Addy Osmani
Engineering leader on Google Chrome; writes about web performance and, lately, about working with AI coding tools.
Addy Osmani did not write this page. ppll compiled it from their own words, each linked to the source. Is this you? · Report a mistake
Latest x.com
What they believe
-
Computing is an onion, and the more layers of it you can peel back and understand, the better you can build for the constrained machines real people are using.
if you treat computing as this onion where you just keep peeling back the layers, there's always something interesting behind the scenes. And the more of those layers that you can peel back and understand, I think in some ways the better you can optimize for that world.
youtube.comSaid 19 Aug 2026This on korrents
The bold line is korrents’ wording; the quote is theirs. These one: what they wrote before what they said aloud, one per source, newest first.
Feed
As its own page →Hiding
1 November
28 October
27 October
25 October
23 October
21 October
18 October
17 October
11 October
10 October
9 October
7 October
5 October
14 September
5 September
31 August
27 August
24 August
21 August
19 August
From one piece From Chrome DevTools to AI Engineering, with Addy Osmani 15 beliefs · youtube.com
-
Their words
And it's effectively the next step of, you know, every phase of of software evolution is just like a rising tide of abstractions. This is the next abstraction.
-
korrents.com
Letting automated loops build everything without guardrails on blast radius and on quality is a recipe for disaster.Their words
But, simply just having your loops build everything without having some guardrails around the blast radius without having guardrails around how you think about quality, I think is a recipe for disaster.
-
Their words
I'm a really big fan of this idea of mutual amplification. If you are working with an agent, a coding agent, there are a lot of things that you can do to make sure that the agent is getting better every day and you as an engineer are getting better every day.
+ 12 more
-
Their words
The first thing I do is I try to make sure that if there is a summary at the very end, here are all the decisions that were made, I will read through that end to end. If there hasn't been, I will prompt for that decision process.
-
Their words
I think that we still need to understand enough about how things work so that if something does go wrong, we're actually able to fix it and not just hope and pray that the agent is able to figure things out.
-
Their words
the first is cognitive debt. So the more that you use AI, it's it's sort of the erosion of your ability to have good memory and have good understanding of the problems that you're working on. And the natural follow-up to that is cognitive surrender, which is where you, you know, you blindly give in to whatever the AI says as your answer.
-
Their words
One of the things I found most exciting in the last couple of years was seeing as model qualities gotten better and and harnesses and tools have gotten better, how many people um that were directors or VPs or SVPs or any of these levels were actually rolling up their sleeves and trying things out.
-
Their words
you want to get to a point where you know, your your machine, your org is self-sufficient enough that you know, you just need to occasionally tap the blimp, make sure that things are working. You can course correct if it's not, but that frees you up to then focus on the next important sets of problems that the org needs to, you know, tackle heads on.
-
Their words
And that embrace of people are going to use whatever tech they want to use. You can't tell people what to use very often. They're going to use whatever they want, and your job is to help them be successful on your platform, and to help your users have a great time.
-
Their words
even if you set aside, you know, maybe there's some validity around a business needs to monetize, those things shouldn't cause a really bad experience.
-
Their words
Humans are humans are are shockingly simple, you know, uh, if you think about the experience you have with somebody that's just trying to cross the street, if, uh, you know, if the light doesn't turn, you know, doesn't say they can walk fast enough, they'll just keep hitting that button. That's the same experience they have on the internet.
-
Their words
I think for a very long time we had this this almost nebulous way of thinking about page load times. And the team felt like it was finally time to come up with a more nuanced perspective around how we reason about performance.
-
Their words
I remember anytime I would work with a big site on their performance problems, uh you could easily spend half a day, um you know, just looking at traces before you've even written any fixes at all. And now that we have LLMs, it's very quick to like reason through massive stack traces and actually be able to get down to fixes you can make.
-
Their words
I would say, sometimes, I don't know if it's controversial, that very few developers understand memory management. And that makes it even harder to debug memory problems. And so the state of the art around memory debugging hasn't evolved all that much over the years.
-
Their words
if you treat computing as this onion where you just keep peeling back the layers, there's always something interesting behind the scenes. And the more of those layers that you can peel back and understand, I think in some ways the better you can optimize for that world.
16 August
14 August
9 August
8 August
20 July
16 July
15 July
14 July
6 July
24 June
20 June
1 May
Nothing matches.
Beliefs
Korrents What they believe 22 beliefs — each backed by an exact quote.
Each is a — compiled by korrents.com, not by them: the one-line wordings are korrents', the quotes are theirs.
