SIGNAL

Wednesday, August 19, 2026

18 stories · 5 min read

★ Must ReadAn LLM wiki changed how I work

And everything else I learned about productivity this year

Platformer· Casey Newton· 2 min· 6d agoSource →

01QUICK HITS

01

Field measurements of neighborhood-scale air temperature impacts of data centers

Trending on Hacker News with 287 points and 425 comments.

Hacker News · 1 min

→02

Claude Code May–August 2026 weekly limits promotion

Trending on Hacker News with 263 points and 238 comments.

Hacker News · 1 min

→03

Norway should buy OpenAI

Trending on Hacker News with 222 points and 235 comments.

Hacker News · 1 min

→04

Universal health coverage could save $1T and 114k lives a year: study

Trending on Hacker News with 746 points and 813 comments.

Hacker News · 1 min

→05

Anthropic Risk Report: August 2026

I am grateful that Anthropic is producing periodic Risk Reports.

Zvi Mowshowitz · 2 min

→

03THE TENSION2▸

PERSPECTIVE A

“Google’s biggest mistake?”

— Gary Marcus

A theory about what went wrong, and some open questions about what might happen next

Source →

♦vs♦

PERSPECTIVE B

“Using the railway network as a flatbed scanner”

— Hacker News

Trending on Hacker News with 411 points and 66 comments.

Source →

04RESEARCH3▸

★ Must ReadBreaking: U.S. young adults are now more concerned about AI than enthusiastic

Brief update to last’s week’s note re: the growing backlash.

Gary Marcus· 3 min· 6d agoSource →

The hyping of Anthropic’s IPO

Gary Marcus — The hyping of Anthropic’s IPO

Gary Marcus· 3 min· 8d agoSource →

Import AI 469: Science AI; RSI simulator; and Zuck's technological pessimism

The new frontier of AI is developing capable autonomous researchers

Import AI· Jack Clark· 3 min· 7d agoSource →

05BUILDERS3▸

Qwen 3.8 27B is excellent, but it defaults to wildly overthinking things

Plus Gemini 3.

Simon Willison· 2 min· 8d agoSource →

Full Fine-Tuning vs LoRA

AI by Hand — Full Fine-Tuning vs LoRA

AI by Hand· Prof. Tom Yeh· 2 min· 7d agoSource →

Launch HN: machine0 (YC S26) – Persistent CPU and GPU VMs from the CLI

Hi HN! I’m Barnaby, founder of machine0 ( I’m building a CLI for long horizon agent compute: `machine0 new mybox` gives your agent a persistent cloud VM, billed by the minute, from $0. 013/hr up to 60 vCPU / 240 GB RAM and GPUs (H100s, H200s etc), with 99.

Hacker News· 1 min· 6d agoSource →

06GO DEEP1▸

★ Must ReadFrontier Model Cost and Open-Weights Popularity is Driving Demand for Model Routing

Glean CEO Arvind Jain explains why model routing helps control AI costs for organizations, and how human feedback loops at scale improve its routing systems.

Latent Space· Richard MacManus· 3 min· 6d agolong readSource →

07VOICES3▸

“An LLM wiki changed how I work”

— Casey Newton, Platformer