CriticGPT: Finding GPT-4's mistakes with GPT-4
CriticGPT, a model based on GPT-4, writes critiques of ChatGPT responses to help human trainers spot mistakes during RLHF
Topics /
The most popular blog posts about software engineering on Hacker News. 419 posts.
CriticGPT, a model based on GPT-4, writes critiques of ChatGPT responses to help human trainers spot mistakes during RLHF
Lessons learned from our challenges reducing LLM output: Discover everything that went wrong when trying to make our models say less.
A deep dive into the complexities of optimizing code for SIMD instruction sets across multiple platforms.
A deep dive into merykitty's 8-line SWAR temperature parser from the One Billion Row Challenge, using bitwise tricks and ALU magic with no loops or branches.
Bill is driven by the challenge of how best to quantify valuable questions that defy quantification. It's possible this instinct may have been awakened in Bill at age 14, when he won a soft, stuffed bunny at the orthodontist for guessing the number of jelly beans in the gumball machine.
From API contract to documentation portal: streamline your workflow and deliver the best API experience. Built for engineers and tech writers.
So, there's this function. It's called a lot. More importantly, all those calls are on the critical path of a key user interaction. Let's talk about making it fast.
We discuss constraint programming, an approach to solve discrete optimization problems, and illustrate how it works with a scheduling code example implemented in Python with the CP-SAT solver.
If you love exploit mitigations, you may have heard of a new system call named mseal landing into the Linux kernel’s 6.10 release, providing a protection called “memory sealing.” Beyond notes from the authors, very little information about this mitigation exists. In this blog post, we’ll explain what this syscall is…
Nine months ago, we switched to a monorepo. Here's a detailed look at what led to the change, how we approached it, and what the experience has been like.
Dr. Brian Kernighan is a Canadian computer scientist who contributed to the development of UNIX at B
To celebrate ChipWits' 40th Anniversary, the original FORTH source code has been open sourced! Learn about the game's history and source code
This is a simple experimental Lisp compiler, written in uLisp, that will compile a Lisp function into RISC-V machine code. You can run the compiler on the RISC-V core of a Raspberry Pi Pico 2 (or another RP2350-based board):
GitHub Actions now offers Arm-hosted runners with images built by Arm for developers to begin building on the latest and most sustainable processors on the market.
Introducing the search capabilities of ParadeDB.
Virtualization is a powerful technique for code obfuscation, and reversing it can be challenging. In this post, we cover the work done during an internship on developing an automated devirtualization tool. We explore a simplified taint-based approach and discuss its limitations. For a more in-depth analysis, the full…
We recently shared a number of updates with our community of users, and now we want to share them here: At Mozilla, we work hard to make Firefox the best b
From the Zed Blog: In this episode of Zed Decoded, Thorsten talks to Mikayla, who's been leading the effort to Zed working on Linux, about the Zed's Linux version and how it's taking shape
As for dangling pointers and for ownership, this model detects all possible errors . This means that we can guarantee that a program is free of uses of invalidated pointers.
Demystifying code AI: Learn how Cody analyzes context, retrieves references, prompts models, and refines completions behind the scenes.
Java users on macOS 14 running on Apple silicon systems should consider delaying the macOS 14.4 update
C2y makes memcpy(NULL, NULL, 0) and other zero-length operations on null pointers well defined. This removes a source of bugs that may result in security issues.
Pharo is a pure object-oriented programming language and a powerful environment focused on simplicity and immediate feedback.
Learn how Firezone uses the sans-IO pattern to build testable, deterministic Rust networking code. A deep dive into our connlib design.
How we implemented fair queueing strategies for Postgres-backed task queues to ensure fair execution time across tenants.
This article explores the design decisions of typing lists and tuples in Elixir within a sound gradual type system.