Recent
Computing is an onion, and the more layers of it you can peel back and understand, the better you can build for the constrained machines real people are using.
if you treat computing as this onion where you just keep peeling back the layers, there's always something interesting behind the scenes. And the more of those layers that you can peel back and understand, I think in some ways the better you can optimize for that world.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
Very few developers understand memory management, and that is why the state of the art in memory debugging has barely moved in years.
I would say, sometimes, I don't know if it's controversial, that very few developers understand memory management. And that makes it even harder to debug memory problems. And so the state of the art around memory debugging hasn't evolved all that much over the years.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
Performance work used to mean half a day reading traces before writing a single fix, and language models have collapsed that step to minutes.
I remember anytime I would work with a big site on their performance problems, uh you could easily spend half a day, um you know, just looking at traces before you've even written any fixes at all. And now that we have LLMs, it's very quick to like reason through massive stack traces and actually be able to get down to fixes you can make.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
Show 19 more
"Is the page loaded?" was never a real question, and web performance only became measurable once it was broken into the distinct moments a user actually notices.
I think for a very long time we had this this almost nebulous way of thinking about page load times. And the team felt like it was finally time to come up with a more nuanced perspective around how we reason about performance.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
People treat an unresponsive interface exactly the way they treat a slow pedestrian crossing: they keep hitting the button.
Humans are humans are are shockingly simple, you know, uh, if you think about the experience you have with somebody that's just trying to cross the street, if, uh, you know, if the light doesn't turn, you know, doesn't say they can walk fast enough, they'll just keep hitting that button. That's the same experience they have on the internet.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
A site's need to monetise is not a defence of an experience that shifts the page under the reader; those are separate questions.
even if you set aside, you know, maybe there's some validity around a business needs to monetize, those things shouldn't cause a really bad experience.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
A platform cannot tell developers what technology to use; its only job is to help them succeed with whatever they have already chosen.
And that embrace of people are going to use whatever tech they want to use. You can't tell people what to use very often. They're going to use whatever they want, and your job is to help them be successful on your platform, and to help your users have a great time.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
The point of building a self-sufficient organisation is not to step back from it, but to free yourself to work on the problem the organisation has not reached yet.
you want to get to a point where you know, your your machine, your org is self-sufficient enough that you know, you just need to occasionally tap the blimp, make sure that things are working. You can course correct if it's not, but that frees you up to then focus on the next important sets of problems that the org needs to, you know, tackle heads on.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
AI tooling has done something no earlier technology managed: pulled directors, VPs and SVPs back into building things with their own hands at the weekend.
One of the things I found most exciting in the last couple of years was seeing as model qualities gotten better and and harnesses and tools have gotten better, how many people um that were directors or VPs or SVPs or any of these levels were actually rolling up their sleeves and trying things out.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
Heavy AI use erodes understanding first and judgement second: cognitive debt turns into cognitive surrender, where the model's answer simply becomes your answer.
the first is cognitive debt. So the more that you use AI, it's it's sort of the erosion of your ability to have good memory and have good understanding of the problems that you're working on. And the natural follow-up to that is cognitive surrender, which is where you, you know, you blindly give in to whatever the AI says as your answer.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
You have to keep enough understanding of how a system works to fix it yourself, because the alternative is hoping and praying the agent can figure it out.
I think that we still need to understand enough about how things work so that if something does go wrong, we're actually able to fix it and not just hope and pray that the agent is able to figure things out.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
Reading every agent trajectory has stopped being possible, and what replaces it is demanding an end-to-end account of the decisions that were made.
The first thing I do is I try to make sure that if there is a summary at the very end, here are all the decisions that were made, I will read through that end to end. If there hasn't been, I will prompt for that decision process.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
Working with a coding agent should leave both of you better every day, and that only happens if you deliberately record what was learned and decided in the session.
I'm a really big fan of this idea of mutual amplification. If you are working with an agent, a coding agent, there are a lot of things that you can do to make sure that the agent is getting better every day and you as an engineer are getting better every day.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
Letting automated loops build everything without guardrails on blast radius and on quality is a recipe for disaster.
But, simply just having your loops build everything without having some guardrails around the blast radius without having guardrails around how you think about quality, I think is a recipe for disaster.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
Building a system that does the prompting for you is not a break with software history but the next rung on the same ladder of rising abstraction.
And it's effectively the next step of, you know, every phase of of software evolution is just like a rising tide of abstractions. This is the next abstraction.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
An agent can tell you whether a thing looks correct against a spec; it cannot tell you whether the thing is any good.
