3991
New
The latest blog posts to take off on Hacker News. 6,220 posts.
3992
1,145 pull requests per day
There is a recent video of Patrick Collison at Stripe Sessions 2025 stating that in 2024 Stripe did on average 1,145 pull requests per day. Not just creating them, but actually finishing them; "fully shipped into production". All whilst having less than a minute of API unreliability for the
3993
Loading Pydantic models from JSON without running out of memory
Pydantic’s JSON loading uses a huge amount of memory; here’s how to reduce it.
3994
The "AI 2027" Scenario: How realistic is it?
Everyone’s been talking about it. JD Vance read it. What does it actually tell us?
3995
We’ll be ending web hosting for your apps on Glitch
We’ve got an important update for the Glitch community today: We’ll be ending web hosting for your apps on Glitch.
3996
Claude 4
Discover Claude 4's breakthrough AI capabilities. Experience more reliable, interpretable assistance for complex tasks across work and learning.
3997
That fractal that's been up on my wall for years
Warning: Math, Handwaving I spent a lot of time doodling in middle school instead of whatever middle schoolers are supposed to do. Somewhere between the Cool S’s and Penrose triangles I found a neat way to fill up graph paper by combining and copying squares.
3998
MCP explained without hype or fluff
Model Context Protocol, like most protocols, solves the M ⨯ N integration problem by turning it into an M + N integration problem.
3999
I Built My Own Audio Player
In 2025, playing your own music on an iPhone is surprisingly hard, unless you pay Apple or navigate a maze of limitations. So I built my own player from scratch, with full text search, iCloud support, …
4000
Fast Allocations in Ruby 3.5
Many Ruby applications allocate objects. What if we could make allocating objects six times faster? We can! Read on to learn more!
4001
Adventures in Symbolic Algebra with Model Context Protocol
Personal blog of Stephen Diehl - Software engineer writing about technology, programming, and the future
4002
Trading with Claude, and writing your own MCP server
Ever wanted to check on your portfolio and trade some stocks directly with Claude? Well you’re in luck.
4003
Improving performance of rav1d video decoder
*on macOS with an M3 chip *slightly more than 1%, on a specific benchmark, without any new unsafe code A while ago, memorysafety.org announced a contest for …
4004
Planetfall
Gentle readers, I have just wrapped up a fun side project that will be of great interest to a very small number of you. The result of one of the most technically demanding efforts of my career, I a…
4005
Ancient law requires a bale of straw to hang from Charing Cross rail bridge
The scaffolding surrounding the Charing Cross railway bridge has received an addition – two bales of hay/straw – because an ancient law requires it.
4006
Why does Debian change software?
When I wrote Why is Debian the way it is? , a year and a half ago, I was asked to also cover why Debian changes the software it packages. Here's a brief list of examples of why that happens:
4007
You can choose tools that make you happy
Stop falsifying your motivations.
4008
The scientific “unit” we call the decibel
Celebrating a rare dumpster fire in the kingdom of science.
4009
Getting a paper accepted
Page 1 Accepts, the Rest Avoids Rejection
4010
Gemini Diffusion
Another of the announcements from Google I/O yesterday was Gemini Diffusion, Google's first LLM to use diffusion (similar to image models like Imagen and Stable Diffusion) in place of transformers. …
4011
I have tinnitus. I don't recommend it
i went to a show in november of last year (it’s may right now) and ever since, i have tinnitus. i don’t recommend it.
4012
Rocky Linux 10 Will Support RISC-V
🏗️ Rocky Linux 10 Will Support RISC-V! We're excited to announce that Rocky Linux 10 will officially support RISC-V! Thanks to the incredible work of...
4013
An upgraded dev experience in Google AI Studio
Explore the upgraded Google AI Studio with Gemini 2.5 Pro, agentic tools, and enhanced multimodal capabilities for developers to build sophisticated AI apps.
4014
LLM function calls don't scale; code orchestration is simpler, more effective
TL;DR: Giving LLMs the full output of tool calls is costly and slow. Output schemas will enable us to get structured data, so we can let the LLM orchestrate ...
4015
A letter from Sam & Jony
Building a family of AI products for everyone.
4016
Collaborative Text Editing Without CRDTs or OT
Matthew Weidner | May 21st, 2025 Home | RSS Feed Keywords: text editing, server reconciliation, Articulated
4017
By default, Signal doesn't recall
Signal Desktop now includes support for a new “Screen security” setting that is designed to help prevent your own computer from capturing screenshots of your Signal chats on Windows. This setting is automatically enabled by default in Signal Desktop on Windows 11. If you’re wondering why we’re on...
4018
The Era of the Business Idiot
Fair warning: this is the longest thing I've written on this newsletter. I do apologize. Soundtrack: EL-P - $4 Vic Listen to my podcast Better Offline. We have merch. Last week, Bloomberg profiled Microsoft CEO Satya Nadella, revealing that he's either a liar or a specific kind of
4019
Introducing the Llama Startup Program
We’re excited to announce the Llama Startup Program, a new initiative to empower early stage startups to innovate and build generative AI applications with Llama.
4020
ZEUS – A new two-petawatt laser facility at the University of Michigan
Hitting 2 petawatts, the NSF-funded ZEUS facility at U-M enables research that could improve medicine, national security, materials science and more.