Bài đăng

New top story on Hacker News: Qwen3.6-35B-A3B on my laptop drew me a better pelican than Claude Opus 4.7

Qwen3.6-35B-A3B on my laptop drew me a better pelican than Claude Opus 4.7 48 by simonw | 4 comments on Hacker News. April 17, 2026 at 12:37AM simonw 48 https://ift.tt/1oY2ZfG Qwen3.6-35B-A3B on my laptop drew me a better pelican than Claude Opus 4.7 4 https://ift.tt/N41vZHX

New top story on Hacker News: Where the DOGE Operatives Are Now

Where the DOGE Operatives Are Now 9 by droidjj | 1 comments on Hacker News. April 17, 2026 at 12:13AM droidjj 9 https://ift.tt/92xgmZT Where the DOGE Operatives Are Now 1 https://ift.tt/zUw9Ykj

New top story on Hacker News: Codex for Almost Everything

Codex for Almost Everything 43 by mikeevans | 9 comments on Hacker News. April 17, 2026 at 12:12AM mikeevans 43 https://ift.tt/bw8Wc1h Codex for Almost Everything 9 https://ift.tt/z8pUNgx

New top story on Hacker News: Japan implements language proficiency requirements for certain visa applicants

Japan implements language proficiency requirements for certain visa applicants 11 by mikhael | 2 comments on Hacker News. April 16, 2026 at 11:53PM mikhael 11 https://ift.tt/DWxeBPC Japan implements language proficiency requirements for certain visa applicants 2 https://ift.tt/1I5PemB

New top story on Hacker News: Show HN: Libretto – Making AI browser automations deterministic

Show HN: Libretto – Making AI browser automations deterministic 12 by muchael | 2 comments on Hacker News. Libretto ( https://libretto.sh ) is a Skill+CLI that makes it easy for your coding agent to generate deterministic browser automations and debug existing ones. Key shift is going from “give an agent a prompt at runtime and hope it figures things out” to: “Use coding agents to generate real scripts you can inspect, run, and debug”. Here’s a demo: https://www.youtube.com/watch?v=0cDpIntmHAM . Docs start at https://ift.tt/i4AwnXd . We spent a year building and maintaining browser automations for EHR and payer portal integrations at our healthcare startup. Building these automations and debugging failed ones was incredibly time-consuming. There’s lots of tools that use runtime AI like Browseruse and Stagehand which we tried, but (1) they’re reliant on custom DOM parsing that's unreliable on older and complicated websites (including all of healthcare). Using a website’s internal...

New top story on Hacker News: Show HN: LangAlpha – what if Claude Code was built for Wall Street?

Show HN: LangAlpha – what if Claude Code was built for Wall Street? 25 by zc2610 | 9 comments on Hacker News. Some technical context on what we ran into building this. MCP tools don't really work for financial data at scale. One tool call for five years of daily prices dumps tens of thousands of tokens into the context window. And data vendors pack dozens of tools into a single MCP server, schemas alone can eat 50k+ tokens before the agent does anything useful. So we auto-generate typed Python modules from the MCP schemas at workspace init and upload them into the sandbox. The agent just imports them like a normal library. Only a one-line summary per server stays in the prompt. We have around 80 tools across our servers and the prompt cost is the same whether a server has 3 tools or 30. This part isn't finance-specific, it works with any MCP server. The other big thing was making research actually persist across sessions. Most agents treat a single deliverable (a PDF, a spre...

New top story on Hacker News: Tracking down a 25% Regression on LLVM RISC-V

Tracking down a 25% Regression on LLVM RISC-V 19 by luu | 2 comments on Hacker News. April 13, 2026 at 02:52AM luu 19 https://ift.tt/8vTxlOr Tracking down a 25% Regression on LLVM RISC-V 2 https://ift.tt/UHPxSJs