← Back to feed

Who we’d like as regrantors

L4 · DeveloperOpen SourceLessWrong AI· 8/14/2026

Builders and developers funding or working on AI safety projects need to understand funding mechanisms and potential grantmakers.

AI Summary

LessWrong AI proposes a list of potential regrantors for funding AI safety projects and outlines their desired qualities (taste, dealflow, hustle).

Excerpt

Following up on Regranting in 2026, here are some people we’d be excited to have as regrantors! Good regrantors have at least one, and preferably more than one, of taste, dealflow, and hustle. We’ve divided them into rough categories with a bit of explanation, though many of these people could fall into a few different categories. If we raise $2m for regranting, we’d love to give each of these people a $50k pot. To be clear, this is not a list of people who’ve agreed to be regrantors! Many have

Read Original
0 upvotes · 0 downvotes · 1 min read

Related Articles

L4 · DeveloperOpen SourceHugging Face Blog
The Open ASR Leaderboard Adds Its First Global South Language

Hugging Face and Voice Arena launched Monsoon benchmark datasets for Hindi and Indian English, adding the first Global South language to the Open ASR Leaderboard with speaker attribute tracking.

L4 · DeveloperOpen SourceHugging Face Blog
Introducing @huggingface/kernels: 200+ WebGPU Kernels for Local AI

Hugging Face releases @huggingface/kernels with 200+ optimized WebGPU kernels and Fleet, a browser-based benchmarking tool for local AI inference performance.

L3 · BuilderOpen SourceSimon Willison's Blog
datasette-mcp 0.2

datasette-mcp 0.2 adds MCP server support to Datasette instances and improves SQL result handling for better LLM compatibility.

L4 · DeveloperOpen SourceSimon Willison's Blog
Introducing wrapture

Graham Dumpleton releases Wrapture, an AI-assisted Python library for tracing and mocking functions with OpenTelemetry support.

L3 · BuilderOpen Source@ClementDelangue
clem 🤗 (@ClementDelangue): raining cats and ducks https://t.co/3A0hZ7iDxk

Rok Benko open-sourced quackd, an LLM-powered brain for Microduck robots that interprets natural language commands and plans multi-step actions.