I continue to believe that context management is the #1 skill to develop if you want to get the most out of coding agents. But why? Modern models have a context limit of 200,000 tokens! That seems like a lot! There's a couple of reasons why... 🧵

This is a great summary of what code review is REALLY for. It also demonstrates why AI-powered code review does not help AT ALL with the most important aspects of code review: aligning the team around software design (#1 and #3 in the hierarchy). blakesmith.me/2015/02/09/c...

AI WON'T turn you into a 10x engineer but it CAN accelerate your day-to-day work. I'm hosting a live workshop early next month to help experienced devs level up from copilot to coding agents. No breathless hype, no "learn this one magical prompt" BS, I promise! blog.thepete.net/ai-assisted-...

One thing you can do with skip levels is ask questions about things you believe you've communicated. You'll be surprised how often people haven't really internalized it.

Slides from my #kubecon talk today Feature Flags Suck! - The Problems With Feature Flagging and How To Avoid Them speakerdeck.com/phodgson/fea...

Blurghed about managing you AI coding agent's context. Proactively trimming and refining the context your agent is working with is probably THE best way to improve the quality of it's work. But there's more to it than firing off a `/compact` every now and then. blog.thepete.net/blog/2025/10...

"Many developers are surprised by the changes they have to make to their workflow when they're using AI. For example, <proceeds to describe eXtreme Programming>"

Okay, for the folks who asked: here's the majority AI view, writing up the reasonable, thoughtful view on AI that the vast majority of people in tech hold, that gets overshadowed by the bluster and hype of the tycoons trying to shill their nonsense. anildash.com/2025/10/17/t... Please share!

1000000%

Super interesting proof-of-concept exploit against an AI agent via GH pull request. Very sneaky prompt injection which convinces an AI coding agent to subtly inject a remote backdoor into a codebase. A great example of @simonwillison.net's Lethal Trifecta! blog.trailofbits.com/2025/08/06/p...

Good example of why it's important to monitor your LLM-powered features in prod. Apparently some bug in Amazon's AI assistant where the product description isn't in it's prompt. Pre-production evals are super important, but wouldn't catch this. Monitoring some key metrics in prod would.

Such a nice way to illustrate the point

they should be taken to The Hague for this graph crime

There are some amazing pro-tips for AI coding in this session: github.com/hellovai/ai-... Specifically, really good advice on how to manage agent context (which can have a big impact on the LLMs performance). If nothing else, just skim the summary and see if anything looks interesting. /🧵...

Hi #AWS -using friends. I’ve been writing some articles about fundamental AWS ops for small-medium companies, all oriented around AWS Organizations. So far I’ve covered org structure, Accounts, managing human users, IaC. I have another 4 or so articles still to come, so am about half way. (1/3)

This continues to shock and confuse me: the proportion of devs who are still not getting any active encouragement from their employers to start experimenting w. AI-assisted coding. Also goes to show how silly the "all devs will be replaced by AI in 2 years" claims are. /via @angiejones.tech

Github have a lot of good general advice on code review in an AI coding world here: github.blog/ai-and-ml/ge... 🧵 I have some thoughts...

Nice post from Korny Sietsma on trying to find the middle ground between the AI coding Maximalists and the AI coding luddites. blog.korny.info/2025/07/19/c... 🧵...

Pete Hodgson continues his excellent writing on Gen AI by proposing an approach to leading engineers towards an AI-assisted future, centered around a the concept of aligned autonomy. blog.thepete.net/blog/2025/06... @thepete.net

“Listen, I’m wondering if our engineers are doing enough with AI…” your CEO says, as the two of you sit down for your 1:1. Wouldn't it be great to have a true strategy in place for adopting AI-assisted coding? I've got you covered! blog.thepete.net/blog/2025/06...

I've seen a surge of discussions recently about large AI-generated change sets that are impossible to review by humans, paired with speculation if we still need to care about the code in the future. I expect to continue to care, especially if I'm on call for it martinfowler.com/articles/exp...

The replies in this thread are fascinating. Way different from the constant breathless hype I'm used to seeing in general. A bit depressing though.

A great way to get started with feature flags!

