Herrington Darkholme

19K posts

Herrington Darkholme

@hd_nvim

🌐 Frontend Vimmer ⚒️Open Source with @typescript @vuejs and @rustlang 🚀 https://t.co/SLjF6No9qD is my hobby project

Katılım Temmuz 2011

2.7K Takip Edilen8.4K Takipçiler

Sabitlenmiş Tweet

Herrington Darkholme@hd_nvim·3 Oca

Announcing the book "𝑀𝑎𝑠𝑡𝑒𝑟𝑖𝑛𝑔 𝑎𝑠𝑡-𝑔𝑟𝑒𝑝"!🦉 The book is a structured deep-dive: we start with the mental models of ASTs, move to rule composition, and end with advanced usage. It’s the place where design principles and philosophy are explained. Link below ↓

English

15.7K

Herrington Darkholme@hd_nvim·3h

@suwakopro 准备做个初音未来 AI 脱口秀

中文

104

Suwako — e/acc@suwakopro·4h

试了一下，真的挺有意思，可惜不是全双工的，效果差一些，对动漫支持的效果也比真人差一点，但考虑到这个还是非常初期，对后面scaling up之后的发展真的很期待。感觉今年底应该就可以有大量的ai直播了

Vidu AI@ViduAI_official

Vidu S1, the World’s Most Advanced Model for Real-Time AI Interaction Vidu S1 introduces a new era of real-time interactive video. Voice-Driven Character Control: Go beyond voice-driven lip sync with real-time behavior generation. Characters understand, react, and respond naturally. Unlimited Real-Time Generation: Experience seamless real-time video generation with unlimited interaction duration. 540P at 25 FPS, built for live interaction: Generate video in real time at 540P / 25 FPS, with support for up to 42 FPS for even smoother experiences. Create Custom Characters from Images & Voices: Customize both appearance and voice. Bring humans, anime characters, pets, and more to life with personalized AI characters. Use code VIDUS1 to get started for free.

中文

333

Herrington Darkholme retweetledi

alphaXiv@askalphaxiv·1d

"Video = World + Event Stream" Wan-Streamer v0.3 treats video as a world plus events. So the world is what stays stable, like the scene, character, voice, and ambience; The events are what change, like speech, motion, gestures, sounds, and reactions. This lets a real-time AI agent not just talk, but act inside a scene with synchronized voice, video, and behavior at low latency.

English

213

10.7K

Herrington Darkholme@hd_nvim·5h

@tcp_handshake don't attack my number, attack my methodology github.com/HerringtonDark…

English

tcp_handshake@tcp_handshake·5h

@hd_nvim even your best fixture landed at 4.94%, so the original 20% was a microbench none of the real inputs reproduce. 2.7% geomean is the number that ships, and the gate is what made the gap visible.

English

Herrington Darkholme@hd_nvim·6h

The gate is decisively positive: all 7 languages improve, with a +2.70% geometric mean across 40 byte-identical fixtures; per-language results range from +1.21% to +4.94%. Peak RSS is effectively neutral/slightly lower. I’m finishing parity and lint gates before committing this

Herrington Darkholme@hd_nvim

oh i think somehow i can improve tree-sitter performance by 20% with ChatGPT

English

1.8K

Herrington Darkholme@hd_nvim·8h

@Escapist_1 E 妹在 LA 吗？

中文

126

Escapist@Escapist_1·17h

恩嘟，也是跟金牌讲师吃上火锅了很喜欢讲师的一句话 “又要到饭了兄弟们” 讲师确实，那么多润人里难得一见站着要饭的 #人类观察

中文

Herrington Darkholme@hd_nvim·13h

@CORALFRISK 生日快乐！

中文

小山内きのこ@CORALFRISK·1d

认真过了生日

中文

1.3K

Herrington Darkholme retweetledi

Herman Jin@ShanghaoJin·14h

@BabybusFL @woooooow_w 砸到短期momentum 集中做空半导体，半导体在里面变成loser 然后拿出业绩一打脸，反向爆仓，又是去杠杆。这将比三月份还波澜壮阔

中文

220

47.3K

Herrington Darkholme@hd_nvim·1d

hey what’s up guys

English

401

Herrington Darkholme@hd_nvim·1d

@njukidreborn There are two versions of perf improvements /goal improve the perf by 20% and me after understanding the nitty-gritty of GLR/tree-sitter The code is messy here, but you can somehow get a feel of it github.com/HerringtonDark…

English

122