I considered it a possibility.

We recently released v0 for iOS, Vercel’s first mobile app.

Our newest model, Claude Opus 4.5, is available today.

Job aggregator

The internet is flooded with very impressive vibe-style coding demos, but in my day-to-day job at Filestage we rarely start codebases from scratch and we have to deal with hundreds of thousands of lines of code and their dependencies.

Lessons from over 2,500 repositories

Twice a year, I produce a big presentation exploring macro and strategic trends in the tech industry.

Less dependencies because of AI

After months of agentic coding frenzy, Twitter is still ablaze with discussions about MCP servers.

Today we're releasing GPT‑5.1 in the API platform, the next model in the GPT‑5 series that balances intelligence and speed for a wide range of agentic and coding tasks.

We've been making photo apps for iOS for long enough that we have gray hairs now, and using our experience we ran over 600 image generations to compare which AI models work best for which image edits.

I'll start off by saying that I am not at all an AI doomer - by any stretch.

From micro-manager to macro-manager: coding's asynchronous future

I want to take stock of how I'm currently using AI — at least, the cool stuff like LLMs that we call AI.

This is a list of writing and formatting conventions typical of AI chatbots such as ChatGPT, with real examples taken from Wikipedia articles and drafts.

Some concepts are easy to grasp in the abstract. Boiling water: apply heat and wait. Others you really need to try. You only think you understand how a bicycle works, until you learn to ride one.

A comparison of various SOTA generative image models on specific prompts and challenges with a strong emphasis placed on adherence.

Useful strategies

Tech debt and comprehension debt

From Shallow Loops to Deep Agents

Anthropic this morning introduced Claude Skills, a new pattern for making new abilities available to their models.

for Copilot Enterprise subscribers

Use the C.R.A.F.T.E.D. prompt framework to save time with common SWE tasks!

Anthropic released Claude Haiku 4.5 today, the cheapest member of the Claude 4.5 family that started with Sonnet 4.5 a couple of weeks ago.

I recently shipped a non-trivial Ghostty feature (unobtrusive macOS automatic updates) that was largely developed with AI.

As coding agents become more capable and long-running, they don't remove the human job of thinking.

Everyone is worried that AI will replace programmers. They’re missing the real revolution: AI is turning everyone into one.

All around AI

On AI, Creativity, and Craft

Results for 11 Formats

Most devs are underutilizing Claude Code. These 5 advanced strategies will turn it into your most powerful engineering teammate

Three months ago I said that AI changes everything.

Context is a critical but finite resource for AI agents. In this post, we explore strategies for effectively curating and managing the context that powers them.

Introducing Chrome DevTools MCP

If you ever watch someone coding, you might see them spending far more time staring into space than typing on their keyboard.

A two-week experiment building an app with full AI assistance, exploring both the promise and frustrations of LLM-based development workflows.

One of the tasks that I do most often is to review code. I've written a review command that asks an AI to review a code sample, and I've gotten a lot of value out of it.

How HubSpot Scaled AI Adoption

It seems pretty well-accepted that AI coding tools struggle with real production codebases.

We’re bringing the power of GitHub Copilot coding agent directly to your terminal. With GitHub Copilot CLI, you can work locally and synchronously with an AI agent that understands your code and GitHub context.

Using AI agents correctly is a process of reviewing code. If you’re good at reviewing code, you’ll be good at using tools like Claude Code, Codex, or the Copilot coding agent.

A new web search API is now available in Ollama.

The return of prompt engineering?

Confession: I’ve been using Claude Code to write all my code for me. And I think it’s making me worse at the thing I’ve loved doing for twelve years.

Claude can now create and edit Excel spreadsheets, documents, PowerPoint slide decks, and PDFs directly in Claude.ai and the desktop app.

Last month we launched Product Intelligence to help product teams handle incoming issues and put them in the right place in the backlog.

We're witnessing the final generation of people who translate ideas into code by hand.

I spend a decent amount of time reviewing code I didn't write.

101 architectural blueprints

Frameworks for LLM Agents have been surprisingly disappointing. I want to offer some principles for building agents based on our own trial & error, and explain why some tempting ideas are actually quite bad in practice.

A staff engineer's 6-week journey with Claude Code

I wrote a small tool for myself as my biannual routine check of where llms are currently at.

Fastly’s July 2025 survey of 791 developers found a notable difference in how much AI-generated code is making it into production.

Eight ways I use DeepWiki to speed up real coding work

Most models are excellent, so cost and latency dominate.

Ghostty + Claude Code + minimal tooling = maximum productivity.

This post details my experiences creating software with LLM coding agents, emphasizing that what you do with AI agents is ‘creation’, not just 'coding,' and sharing what worked for me.

This weekend at the YC Agents hackathon, we asked ourselves: what’s the weirdest way we could use a coding agent?

