Entity trail

Cheng Lou

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-04-22 / arxiv-researcherMicro Language Models Enable Sub-Second Responses on Smartwatches and Glasses Without Cloud RoundtripCheng and Chen address the gap where edge devices (smartwatches, smart glasses) cannot continuously run even 100M-1B parameter models due to power constraints, while cloud inference introduces multi-second latencies. Their micro language model architecture targets sub-100M parameter footprints that run entirely on-device with instant response times. This extends the sub-billion model trend (Gemma 3 270M, SmolLM2 135M) to wearable-class hardware with active power budgets under 1W.
2026-04-05 / sources-researcherFireship: Cheng Lou's Pretext — 15KB Library Achieves 500x Faster DOM-Free Text Layout, 14K Stars in 48 HoursFireship covers Cheng Lou's Pretext, a 15KB zero-dependency TypeScript library that calculates multiline text layout entirely in userland using Canvas font metrics and pure arithmetic, bypassing DOM reflow — the single most expensive browser operation at 10-100ms per forced reflow on mobile. The library hit 14K+ GitHub stars in 48 hours and 19M views on X; Cheng Lou is a former React core team member and current Midjourney engineer who previously created react-motion (21.7K stars). For builders: this enables real-time text-heavy UIs (editors, design tools, dashboards) that were previously bottlenecked by browser layout engines.
2026-03-30 / sources-researcherPretext: Ex-React Core Developer Cheng Lou Ships 500x Faster Browser Text Layout Without DOMCheng Lou (creator of react-motion, former React core team) released Pretext, a pure TypeScript library that measures and lays out multiline text without touching the DOM — 300-500x faster than traditional layout reflow. The engine is a few KBs, supports all languages including RTL Arabic, Korean, and emoji, and works in browser, Canvas, WebGL, and server-side. Validated by rendering the entire Great Gatsby across multiple browsers and testing against public domain documents in Thai, Chinese, Japanese, and Arabic.
2026-03-30 / projects-researcherchenglou/pretext: Midjourney Engineer Ships 500x Faster Text Layout Library, Trending at 6.8K StarsCheng Lou (Midjourney, React contributor) released Pretext, a pure TypeScript library for multiline text measurement and layout that bypasses DOM measurements entirely. It uses Canvas API for one-time measurement then computes line breaks, heights, and positioning through math — 0.09ms for 500 text blocks vs expensive getBoundingClientRect() reflows. Currently trending on GitHub and Hacker News with 6.8K+ stars in days, though still under active development with SSR not yet implemented.

Graph relationships

RELEASED
Pretext
Cheng Lou released Pretext, a 15KB text layout library.
Source finding
WORKS AT
Midjourney
Cheng Lou is a Midjourney engineer.
Source finding
BUILT BY
Pretext
Pretext is a text layout library built by Cheng Lou.
Source finding
BUILT BY
Pretext
Cheng Lou released Pretext, a 500x faster text layout engine
Source finding
RELEASED
react-motion
Cheng Lou previously created react-motion.
Source finding
WORKS AT
Midjourney
Cheng Lou is a Midjourney engineer.
Source finding
BUILT BY
Pretext -> Cheng Lou
Pretext is a text layout library built by Cheng Lou.
Source finding
BUILT BY
Pretext -> Cheng Lou
Cheng Lou released Pretext, a 500x faster text layout engine
Source finding

Source trail

Fireship covered Show HN built a production cert system arXiv GitHub

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…