与「LETS」相关的搜索结果

SOMEONE BUILT A WRAPPER THAT LETS YOU USE CLAUDE CODE FOR FREE. NO API KEY. NO SUBSCRIPTION. NO $20/MONTH. IT JUST WORKS. 43.9k stars. 7.3k forks. 871 commits. i opened the repo expecting a catch. a hidden cost. a "free tier" limit. there is none. it grabs the session from the web interface and pipes it to your terminal, routing through 31 cloud and local providers. works in vscode. works in jetbrains. works on your phone. voice input included. you talk, it codes. the $20/month barrier didn't get lower. it got deleted for anyone willing to run a python script. i've been rationing tokens like water in a desert. turns out the tap was never locked. free. mit license. 31 providers. local terminal. voice. zero permission needed.

RAG system that skips HTML parsing entirely! PixelRAG is an open-source visual RAG framework that renders documents as screenshots instead of parsing them into text. Most RAG pipelines start by converting HTML to text. Tables flatten into unstructured rows. Charts disappear. Layout context is gone before the LLM ever sees it. The paper measured this directly: HTML-to-text conversion accounts for 36.6% of retrieval failures on SimpleQA. PixelRAG skips that step entirely. It renders pages as screenshot tiles using Playwright, embeds those tiles with a fine-tuned Qwen3-VL-Embedding model, builds a FAISS index, and passes retrieved images directly to a VLM reader. No text abstraction in between. Benchmarked across six datasets against the strongest text-based baselines: - SimpleQA: 78.8% vs 71.6% (+7.1 points) - NQ-Tables: 48.8% vs 42.5% (+6.3 points) - EVQA: +15.5 points - LiveVQA: +11.3 points One honest caveat from the authors: this requires Qwen3-VL-4B class models or larger to see the benefit. Smaller models trail text retrieval. The authors also recommend using PixelRAG as an enhancement layer alongside existing text systems rather than a full replacement. Ships with a pre-built Wikipedia index covering 8.28M articles across 28.1M screenshot tiles. A Claude Code plugin lets Claude take screenshots of any URL and reason over the visual content directly. Key capabilities: • Renders web pages, PDFs, and images as screenshot tiles via Playwright • Fine-tuned Qwen3-VL-Embedding model for visual retrieval • FAISS index for fast vector search • Pre-built Wikipedia index: 8.28M articles, 28.1M tiles • 3x token cost reduction via image compression • Claude Code plugin for direct URL screenshot and visual reasoning • LoRA fine-tuning support via pixelrag-train 100% open source. I've shared the link in the replies!

Grok Build keeps getting smarter.....this update expands Auto mode, gives developers more control over Bash permissions, speeds up /btw, and makes long conversations easier to navigate Release Notes: v0.2.119 Features: • Always allow for bash commands now lets you edit a free-form glob pattern instead of only word-prefix scopes. • Long responses now show a clickable arrow that jumps back to the start of the answer. • Auto mode now auto-approves more common read-only git commands and harmless file appends. • Plan previews now show Mermaid diagram buttons (Open Image, Copy Image Path, Copy Source). Bug Fixes: • Gateway connections now detect and recover from dead sockets more reliably. • Question cards now let you Tab through answers instead of losing focus to the scrollback. • Resume picker no longer tries to load a session from pasted garbage when you press Enter. • Background task completion messages no longer grow unbounded when the task produced a huge log. • Plan viewer scrollbar now responds to clicks on the border column and renders without dark stripes in • Expired external auth provider credentials now correctly trigger the interactive sign-in flow instead of a silent 401 loop. Performance: • /btw side questions now reuse the parent session’s cached prefix for faster responses. • Doctor and tmux-backed startup are now faster when no live tmux processes remain.

$1.6 million dollars in Bitcoin was drained from my account on July 29th in the Cold Card wallet hack. My Bitcoin was in cold storage. My keys were on a ColdCard device kept in a safety deposit box that had never been connected to the internet. This part's nerdy, but here's what happened: Hackers discovered a vulnerability in the part of the hardware wallet code used to create seed phrases. This allowed them to use AI to brute force guessing seed phrases. I was at our cottage and heard about the hack today. "No way this affects me." I thought. I logged into Wasabi––software that lets me view my bitcoin wallets online. Right away I saw lines of red transaction–withdrawals–and I knew. From 9:36pm - 9:43pm on July 29th, every wallet I had had been emptied. 18.25245043 btc gone. That's just over $1.6 million dollars CAD. Perhaps the hardest part about this is that I did everything right. I never shared my seed phrase with anybody. My devices never touched the internet. Everything was kept in multiple safes and safety deposit boxes. None of it mattered. All because the hardware that created the seed phrase originally had one line in their code from 2021 that had a vulnerability. I'm filing a police report and a report with the Ontario Securities Commission. But I don't expect to recoup anything. A part of me is trying to make sense of what just happened. Or try to figure out a lesson in it. I'm struggling. $1.6 million is a staggering amount of money to have stolen. I guess all that I can think about right now is that I'm so damn happy that I'm an entrepreneur and that my earning potential is under my control. Mark my damn words. I'll recover.

Google Earth now lets you create AI images using real-world locations. With the new Nano Banana AI feature, you can pick any place in Google Earth, click “Create Image,” and describe what you want to see. The AI can turn that location into a historical scene, a futuristic city or a fantasy world The feature uses Google Earth’s satellite and 3D map data as the base, so the AI-generated images are built from real locations.

Software quality now depends on the constraints you set around your agents. When humans manually wrote most of the code we could look at the code itself for signs of quality. Is it clean? Is it thoughtful? Is it fast? Can another engineer understand it? Does it have tests? Agents can now generate more code than people can read. When code generation scales beyond review, quality - checks for one or more of correctness, maintainability, security, performance etc - increasingly has to live somewhere else. It moves into the harness, environment and operating system around the agent. This can be the tests and deterministic checks that decide what the system is allowed to do (amongst others). Your constraints are what may eventually enable loops of agents to deliver production software reliably. They can include unit tests, property tests, acceptance tests, mutation testing and quality metrics. This back-pressure lets the system resist bad work before it becomes somebody elses problem. Set your constraints. They decide whether the code your agents generate is good enough to ship.

One of the most powerful ways to use Pi is through the SDK, which lets you embed Pi into your apps. You can build custom interfaces or integrate with automated workflows. Have you tried the Pi SDK? What have you built with it? If not, here’s the code to get started💻