4561
LLMs predict my coffee
Why not benchmark with physical experiments?
Top posts
Every blog post that took off on Hacker News, ranked by the votes of its best submission. 5,962 posts.
4562
Why Mathematica does not simplify sinh(arccosh(x))
When Mathematica seems to not to know how to simplify an expression, it may be doing the right thing, being more sophisticated than you expect.
4563
Starting from scratch: Training a 30M Topological Transformer
Tauformer is a topological transformer (see paper ) that replaces dot‑product attention with a Laplacian-derived scalar (taumode) per token/head, then attends using distances in that scalar space. Below is a post-style overview of the idea and the first training signals from a 30M-parameter run.
4564
21 years and counting of 'eight fallacies of distributed computing' (2025)
Eight long held and common beliefs about the network have been shown, time after time, to be false. What are they, and what do they mean?
4565
4566
Competition is not market validation
How to not fall into the fallacy of seeing competition as a proof of Product-Market Fit
4567
Formal Verification Gates for AI Coding Loops
Formal verification gates for AI coding loops, in the language you're already shipping.
4568
Introducing: ShaderPad
Riley J. Shaw is a programmer and designer. Rooted in repair and craft, he creates transparent, remixable tools for a sustainable and accessible future.
4569
How Quickly Can the DoD Rebuild and Recast the Munitions Industrial Base?
The U.S. munitions industrial base should be rebuilt and recast for modern wars of precision and attrition. What is a realistic timeline for this goal? By analyzing DOD budget documents and industry announcements, this analysis provides a clearer picture of the planned future.
4570
Private healthcare makes industries less innovative. It's time for change
1 in 4 American workers stay in their jobs because of the healthcare. It suppresses wages and working conditions - and prevents innovators from heading out on their own.
4571
I made 20 GDPR deletion requests. 12 were ignored
I made 20 GDPR data deletion requests in the past year. 2 complied immediately. 6 complied after I filed complaints with data protection offices. 12 still haven't. The regulation exists, the enforcement doesn't.
4572
The Agent SDK
The Agent SDK is the framework we use to build our own managed agents like Bugbot, Security Reviewer, and Rollouts. Define agents in code and deploy them with one command.
4573
They lied to you. Building software is hard
Every week there seems to be a new tool that promises to let anyone build applications 10x faster. The promise is always the same and so is the outcome.
4574
British Columbia, Time Zones, and Postgres
This year, British Column has moved to year-round Pacific Time. How does that affect date data?
4575
The future of independence is interdependence
Every one of us depends on others; the only difference is how visible that dependence is.
4576
OpenAI Frontier
OpenAI Frontier is an enterprise platform for building, deploying, and managing AI agents with shared context, onboarding, permissions, and governance.
4577
I don't know Apple's endgame for the Fn/Globe key–or if Apple does
The origin and the evolution of the most confusing modifier key
4578
Just Put It on a Map
An underrated strategy for urbanist persuasion, powered by open source tools
4579
An autopsy of AI-generated 3D slop
Thinking of using AI for your Shopify 3D models? Read this first. We compare AI-generated 3D 'slop' vs. handcrafted models to show why the human touch is very much required to attain positive ROI.
4580
So you want to write an “app” (2025)
I touched "every" OS so you don't have to
4581
Too Much Color
I spent too much time looking at too many colo(u)rs to try and optimise them for csskit. Here are some interesting findings.
4582
4583
Hackers influence ChatGPT and Gemini to direct users to scam centers
Think you can trust AI to give you correct answer? Think again.
4584
Meta's Muse appears to use an OpenAI model labeled muse-special
While Muse was building my website, one subagent ran on a model called azure/muse-special. So I dug through the filesystem again.
4585
Ramp's Sheets AI Exfiltrates Financials
A vulnerability in Ramp's Sheets AI allowed the agent to insert formulas that made external network requests without user approval, creating the risk of data exfiltration via indirect prompt injection.
4586
Job queues are deceptively tricky
A worked out example of different job queueing semantics and how they affect workloads.
4587
Why are neural networks and cryptographic ciphers so similar? (2025)
Several neural network design principles are almost identical to cipher design principles.
4588
The RISE RISC-V Runners: free, native RISC-V CI on GitHub
Today, we are excited to announce the Early Availability of the RISE RISC-V Runners , a free, managed GitHub Actions runner service that gives any open source project access to real RISC-V hardware in their CI pipelines. No emulation, no cross-compilation hacks, no waitlist. Install a GitHub App, change one line in…
4589
4590
Announcing Rust 1.96
Empowering everyone to build reliable and efficient software.