The famous o3 "GeoGuessr" prompt did not work
Read it on seangoedecke.com In April last year, Kelsey Piper discovered that OpenAI’s o3 model was surprisingly good at figuring out where a photo was taken...
Prompts are technical debt too
Read it on seangoedecke.com It’s common and correct to say that “all code is technical debt”. Adding code is a necessary evil for developing new features:...
The just-say-no engineer was a ZIRP phenomenon
Read it on seangoedecke.com The engineer who says no all the time is a real archetype among senior and staff engineers. Their role is to slow things down, to...
How I use LLMs as a staff engineer in 2026
Read it on seangoedecke.com A bit over a year ago I wrote How I use LLMs as a staff engineer. Here’s a brief summary of what I used AI for last year: Smart...
DeepSeek-V4-Flash means LLM steering is interesting again
Read it on seangoedecke.com Ever since Golden Gate Claude I’ve been fascinated with “steering”: the idea that you can guide LLM outputs by directly...
AI datacenters in space do not have a cooling problem
Read it on seangoedecke.com This year Elon Musk has started banging the drum about building AI datacenters in space. As the only person who owns a successful...
Thinking Machines and interaction models
Read it on seangoedecke.com Thinking Machines just released Interaction Models. This is their first real AI model release1 after a year of work and two...
The left-wing case for AI
Read it on seangoedecke.com In Many anti-AI arguments are conservative arguments I argued that left-wing anti-AI sentiment1 is partly a backlash to two...
AI makes weak engineers less harmful
Read it on seangoedecke.com Like other kinds of puzzle-solving, software engineering ability is strongly heavy-tailed. The strongest engineers produce way...
Notes on incidents
Read it on seangoedecke.com Incidents are boring. Most of what you actually do during an incident is wait: for some other team to investigate, or for a...
Why hasn't longer-horizon training slowed AI progress?
Read it on seangoedecke.com Dwarkesh Patel1 recently posted an award for the best answers to four key questions about AI. It’s partly a challenge and partly...
Why I don't like the "staff engineer archetypes"
Read it on seangoedecke.com The most influential piece of writing about staff engineers in the last decade has to be Will Larson’s Staff engineer archetypes....
Software engineering may no longer be a lifetime career
Read it on seangoedecke.com I don’t think there’s compelling evidence that using AI makes you less intelligent overall1. However, it seems pretty obvious...
Luddites and AI datacenters
Read it on seangoedecke.com Is it time to start burning down datacenters? Some people think so. An Indianapolis city council member had his house recently...
Many anti-AI arguments are conservative arguments
Read it on seangoedecke.com Most anti-AI rhetoric is left-wing coded. Popular criticisms of AI describe it as a tool of techno-fascism, or appeal to...
Programming (with AI agents) as theory building
Read it on seangoedecke.com Back in 1985, computer scientist Peter Naur wrote “Programming as Theory Building”. According to Naur - and I agree with him -...
Working on products people hate
Read it on seangoedecke.com I’ve worked on a lot of unpopular products. At Zendesk I built large parts of an app marketplace that was too useful to get rid...
Engineers do get promoted for writing simple code
Read it on seangoedecke.com It’s a popular joke among software engineers that writing overcomplicated, unmaintainable code is a pathway to job security....
Big tech engineers need big egos
Read it on seangoedecke.com It’s a common position among software engineers that big egos have no place in tech1. This is understandable - we’ve all worked...
I don't know if my job will still exist in ten years
Read it on seangoedecke.com In 2021, being a good software engineer felt great. The world was full of software, with more companies arriving every year who...
