frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: Send Claude Code tasks to the Batch API at 50% off

https://github.com/s2-streamstore/claude-batch-toolkit
2•misker1•1h ago
Hey HN. I built this because my Anthropic API bills were getting out of hand (spoiler: they remain high even with this, batch is not a magic bullet).

I use Claude Code daily for software design and infra work (terraform, code reviews, docs). Many Terminal tabs, many questions. I realised some questions are ok to wait on and with that comes some cost savings. So here is a small MCP that lets you send work directly to Anthropic's Batch API from inside Claude Code, for the same quality responses just 50% cheaper, results come back in ~30min-1hr.

  How it works: you type /batch review this codebase for security issues, Claude gathers all the context, builds a self-contained prompt, ships it to the Batch API via an MCP server, and you get notified in the status bar when it's done (optional). 

  The README has installation instructions, which were mainly generated by claude. I removed the curl | bash setup and at this stage of the project i feel more confident sharing the manual setup instructions. 

 My main hope with this project is to monetize it. Not by asking for money, rather I am hoping others have ideas or improvements  to add and use those to save more on cost.

The box the AI can't think about

https://www.researchgate.net/publication/400670258_Self-Referential_Quantum_Barriers_for_AGI_Cont...
1•JakubCwi•48s ago•1 comments

Frequent Diet Soda Intake Linked to Fourfold Increased Dementia Risk

https://www.medscape.com/viewarticle/frequent-diet-soda-intake-linked-fourfold-increased-dementia...
2•wjb3•7m ago•1 comments

GPT-5 outperforms federal judges 100% to 52% in legal reasoning experiment

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6155012
2•droidjj•7m ago•0 comments

LLM "reasoning" continues to be deeply flawed

https://garymarcus.substack.com/p/breaking-llm-reasoning-continues
3•petethomas•7m ago•0 comments

Show HN: Undertale Text Box Generator

https://undertaletextboxgenerator.online
1•ethanYIAI•8m ago•0 comments

Rare, dangerous side effects of some Covid-19 vaccines explained

https://www.science.org/content/article/rare-dangerous-side-effects-some-covid-19-vaccines-explained
2•epistasis•9m ago•0 comments

But What to Do About It?

https://andys.blog/what-to-do/
1•andytratt•11m ago•0 comments

One with the Environment

https://andys.blog/environment/
1•andytratt•11m ago•0 comments

7 Days in the NICU

https://www.sonyasupposedly.com/nicu-diary/
1•exolymph•11m ago•0 comments

Claude Code Skill That Shares Noteworthy Moments to Slack

https://quickchat.ai/post/claude-code-skill-watches-coding-sessions-shares-to-slack
1•piotrgrudzien•11m ago•0 comments

Few people inevitably rise to the top

https://andytrattner.com/
1•andytratt•12m ago•0 comments

Memory Layer for AI Agent

https://github.com/obicho/mem
1•obicho•14m ago•1 comments

Algorithmic Liability: How Social Media Platforms Disclaim Responsibility

https://tostracker.app/analysis/algorithmic-liability
1•tldrthelaw•14m ago•0 comments

How The New York Times uses a custom AI tool to track the "manosphere"

https://www.niemanlab.org/2026/02/how-the-new-york-times-uses-a-custom-ai-tool-to-track-the-manos...
3•thetrustworthy•16m ago•0 comments

We built Vortex to fix team activation in PLG products

https://www.vortexsoftware.com/
1•santiaparicio•17m ago•1 comments

Death of Software. Nah

https://hardcoresoftware.learningbyshipping.com/p/238-death-of-software-nah
2•whobre•18m ago•0 comments

Show HN: Ignlnk, Hide the content to agents, not the files

https://github.com/No3371/ignlnk
1•3371•20m ago•0 comments

How the hardware/software interface works (Linux Kernel)

https://offlinemark.com/how-the-hardware-software-interface-works/
2•medbar•23m ago•0 comments

Most Americans don’t pay for news and don’t think they need to

https://www.niemanlab.org/2026/02/most-americans-dont-pay-for-news-and-dont-think-they-need-to/
2•jaredwiener•24m ago•0 comments

Leo thought Guthrie's smart cam was disconnected, Google still had the tape

https://fortune.com/2026/02/11/privacy-concerns-nancy-guthrie-google-nest-camera-footage-kidnapping/
1•c420•26m ago•1 comments

The curious case of binfmt for x86 emulation for ARM Docker

https://gergely.imreh.net/blog/2025/04/the-curious-case-of-binfmt-for-x86-emulation-for-arm-docker/
1•pvtmert•34m ago•0 comments

A "QuitGPT" campaign is urging people to cancel their ChatGPT subscriptions

https://www.technologyreview.com/2026/02/10/1132577/a-quitgpt-campaign-is-urging-people-to-cancel...
1•amai•34m ago•1 comments

With Boran, First AI Agent Protagonist in Next-Generation Video Games

https://store.steampowered.com/app/2451190/Kebabstar/
1•akbr•34m ago•1 comments

Prime Intellect Lab: a full-stack platform for training your own models

https://www.primeintellect.ai/blog/lab
1•anacleto•36m ago•0 comments

Self-contained RTL to GDS flow for simple chip designs

https://github.com/htfab/microlane
2•ua709•36m ago•0 comments

Language a Wood for Thought: Susan Howe's Work

https://www.poetryfoundation.org/articles/1769037/language-a-wood-for-thought
1•apollinaire•37m ago•0 comments

Show HN: Terminal Anywhere – Browser terminal via Cloudflare tunnels

1•alamas•37m ago•1 comments

Measles Is Raging Worldwide

https://www.nature.com/articles/d41586-026-00367-5
3•anigbrowl•40m ago•2 comments

Show HN: Japanese ma is the most underrated design system in front end

https://opale-ui-mugen.vercel.app/
1•Dontizi•40m ago•0 comments

Seedance 2.0 makes it hard for agencies and production houses to ignore AI now

https://twitter.com/heydin_ai/status/2021529833197773204
1•DustinEchoes•41m ago•1 comments