An agent can tell you if a thing looks correct, if it's matching a spec. It doesn't necessarily mean it can tell you what's good.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
Whatever humans are still better at than models is alpha that decays with every model release, so the skill worth naming today will not be the one worth naming next year.
what what is the current thing that models are not very good at doing? Alpha is going to decay in some way with every model release or every series of model releases. So, it's going to change over time.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
What engineers will stay valuable for is being answerable for a system — the named person who decides what ships, what is blocked and what is deferred, whoever or whatever wrote the code.
But they're the person that's on the hook for understanding, for gating, for making sure that someone is deciding what ships, what's blocked, what do we defer. And so, I think that that is something that engineers are going to continue to be valuable for.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
Every time software has become easier to create, the world has created exponentially more of it, and this time will be no exception.
every time that we've made it easier to create software, we've exponentially created more of it.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
Now that anyone can generate a body of text, the only thing that justifies asking for a reader's time is whether the idea itself is worth reading.
now that it is very easy for anyone to use an agent to create a body of text, I think it's more important than ever for us to make sure that the ideas we're putting out in the world are actually worth people reading.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
Editing your writing with a model pulls it towards one homogeneous style, far enough that you lose track of what your own voice was.
the models are going to in some cases, I think, provide a homogeneous take on what writing looks like. And you start to feel like, wait, what was my writing style, you know?
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
Engineering, product and UX roles are converging, so the career advice that matters now is: do not just be an engineer.
And if you can show employers that you are not just a builder, but you are someone that can help them as these roles start to become a little bit fuzzier. I think that you can be successful in these times. Don't just be an engineer.
From Chrome DevTools to AI Engineering, with Addy Osmani Said 19 Aug 2026
What is a korrent?
A korrent is a belief a person has stated in their own words: one sentence stating the claim, backed by a quote and a source, kept at korrents.com.
Under a name here, the quoted block is what they actually said. The korrent beneath it is the claim those words support, in korrents' wording — tap it to see the record, its source, and who else holds it.
Nobody here wrote their own korrents. They are compiled from public statements, and a person can change their mind, which is recorded too.
Recommends rcmnd
rcmnd What they recommend 7 things — each backed by an exact quote.
Compiled by rcmnd.app, not by them: the quotes are theirs, and each links to the post it was copied from.
Recent
Recommends The Innovator's Dilemma
Disruptive innovations often come from outsiders and new entrants, not incumbent leaders. Focus on customers’ evolving jobs-to-be-done, not existing product categories. Be ready to cannibalize your own products.
addyosmani.com 8 Oct 2023
Recommends The Culture Code
Build psychological safety on teams so people feel comfortable taking risks. Encourage small signals of inclusion and caring between members.
addyosmani.com 8 Oct 2023
Recommends Radical Candor
Directly challenge people when their work needs improvement - be caring personally but candid professionally. Praise good work clearly and sincerely. Show you notice and appreciate their efforts.
addyosmani.com 8 Oct 2023
Show 4 more
Recommends Thinking, Fast and Slow
Understand the two systems driving decisions - fast intuitive vs. slow analytical thinking. Know their biases. For complex analysis, create space from intuitions and emotions. Slow down and use structured reasoning.
addyosmani.com 8 Oct 2023
Recommends Multipliers
Empower and challenge team members instead of micromanaging. Get people thinking and taking initiative. Create space for people to contribute beyond their official roles. Draw out their skills.
addyosmani.com 8 Oct 2023
Recommends Atomic Habits
Break down big goals into small, manageable habits. Focus on processes, not outcomes. Create obvious cues and routines for habits to make them easy to start.
addyosmani.com 8 Oct 2023
Recommends Deep Work
Minimize distractions and shallow work. Carve out long stretches for intense focus and complexity. Build rituals and routines to regularly do deep work. Protect this time.
addyosmani.com 8 Oct 2023
Where they publish
Blog addyosmani.com His blog on web performance, JavaScript and AI-assisted engineering. Has a feed.
Recent
- Brownfield Agentic Engineering 14 Sept 2026 What it takes to run agents in a codebase older than the team
- Agentic Skill Decay 31 Aug 2026 Agents can finish the task without teaching you anything. Building expertise now has to be deliberate.
- Audit your Agent files 27 Aug 2026 A practical guide to auditing what your coding agent still needs.
Show 7 more
- Human judgment doesn't leave the software factory. It relocates. 21 Aug 2026 A field guide to building a software factory that still has an owner.