Days like today I can't believe I get to live here

Cool, so we’ll get amazing productivity boosts with agentic AI, we just have to build Earth 2 as a test environment first…

The rule holds: if it wasn't DNS, then it was probably a config change. #hugops status.cloud.google.com/incidents/ow...

You can already read the full, unedited version of the book online at craftingengstrategy.com (most of those have also come out on my blog over the past 18 months as I wrote)

*Lots* of really good guidance and concrete tips on AI-assisted coding in this post!

Zeno: okay team, we're half way through the sprint

Yep, same. I had the exact same experience, going from really enjoying it to just not wanting to use it at all. I switched to Gemini's voice chat 🤷

Interesting analysis of the commit history for a cloudflare library where the (AI skeptic) developer tried to use Claude Code as much as possible. Seems like it went surprisingly well. www.maxemitchell.com/writings/i-r...

Detailed dissection of an async coding agent (Codex) autonomously working through a (v. simple) task. I'm perhaps beating this metaphor to death, but it really does feel like the inner monologue of a junior engineer working solo through their first ticket after joining a new team. Act accordingly.

How do you move fast, with safety? Out today, our new eBook brings together some of the brightest minds in software development from organisations like @openfeature.dev and eBay to answer this question with a practical guide to introducing feature flags at scale. www.flagsmith.com/ebook/scalin...

💯 seems inevitable. Personally I'm not sure what the solution will be, beyond making MCP tools way more limited in scope. There's no way we'll stop humans from just saying Allow All more and more, and I can't see any indication that AIs won't continue to be fundamentally gullible.

Really nice writeup of specific patterns for pairing with an AI agent to write code. It's partly a pitch for an agentic coding tool, but the playbook is pretty much tool-agnostic: almost all the techniques described are ones I've also used with Claude Code. ampcode.com/how-i-use-amp

best unicode character is "ꙮ" without a doubt. used one time in a single 15th century manuscript to emphasise "many-eyed seraphim" and unicode thought it was really important that we can type that whenever we want

Here's the full workshop handout plus annotated slides from "Building software on top of Large Language Models", a three hour tutorial I presented yesterday at PyCon US #PyConUS simonwillison.net/2025/May/15/...

one cool thing about writing is how every second of it is miserable unless you suck at it

I haven't thought about _why for ages. I'm so grateful to have spent the early years of my career in a time and subculture that was exemplified by people like _why. Like there was so much less posturing and drama and manufactured conflict, just a bunch of weirdo nerds who liked dynamic typing

Loads of detailed practical advice in here: addyo.substack.com/p/vibe-codin...

Engineering leaders at startups sometimes have to fake it as product leaders too 😅 I wrote a short guide on what a useful, succinct product strategy looks like, how to build it, and how to use it to identify and prioritize and roadmap of engineering work. blog.thepete.net/blog/2025/04...

The platonic ideal of a Pacific Northwest spring weekend. Sucia Island!

full-on vibe-coding doesn't work for real-world codebases, but we can still get many of the benefits if we take a more iterative approach. Chain-of-vibes is a workflow that lets you scale up agentic AI to non-trivial coding tasks, by keeping an engineer in the loop blog.thepete.net/blog/2025/04...

When I write a big blog post I sometimes spend way too long trying to come up with a clever metaphorical cover image. For the last few days I've been stuck doing the opposite - trying to figure out what blog post I can write that could use this as a clever metaphorical cover image

Someone once said NextDoor is the social network for dogs if they could type. It’s always “what’s that noise?” and “who’s that stranger?” which is exactly what I imagine my dogs would text other neighborhood dogs.

Great thoughtful post on how the coming flood in AI-generated changes will impact our delivery pipelines, including some strategies to handle it. I particularly like that Galo draws on analogous past case studies (i.e. Slack going through hyper-growth, Google introducing automated code changes)

I think this thing is named after the sound it makes you make

Can't remember who I cribbed these from, but they're a banger: alias gr='f() { git reflog | grep checkout | cut -d " " -f 8 | uniq | head ${1} | cat -n };f' alias gg='fn() { gr -${1} | tail -1 | cut -f 2 | xargs git checkout }; fn'