focus, we have made the decision to discontinue our search business, which includes Ask.com. After 25 years of answering the world's questions, Ask.com officially closed on May 1, 2026.
The TI-84 Evo is the most advanced graphing calculator in the TI-84 series. Ideal for algebra, geometry, calculus, and advanced math courses in secondary schools. Learn more.