- Practical Loop Engineering 14 Aug 2026 Goals, loops, and the discipline of not delegating your judgment
- Agentic Code Quality 8 Aug 2026 Quality now depends on the constraints you set around your agents.
- Software Factories, Light and Dark 20 Jul 2026 A software factory is harnessing loops at scale - the system that builds your software instead of you. You can run the loop with humans in it (light factory) or you can ignore the humans (dark factory) and let those age…
- Own the Outer Loop 15 Jul 2026 Engineers need to own the outer loop, the accountability for these systems. A written version of my AI Engineer World's Fair 2026 closing keynote: quality, verdict, and answerability, the three hidden costs of delegatio…
- Earning taste and judgment 14 Jul 2026 Taste used to be a byproduct of the reps. Agents took the reps. So if you're junior you now have to go get the taste and judgment on purpose. A builder's look at the weakening entry level, the two debts we now owe, and…
- The Agent-Era Career 6 Jul 2026 AI gets good at anything with an answer key. Your career is everything that doesn't have one.
Link verified 20 Sept 2026. Recent items update automatically from the channel.
x.com @addyosmani Their X account.
Recent
- Claude Code tip: if a task turns out harder than you expected, type ultrathink while Claude is working. It raises the effort for the rest of that turn. In the terminal on the latest release. 1 Nov 2026
- "Hunger beats tenure." Great write-up by Felix on why his team at Anthropic dropped coding interviews: newer engineers are flying past stronger, more tenured coders. Execution isn't expensive anymore. Inside a company, the signal is picking an idea that moves the business and prototyping it. Outside, it's a portfolio of small prototypes. Both should show off your taste and judgment. Claude Code Desktop started as a weekend prototype by two newer engineers. 28 Oct 2026
- Before adding another subagent ask what problem it's solving. A bloated system prompt is often the issue. Skills and tool search can fix that without splitting context across agents. My rule of thumb is to keep a subagent when the work is parallel, needs fresh context or one agent really can't pick the right tools. Otherwise, fold it into the main loop, turn its prompt into a skill, run your evals and keep whatever wins. Good write-up: 27 Oct 2026 claude.com
Show 12 more
- Want better quality output from your agent? Give it a way to verify its own work. Love what Warp is doing here: every agent gets a computer and brings back a video of what it did. I do something similar in my workflow - on top of E2E tests, Claude gives me recorded videos + screenshots as part of the evidence work is "done". My bet is verification becomes one of the most important problems to work on this year. 23 Oct 2026
- Claude Duet is our new app where Claude is your proactive agent! Runs on Opus 5.5 across 6,000 connectors/plugins. Duet lets you know what needs your attention and keeps going when you step away. Join the waitlist for the research preview! 21 Oct 2026 claude.com
- Give your agent a way to verify its work. This can dramatically improve the quality of output. With this process in place, Claude Code was able to write 90%+ of its code itself. When generating software at a faster rate than you can review it, the bottleneck shifts to proving that it works. @bcherny and team invested enough in verification, we've been able to move fast while maintaining stability. Code reviews aren't going away, but you need to think of it differently. One human reading through line-by-line is no longer going to work as a gate, it needs to evolve into a multi-layered verifica… 17 Oct 2026
- Agents write code "much better on average at scale" is right. The "at scale" part comes from verification. Agents make checking cheap. The stuff that made "Good Code" expensive (tests, benchmarks, specs, real e2e runs) is now a cheap thing you can ask an agent for. You can improve the checks. So our work is less "review every line" and more "build loops the agent can run itself". Give the agent a way to verify the work. That loop is where quality comes from. And make taste compound. When the agent gets something wrong, write it down (shared config, a hook, a test) so it doesn't happen twice.… 11 Oct 2026
- Verification is one of the most important things for getting great results out of your agent. Clearly define "done" before code review starts. A reviewer, human or agent, asked to find bugs without this may still find something that could be improved. Give Claude checks it can run (bash commands, tests, clicking through the app in a browser) and have it pass that too before it says it's finished. When nothing blocks merge and ship. 11 Oct 2026
- "your edge right now is understanding the possibilities and being curious" Using AI for software engineering is now the default. The engineers I've seen who have the edge are ambitious and have been grinding with agents for years on the job at this point. It's a compounding effect. For teams still unsure or catching up, the smartest thing I can advise is to be even more ambitious and aggressively share what you learn with your peers. If you do all this work privately, only you benefit. If your team only gets to experiment with AI after hours, you've quietly told them using it doesn't count. S… 10 Oct 2026
