Android 17 APIs, Cloudflare Quick Tunnels, and passkeys debate
First things first
People are buzzing about new tech quirks and big shifts, from AI mishaps to privacy debates. It’s like we’re riding a rollercoaster of innovation, with some wild stories about AI design, security, and even the human brain. Things are changing fast, and it’s kinda wild to keep up.
Meanwhile, folks are also talking about global tension, data laws, and how AI influences everything from chips to military stuff. It’s a mix of excitement, concern, and curiosity - like the world’s on the edge of some big moves.
The top 6
Android 17 is the first since 3.x to add new APIs without releasing to the AOSP
Cloudflare Quick Tunnels
Cloudflare Quick Tunnels allows developers to securely share local applications globally with a single command, without needing accounts or complex setup. Built on Cloudflare's global network, it provides automatic HTTPS, DDoS protection, and fast, reliable connections for testing and collaboration.
Claude Code now reads AGENTS.md if there is no Claude.md
US Military had close call after using AI for hallucinated intelligence report
Bend 2 and the Vibe-Coding Trap
Support our sponsor
Your Newsletter Deserves Better Than a Free Plan
Quote of the day
Thousands of candles can be lighted from a single candle, and the life of the candle will not be shortened. Happiness never decreases by being shared.
Buddha
Also big today
I Vibed a Proof of Conway's Conjecture
Saving another 100TB of RAM with math (and Rust)
Korea raises data breach fines to 10% of revenue
Human brain is two separate organs, Stanford Medicine-led research finds
Border agents can search cellphones without a warrant or reasonable suspicion
North Korean nuclear test sets off years of earthquakes – Science – AAAS
Pass it on
Enjoying the digest?
Forward it to someone who'd like it, or send them the link to subscribe.
Share hn.todayOne for the road
What do lions use to look at their manes?
Mirroars.
Worth a look
There's no point at which turning your brain off will work
Using large language models often leads to people turning off their critical thinking and blindly trusting the outputs, which can result in errors and ineffective results. As LLMs improve, relying solely on them without active oversight becomes increasingly problematic, especially for high-value tasks.
US troop deaths during Iran war exceed Pentagon count by at least four
San Francisco Onion Futures Company
US Treasuries Have Become Unappetizing for Foreign Central Banks and Governments
Foreign central banks and governments have reduced their Treasury holdings to $3.77 trillion, the lowest since 2012, as Treasuries become less attractive. Meanwhile, US hedge funds and companies increasingly hold Treasuries through opaque foreign financial centers, making the US less dependent on official foreign investors.
An Empirical Study of Harness Design for Coding Agents
A study evaluates components of coding harnesses that improve autonomous coding agents' performance, focusing on planning, action space, and context management. Results show context management is crucial for small context windows, and different strategies impact efficiency and cost depending on model strength.
Photon-Emission-Guided Laser Fault Injection Enables RP2350 Secure Debug
Song of the day
Deep cuts