hackernews

1. The privacy nightmare of browser fingerprinting Total comment counts : 38 Summary Many de-Googling efforts aim at privacy, but browser fingerprinting makes tracking harder to block than cookies. Third‑party cookies were the historical privacy villain and are now mitigated by browsers and laws, but fingerprinting operates without cookies and can survive VPNs. It builds a unique identifier from data like browser version, OS, language, time zone, and harder signals such as fonts, extensions, hardware, and canvas-derived pixel data—canvas fingerprints are highly unique, with roughly 1 in 1000 browsers sharing the same fingerprint....

1. Show HN: Wealthfolio 2.0- Open source investment tracker. Now Mobile and Docker Total comment counts : 37 Summary An open-source, privacy-first portfolio tracker that runs on-device with no subscriptions and an optional one-time payment. It consolidates all investments (stocks, ETFs, crypto, savings) in one place and supports CSV imports from brokers/banks. It provides clear charts, asset allocation, and performance views, with benchmarks against the S&P 500 and popular ETFs. It tracks dividends, interest, and passive income; monitors fees; and shows past performance trends....

1. Nano Banana Pro Total comment counts : 67 Summary Google DeepMind unveils Nano Banana Pro, a leading image generation and editing model built on Gemini 3 Pro. It produces accurate visuals with legible text in multiple languages and up to 4K resolution. Using Gemini 3 Pro’s advanced reasoning and real-world knowledge, Nano Banana Pro can visualize ideas, generate infographics and diagrams from notes, and incorporate real-time information via Google Search....

1. The Death of Arduino? Total comment counts : 28 Summary Qualcomm-owned Arduino quietly updated its Terms of Service and Privacy Policy, marking a shift from open-hardware to a tightly controlled, data-extractive corporate service. Key changes grant users an irrevocable, perpetual license to anything uploaded, enable broad surveillance of AI features, prohibit reverse-engineering without permission, retain usernames for years after deletion, and integrate all user data (including minors) into Qualcomm’s global data ecosystem....

1. Gemini 3 Total comment counts : 133 Summary Google unveils Gemini 3, its most intelligent AI model, with enhanced reasoning and multimodal capabilities to turn ideas into reality. It’s available now across Google products—including the Gemini app, AI Studio, and Vertex AI—and will add Gemini 3 Deep Think for Ultra subscribers soon. Gemini 3 Pro, in preview, outperforms 2.5 Pro in reasoning, multimodality, and coding. The rollout spans AI in Search, the Gemini app, AI Studio, Vertex AI, and Google’s Antigravity platform, marking two years of rapid Gemini progress and broad-scale AI adoption....

1. Compiling Ruby to Machine Language Total comment counts : 3 Summary An excerpt from Ruby Under a Microscope explains YJIT/ZJIT in Ruby 3.x. YJIT measures how often a Ruby block or function runs, using jit_entry and jit_entry_calls; when calls exceed a threshold (default 30 for small apps, 120 for large ones), YJIT compiles the YARV instructions into machine code. The compiled code is stored in YJIT blocks, which cover sections of YARV instructions rather than whole functions, enabling optimization and avoiding unnecessary compilation....

1. Open-source Zig book Total comment counts : 14 Summary A 61-chapter, project-based guide that aims to fundamentally change how you think about software, shifting focus from syntax to a broader philosophy. It emphasizes zero AI and is authored by @zigbook. Overall Comments Summary Main point: The Zigbook is discussed as a promising, high-quality resource for learning Zig and low-level systems programming, while questions linger about its authenticity and accessibility. Concern: The main worry is whether the content is genuinely hand-written or AI-generated, and whether readers can access and navigate the material easily (PDF availability, broken/non-working links, confusing structure)....

1. Our investigation into the suspicious pressure on Archive.today Total comment counts : 42 Summary The FBI is reportedly investigating Archive.is (Archive.today) and has subpoenaed its domain registrar to identify its operator, amid concerns that touch on copyright or CSAM dissemination. A French group, Web Abuse Association Defense, urged blocking archive.today in AdGuard DNS for allegedly refusing to remove illegal content since 2023. Archive.today says it removes illegal content when notified and has received no prior alerts; the group’s claims and identity appear dubious, with “bailiff reports” dated mainly August 2025, not 2023....

1. AI World Clocks Total comment counts : 63 Summary Every minute, a new clock is generated by nine AI models, each limited to 2000 tokens. The project, titled “Generating AI Clocks,” credits Brian Moore (also on Instagram) and cites Matthew Rayfield as the inspiration. Overall Comments Summary Main point: The discussion compares how different LLMs (e.g., Kimi K2, Qwen 2.5, Deepseek, Haiku, Gemini, GPT, Claude) generate clock faces, with Kimi K2 often most consistent and others showing wide variability....

1. Nano Banana can be prompt engineered for nuanced AI image generation Total comment counts : 27 Summary The AI image field is rapidly evolving, with FLUX.1-dev and rivals like Seedream, Ideogram, Qwen-Image, Imagen 4, and ChatGPT’s free image support reshaping 2025. ChatGPT’s gpt-image-1 is autoregressive but slow. In Aug 2025, Google released Nano Banana (Gemini 2.5 Flash Image), an autoregressive model generating 1,290 tokens per image and popularly called Nano Banana....