Loading…

The Overflow

A newsletter by developers, for developers, curated by Cassidy Williams and the Stack Overflow team. Every week, we’ll share a collection of great questions from our community, news and articles from our blog, and awesome links from around the web.

Email settings

Issue 219: Memory safety for national security

This week: tips on how to get your RAG in order, pictures that let you program, and Java inches closer to becoming Rust.

Issue 218: Textbooks for robots

This week: how we're partnering to ensure LLMs acknowledge our community's contributions, why a slow train is like walking into a wall, and what design patterns encourage junk data.

Issue 217: Discussing the discussions

This week: an answer for the perennial question of who owns this service, a question about the name for scissor makers, and names of companies that are also the name of a person.

Issue 216: Functional time-travel

This week: the AI bot that fixes security flaws, the legality of phone lines that are usually unusually busy, and tech jobs lose their shine.

Issue 215: DIY LLM

This week: your company's stock price is taking our jobs, brain development after 25, and why 100% unit test coverage is actually a bad idea.

Issue 213: Is gzip an LLM?

This week: we're celebrating our annual Stack Gives Back event, cutting to the chase in work DMs, and wondering what the rest of the world thinks about Comic Sans.

Issue 212: Socially responsible AI

This week: the incredible shrinking life of developer skills, the future dearth of space pirates, and the impeding ubiquity of IPv6.

Issue 211: Breaking Tetris

This week: creating a role-playing video game in a single day with AI, casting doubt on rocket exhaust casting shadows, and admitting that those cookie banners are really annoying.

Issue 210: The last byte of the internet

This week: we revisit a pair of old favorites, question the intelligence of the morons of Mordor, and desperately try to halt the unstoppable march of a `forEach` loop.

Issue 209: The best of 2023

New year, same old reliable newsletter. This week, we're recapping the top five blogs from last year and sharing some fresh questions and links.

Issue 208: Three ways to program with AI

This week: we discuss how to modernize alerting and incident management, ride a reindeer, and adopt a developer tool.

Issue 207: One weird trick

This week: building an AI platform for business with IBM, refusing to answer questions on a jury, and the fastest way to String in C++.

Issue 206: Inference-as-a-service

This week: Will GenAI generate productivity? Does spite look better than a PhD on your resume? How can you post an email address on the internet without bots getting ahold of it?

Issue 205: Engineering thanks

This week: discussing inference-as-a-service on the edge, preventing your code form being used as training data, and writing clean code to reduce cognitive load on developers.

Issue 202: Accepted

This week: how and why developers are changing jobs, whether you can gamble with stolen money, and why text embeddings matter.

Issue 201: Rise of the coffeebots!

This week: what you need (besides AI) to improve productivity, forcing diplomats to pay back rent, and the downsides of mobile-first web design.

Issue 200: Reducing meeting blast radius

This week: how to integrate an AI tool into your existing workflow, advise PhD candidates smarter than you, and run neural networks in your browser.

Issue 199: Code every Zig

This week: Can infrastructure-as-code be made simpler? Why are blood iron measurements different in your right and left hands? Should academic disciplines have an expiration date?

Issue 198

This week: we're looking back at 15 years of Stack Overflow with a snazzy poster, thinking of critiquing our skip-level boss, and understanding why open source wins.

Issue 197

This week: an IDE built for multiplayer, a new antibiotic that hasn't hit clinical use, and an attempt at an open standard for accelerator programming.

Issue 196

This week: understanding what VC investors are looking for, when SOLID principles are on shaky ground, and why not everything should be lazy loaded.

Issue 195

This week: the dramatic conclusion of our cloud journey, the math behind a good toy train track, and the case against walled gardens.

Issue 194

This week: the first part of our saga of moving Teams to the cloud, the reason that programming languages usually delimit strings, and the end of the argument whether tabs or spaces are better.

Issue 193

This week: how to pay off tech debt by giving it numbers, why type systems stop at a level of expressiveness, and why one developer built a garbage collector for a language that doesn't need it.