The Case for Writing and Coding Manually

What Actually Works in Production

From Zero to Onboarded (For Real)

Note This field is evolving quickly, and we will update this guide as new methods and recommendations arise.

The Universal Agent Configuration File

A while back, I decided to undertake a project to challenge myself: build a web search engine from scratch.

You don't have to spend hundreds of dollars on AI tools to get started.

Bringing the Browser into the Agent

This post covers the biggest risks (with real examples) and how to think about MCP securely.

v0.dev is now v0.app, the AI builder for everyone

I spent two months forcing an unnatural constraint on myself: what if I could only rely on agents writing code for a real-world production service?

Curing Your AI 10x Engineer Imposter Syndrome

Optimize the graph!

Agents need context to perform tasks. Context engineering is the art and science of filling the context window with just the right information at each step of an agent’s trajectory. In this post, I group context engineering into a few common strategies seen across many popular agents today.

What do docs look like if people don't read them?

Ollama’s new app is now available for macOS and Windows.

Practical lessons from building production agentic systems

It's about adopting AI in your engineering org.

The future of software development might just be jazz. Everyone improvising. Nobody following the sheet music.

I've built 12+ production AI agent systems across development, DevOps, and data operations. Here's why the current hype around autonomous agents is mathematically impossible and what actually works in production.

A view on the Metr paper about developer performance using AI tools

I used camelAI with a ClickHouse database of every HN story to do all analysis. You can use it for free with no login here to explore the data interactively yourself.

This guide covers how to enhance AI tools with up-to-date Astro knowledge and provides best practices for building Astro sites with AI assistance.

TLDR: Use claude max subscription

The non-obvious insights, tactics and workflows Shopify used to bring an ambitious memo to life

We’re entering a new phase of software engineering.

Last night I finished a transition from my old AI coding setup I've been using for a while to running Claude Code in Docker using VSCode's "Dev Container" feature. In this post I lay out a few of my thoughts on why I wanted to switch to something in the first place, and also a short guide for those who want to do the same.

notes for AI Engineers

How to win in the era of LLMs, AI Overviews, and GEO — before your traffic disappears

A Guide to llm

How are devs at AI startups and in Big Tech using AI tools, and what do they think of them? A broad overview of the state of play in tooling, with Anthropic, Google, Amazon, and others

Over the past few months, I’ve rediscovered my love for building software.

how to win by riding waves you didn't make.

Why Senior Developers Are More Valuable Than Ever

Upgrading dependencies is the kind of task everyone tries to avoid — especially when it involves a library like Storybook, known for frequent updates and breaking changes.

Part 1 is tough: The Death of the IDE

Keynote for AI Engineer World's Fair

Tracking the performance of the various coding agents.

Vibe Coding Isn’t Just a Vibe

A thorough analysis of how AI changes what is needed from a good API, with real-world examples + some speculations about the future.

How AI is chaging how we analyze data.

You get 1M tokens for cents for some LLMs.

Turn AI coding assistants into more reliable development partners

There’s a lot of talk about “vibe coding”, but is it just a vague term for prototyping, or could vibes change how we build software?

It’s performance-review season and I’m watching managers kneecap their careers. Gleefully they share the best prompts to have ChatGPT write their performance assessments - exactly the sort of shortcut that guarantees they’ll never get better at the job.

As a site owner, publisher, or creator, you may be wondering how to best succeed in our AI search experiences, such as AI Overviews and our new AI Mode.

A cloud-based software engineering agent that can work on many tasks in parallel, powered by codex-1.

I love these ones: the second opinion and the throwaway technique.

The AI-driven approach takes you from idea to app in minutes.

Four months ago, we asked Are LLMs making Stack Overflow irrelevant? Data at the time suggested that the answer is likely yes

Airbnb recently completed our first large-scale, LLM-driven code migration, updating nearly 3.5K React component test files from Enzyme to use React Testing Library (RTL) instead.

Roo Code — formerly Roo Cline, a fork of Cline — is an autonomous coding agent that I’ve been getting a lot of use from over the past few months.

Companies are hacking back because of AI

Agentic AI is introduced in Zed.

Nice future ahead of us ;)

tldr: lack of joy

Claude Code is a command line tool for agentic coding. This post covers tips and tricks that have proven effective for using Claude Code across various codebases, languages, and environments.

Developers and marketers are being told to add llms.txt files to their sites to help large language models (LLMs) understand their content. But what exactly is llms.txt, who’s using it, and—more importantly—should you care?

I enjoy shocking people by telling them I don’t use LLMs.

Writing code with LLMs is fundamentally different from other ways of programming. LLMs are often non-deterministic and always unpredictable. They have a capability that no other technology can match: the ability to interface with natural language. What does that mean for security?

tldr: We’re at AI’s halftime.

Announcing two major changes to Devin

We predict that the impact of superhuman AI over the next decade will be enormous, exceeding that of the Industrial Revolution.