- Don't forget: Claude Max & Team now include monthly API credits. Max 5x: $100, Max 20x: $200 Great for building apps & agents via the API, Managed Agents or Agent SDK. We hope you'll enjoy Opus 5.5 there too. 10 Oct 2026
- Pro and Max users will soon have Claude Code Projects! I've been enjoying using them. You use a project when you have a stream of related work that outlasts one session. Claude coordinates it, runs each task as a parallel cloud thread and you set the instructions, repos and memory once instead of repeating them. You can also hand off a batch of tasks, walk away and come back to the Overview pane showing which threads finished, which PRs are ready and which need your answer. Check out @delba_oliveira's video for more! 9 Oct 2026
- Whether agents feel like liberation or loss depends on which joy of engineering you lean on most: making, knowing, or mattering. I've seen engineers say many conflicting things about agents: some are bubbling with excitement and for the first time ever, they feel like they're in an all-star lineup, while others mourn for the death of their craft, leaving them feeling hollow. I'm someone who works with agents every day, so I understand these sentiments. Rather than take sides, my hope is that I help you understand this in a more nuanced way. Too often we see two camps: people who love the joy… 9 Oct 2026
- Claude Haiku 5.5 is out 🎉 It's a cheap, fast and very capable small model. ~75% cheaper to run than Haiku 4.5 with an adjustable effort setting. My team loves it as a subagent alongside Opus 5.5. 7 Oct 2026
- You need both taste and judgement. You need taste to know what is good, and you need judgement to understand the price of quality. If you ask a model for a component or migration plan, it will create a lot of options, very quickly. Most of those drafts seem OK. So once you have a collection of work that meets the basic criteria, you can no longer use taste to select between them. You must choose. I like to think of this as two filters. There is the taste, and there is the judgement. And I think a lot of people mix those up. The first filter is taste. Taste is the feeling for what makes a thin… 7 Oct 2026
- To make your agent produce high-quality output, equip it with a way to check its work. Tests are part of that. Here are some of the ones I think are worth investing in. - Write end-to-end tests that simulate real user flows and act as the ground truth - Use property-based testing to specify what should never happen and generate thousands of test cases that attempt to break those constraints - If replacing a system, compare the old and new versions using randomly selected inputs - Require tests to run quickly and deterministically, which reinforces the agent to retry when the loop is slow or u… 5 Oct 2026
Link verified 23 Sept 2026. Recent items update automatically from the channel.
GitHub @addyosmani Their code.
Recent
- agent-skills — Production-grade engineering skills for AI coding agents. 1 Nov 2026
- critical — Extract & Inline Critical-path CSS in HTML pages 25 Oct 2026
- addyosmani — GitHub README 18 Oct 2026
Show 14 more
- clarity — Clarity - an Agent skill for clearer writing 5 Sept 2026
- skills.addy.ie — Agent Skills official site 27 Aug 2026
- web-quality-skills — Agent Skills for optimizing web quality based on Lighthouse and Core Web Vitals. 24 Aug 2026
- factory — A reference software factory for Claude Code and Codex 21 Aug 2026
- factory-demo — A worked TMDB demo of Factory's issue-to-verified-draft-PR software development loop. 19 Aug 2026
- save-image-as — Save Image As Chrome Extension 16 Aug 2026
- chatty — ChattyUI - your private AI chat for running LLMs in the browser 9 Aug 2026
- agent-engineer — Agent Engineer - a practical course for software engineers 16 Jul 2026
- rust-browser — Robin - build a web browser engine from scratch in Rust 24 Jun 2026
- adverse — Multi-agent adversarial code review for any coding agent 20 Jun 2026
- agent-house — Lighthouse for agents - score an agent run, then tell it how to get faster and cheaper. 20 Jun 2026
- webpack-lighthouse-plugin — A Webpack plugin for Lighthouse 1 May 2026
- agentic-seo — Agentic SEO / Agentic Engine Optimization (AEO) tools 12 Apr 2026
- oust — Extract URLs to stylesheets, scripts, links, images or HTML imports from HTML 2 Apr 2026
Link verified 23 Sept 2026. Recent items update automatically from the channel.
About the English under a post
Some people here publish in a language other than English. Where they do, this site shows a machine translation beneath the post, in this typeface — the site's own, not theirs.
The post itself is never changed, moved or hidden: what is set in the serif above is exactly what the person published, and it is what to quote them on. A translation can be wrong in ways that matter, especially about tone.
Only the post's own words are translated. A quoted post, a linked article and a belief on korrents.com are left in their original language.