第74期 | karpathy/build-nanogpt
今日摘要
GitHub anthropics：A suite of plugins for legal workflows
GitHub openai：openai/openai-builder-lab recently updated repository.
GitHub karpathy：LLM training in simple, raw C/CUDA
OpenAI Blog：See how finance teams can use Codex to build MBRs, reporting packs, variance bridges, model checks, and planning scenarios from re…
GitHub anthropics：anthropics/financial-services recently updated repository.
总结 + 观点：Lightweight Android capture developer tools for…｜中文观点：openai/snap-o 的核心不在新鲜感，而在它是否能提升工程效率、部署稳定性或开发者…
总结 + 观点：VQVAEs, GumbelSoftmaxes and friends｜中文观点：karpathy/deep-vector-quantization 更值得从实际采用价值来…
总结 + 观点：Transformers: State-of-the-art Machine Learning…｜中文观点：karpathy/transformers 更值得从实际采用价值来判断，而不是只看它有没有…
总结 + 观点：Video+code lecture on building nanoGPT from scra…｜中文观点：karpathy/build-nanogpt 更值得从实际采用价值来判断，而不是只看它有没…
总结 + 观点：Teams use Codex with GPT-5.5 to ship production…｜中文观点：更值得关注的是 How NVIDIA engineers and researchers…
anthropics/claude-for-legal
标签：#github_orgs #extended
作者：
原文：A suite of plugins for legal workflows
openai/openai-builder-lab
标签：#github_orgs #extended
作者：
原文：openai/openai-builder-lab recently updated repository.
karpathy/llm.c
标签：#github_orgs #extended
作者：
原文：LLM training in simple, raw C/CUDA
How finance teams use Codex
标签：#ai_engineering_blogs #core
作者：
原文：See how finance teams can use Codex to build MBRs, reporting packs, variance bridges, model checks, and planning scenarios from real work inputs.
anthropics/financial-services
标签：#github_orgs #extended
作者：
原文：anthropics/financial-services recently updated repository.
openai/snap-o
标签：#github_orgs #extended
作者：
原文：Lightweight Android capture developer tools for macOS
karpathy/deep-vector-quantization
标签：#github_orgs #extended
作者：
原文：VQVAEs, GumbelSoftmaxes and friends
karpathy/transformers
标签：#github_orgs #extended
作者：
原文：Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.
karpathy/build-nanogpt
标签：#github_orgs #extended
作者：
原文：Video+code lecture on building nanoGPT from scratch
How NVIDIA engineers and researchers build with Codex
标签：#ai_engineering_blogs #core
作者：
原文：Teams use Codex with GPT-5.5 to ship production systems and turn research ideas into runnable experiments.
What Parameter Golf taught us about AI-assisted research
标签：#ai_engineering_blogs #core
作者：
原文：Parameter Golf brought together 1,000+ participants and 2,000+ submissions to explore AI-assisted machine learning research, coding agents, quantization, and novel model design under strict constraints.
AutoScout24 scales engineering with AI-powered workflows
标签：#ai_engineering_blogs #core
作者：
原文：Learn how AutoScout24 Group uses Codex and ChatGPT to speed development cycles, improve code quality, and expand AI adoption.
This works really well btw, at the end of your query ask your LLM to "structure your response as HTML", then view the generated file in your...
标签：#x_profiles #extended
作者：
原文：This works really well btw, at the end of your query ask your LLM to "structure your response as HTML", then view the generated file in your browser. I've also had some success asking the LLM to present its output as slideshows, etc. More generally, imo audio is the human-preferred input to AIs but vision (images/animations/video) is the preferred output from them. Around a ~third of our brains are a massively parallel processor dedicated to vision, it is the 10-lane superhighway of information into brain. As AI improves, I think we'll see a progression that takes advantage: 1) raw text (hard/effortful to read) 2) markdown (bold, italic, headings, tables, a bit easier on the eyes) nitter.net/zan2434/status/2046982… There are also improvements necessary and pending at the input. Audio nor text nor video alone are not enough, e.g. I feel a need to point/gesture to things on the screen, similar to all the things you would do with a person physically next to you and your computer screen. TLDR The input/output mind meld between humans and AIs is ongoing and there is a lot of work to do and significant progress to be made, way before jumping all the way into neuralink-esque BCIs and all that. For what's worth exploring at the current stage, hot tip try ask for HTML. Thariq (@trq212) x.com/i/article/205279610060… https://nitter.net/trq212/status/2052809885763747935#m
How ChatGPT adoption broadened in early 2026
标签：#ai_engineering_blogs #core
作者：
原文：ChatGPT adoption surged in Q1 2026, with fastest growth among users over 35 and more balanced gender usage, signaling broader mainstream AI adoption.
How enterprises are scaling AI
标签：#ai_engineering_blogs #core
作者：
原文：How enterprises scale AI: from early experiments to compounding impact through trust, governance, workflow design, and quality at scale.
链接：https://openai.com/business/guides-and-resources/how-enterprises-are-scaling-ai
huggingface/smol-course
标签：#github_orgs #extended
作者：
原文：A course on aligning smol models.
langchain-ai/docs
标签：#github_orgs #extended
作者：
原文：Unified LangChain documentation.
langchain-ai/chat-langchain
标签：#github_orgs #extended
作者：
原文：langchain-ai/chat-langchain recently updated repository.
I got tired of realizing "discounts" weren't discounts
标签：#research_community #extended
作者：
原文：Article URL: https://apps.apple.com/us/app/silicon-ai-price-comparison/id6764054291 Comments URL: https://news.ycombinator.com/item?id=48117483 Points: 1 Comments: 0
链接：https://apps.apple.com/us/app/silicon-ai-price-comparison/id6764054291
Show HN: SelfCertForge, manage root CAs and self-signed certs on macOS/Windows
标签：#research_community #extended
作者：
原文：Hi, HN! I recently published SelfCertForge, an open source desktop app for creating and managing self-signed certificates on macOS and Windows: https://github.com/rbonestell/SelfCertForge/ I know, I know... this isn't an earth-shattering invention. This came from a place of years of personal annoyance. I’ve been maintaining scripts that wrapped OpenSSL for local certificate workflows: create a long-lived root cert, add it to the trust store, then generate and sign child certificates for local services. The scripts work but the workflow was clunky. Used infrequently enough that it's easy to forget the flags, not much easier to explain to anyone than OpenSSL itself, and irritating to repeat across machines. Every time I'd share them with colleagues and friends there were always questions and feature requests, so I used Claude Code to amplify my UI/UX design skills (well, I can't amplify 0...) and turned the functionality into a cross-platform GUI. SelfCertForge can generate root CAs and add them to the system's trust store, generate and sign child certificates, and export DER, PEM, PFX, and P7B formats. It also supports common X.509 fields like Subject, SANs, and Key Usage. Very important caveat: this is not suitable for production environments or public-facing SSL/TLS. Comments URL: https://news.ycombinator.com/item?id=48117476 Points: 1 Comments: 0