Leveraging Experience for Better Results

I chose to make using AI a manual action, because I felt the slow loss of competence over time when I relied on it, and I recommend everyone to be cautious with making AI a key part of their workflow.

Rethinking How We Build AI Systems

A full-remote security startup nearly hired a backend engineer who doesn’t exist, after a candidate used an AI filter as an on-screen disguise in video interviews. Learnings for tech companies

In this post, we’ll showcase the potential of Kuzu-Wasm by building a fully in-browser chatbot that answers questions over LinkedIn data using an advanced retrieval technique.

I tested 12 LLMs — 10 running locally and 2 cloud-based — to assess their accuracy in generating alt-text for images.

Developers worldwide now get free AI-assisted coding help with the highest usage limits available, as well as code review assistance.

tl:dr; Brainstorm spec, then plan a plan, then execute using LLM codegen. Discrete loops. Then magic.

It’s a strange time to be a software engineer. Large language models are very good at writing code and rapidly getting better.

I’m not a natural “doomsayer.” But unfortunately, part of my job as an AI safety researcher is to think about the more troubling scenarios.

Why Firing Programmers for AI Will Destroy Everything

There is so much debate online about the usefulness of LLMs. While some people see giant leaps in productivity, others don’t see what the fuss is about. Every relevant HackerNews post now comes with a long thread of folks arguing back and forth. I call it the new Great Divide.

I'm hesitant to give this advice away for free, but I'm gonna push past it and share it anyway. You're using Cursor incorrectly.

Chunking PDFs—converting them into neat, machine-readable text chunks—is a major headache for any RAG systems. Both open-source and proprietary solutions exist, but none have truly achieved the ideal combination of accuracy, scalability, and cost-effectiveness.

There’s a lot of chatter in the media that software developers will soon lose their jobs to AI. I don’t buy it.

AI will change programming. I’m convinced of it now.

A field guide that also covers why we need to rethink our expectations, and what software engineering really is. A guest post by software engineer and engineering leader Addy Osmani

A lot has happened in the world of Large Language Models over the course of 2024. Here’s a review of things we figured out about the field in the past twelve months, plus my attempt at identifying key themes and pivotal moments.

A HN user asked me0 how I run LLMs locally with some specific questions, I'm documenting it here for everyone.

Claude is Anthropic’s AI, like ChatGPT but more capable. I was a casual user until the 23 October release (informally “Claude 3.6”), when it crossed a quality threshold I didn’t even know was there. It is really, really good. I have been using it a lot more since, and I got curious as to how much more.

Is it truly better than GPT-4o and 3.5 Sonnet?

Over the past year, we've worked with dozens of teams building large language model (LLM) agents across industries. Consistently, the most successful implementations weren't using complex frameworks or specialized libraries. Instead, they were building with simple, composable patterns.

Using transformers.js to summarize a text.

A framework to get accurate and relevant outputs from LLMs.

If you want to dig into local LLMs you should read this first.

Starting today, developers can build on our Grok foundation models using our newly released API.

New AI tool from Github

Get fast, timely answers with links to relevant web sources.

With this public preview, we’re unlocking the context of your code and collaborators—and taking the next step in infusing AI into every developer’s workflow.

In the next couple of decades, we will be able to do things that would have seemed like magic to our grandparents.

How tired are we collectively about AI? Are we over it yet? Can you even hear about it anymore? I'm sorry if you can't, because this essay will be entirely about it.

A proposal to standardise on using an /llms.txt file to provide information to help LLMs use a website at inference time.

We are enabling the rise of the AI engineer with GitHub Models–bringing the power of industry leading large and small language models to our more than 100 million users directly on GitHub.

This will go wild as soon as available for all.

A practical guide to building successful LLM products.

Welcome to the Copilot-native developer environment

Not (that) web dev relevant, but anyways a good read how AI and Blockchain can benefit from each other

Good quick read about that AI won't easily replace all the humans.

I know, the blog is not on https, but the article is really good, though very long!

The Microsoft Copilot app lets you ask questions, draft text, and generate images using AI.

Summarized in 10 points

AI + web3 = <3

AI created wallpapers in van Gogh style

In the evolving world of Large Language Models (LLMs), crafting effective prompts has become an essential skill. That's why I've created this collection, showcasing the most impactful prompts of the year across various intriguing domains.

The generated images look amazing.

Artificial intelligence learned how to generate text and art better than ever before, while computer scientists developed algorithms that solved long-standing problems.

Nothing to add...

Valuated at roughly $2 billion

List of leaked GPTs

Since it arrived a year ago, ChatGPT has been shipping at breakneck speed.

Pitch is such a wonderful tool, now with AI

If you want to build AI products that are unique, valuable, and fast, don't do what everybody else is doing.

LangChain Templates are the easiest and fastest way to build a production-ready LLM application.

Game on!

Instructions how to use Bing chat to create a logo from a drawing

Good tutorial to chat with a pdf

Vercels generative UI tool