This update broke my setup: opening Neovim afterwards showed a wall of attempt to call method 'start' (a nil value) error from nvim-treesitter along with a full stack trace.

This guide summarizes what I did to get it all working again.

{
  'nvim-treesitter/nvim-treesitter',
  branch = 'master', -- [!code --]
  branch = 'main', -- optional as main is default, but I like being explicit // [!code ++]
  ...
},

main = 'nvim-treesitter.configs', -- [!code --]
main = 'nvim-treesitter', -- [!code ++]

opts = {
  highlight = { enable = true }, -- [!code --]
  indent = { enable = true }, -- [!code --]
  -- ...
},
init = function()
  vim.api.nvim_create_autocmd('FileType', { -- [!code ++]
    callback = function() -- [!code ++]
      -- Enable treesitter highlighting and disable regex syntax // [!code ++]
      pcall(vim.treesitter.start) -- [!code ++]
      -- Enable treesitter-based indentation // [!code ++]
      vim.bo.indentexpr = "v:lua.require'nvim-treesitter'.indentexpr()" -- [!code ++]
    end, -- [!code ++]
  }) -- [!code ++]
  -- ...
end,

init = function()
  local ensureInstalled = {
    'lua', 'python', 'typescript',
    -- ... your parsers
  }
  local alreadyInstalled = require('nvim-treesitter.config').get_installed()
  local parsersToInstall = vim.iter(ensureInstalled)
    :filter(function(parser)
      return not vim.tbl_contains(alreadyInstalled, parser)
    end)
    :totable()
  require('nvim-treesitter').install(parsersToInstall)
end,

{
  'nvim-treesitter/nvim-treesitter-textobjects',
  branch = 'master', -- [!code --]
  branch = 'main', -- [!code ++]
  ...
}

Writing to think

Writing is thinking. To write well is to think clearly.

If people cannot write well, they cannot think well; and if they cannot think well, others will do their thinking for them.

I started writing more this year, and its biggest benefit so far has been the mental clarity. Putting my jumbled thoughts down into words has forced them to be clearer and more structured.

The writing process is iterative:

If you're scared of sharing your writing online, note that sharing is completely optional. We don't all have to be influencers. Write for yourself. Write a journal entry, write a technical post, write a romance novel, write about something you don't know much about. Most importantly, write whatever you want.

And if you happen to think that your writing could be helpful for others, share it.

Writing to do great work

We talk a lot about writing well for AI (i.e., prompt engineering) but not enough about writing well for humans.

The most advanced models are still terrible at writing, at least when it comes to writing like an actual human. This is our opportunity: I'm positive that being able to write well will help one's career, especially as AI-written content becomes more ubiquitous.

Think about the last time you read an obviously AI-generated post. It felt icky, didn't it? I hate it, and I know people hate it, too.

So when you write at work -- an email, Slack message, or presentation -- having people actually read and want to read your writing means your voice will be heard and your organizational influence will increase.

Writing to connect

Online discourse in tech is an echo chamber, and it loves fear- and hype-driven topics. For example, this is what shows up on TechCrunch's homepage right now as I'm writing this post:

11 AI or AI-related articles in this screenshot. They covered half of the "Latest News" articles and almost all the "Most Popular" articles. There is nothing wrong with covering AI, but this makes me feel like they're shoving AI down my throat.

Now let me introduce you to my oasis, my RSS feed:

This is a personal newsfeed of writings from my favorite humans like Vicky Boykis, Mitchell Hashimoto, Ty Sloane, and Sean Goedecke.

Sure, they probably all have written about AI, too, but what sets them apart is each person's "voice". Each has a distinct way of conveying their thoughts in words, each making me feel something different as a reader. I feel a sense of closeness and connectedness to each writer in a way that AI-generated content doesn't.

The content that we see online will be more and more AI-generated. AI slop is getting built on top of slop on top of more slop, and VCs are talking about an "agent-first" internet.

What's going to happen to our internet? My ideal version of the internet exists to connect us, to show us the humanity in each other, and I don't think that AI can do that in a genuine way. I'm afraid we'd lose the internet as we know it if we don't share our voices online at all.

Starting in 2026, communication has become the most important skill for software engineers.

It's not writing code, system designs, or having estoric knowledge of a programming language (i.e., Rust).

AI coding agents have gotten very, very good. A year ago, I'd reach out to Cursor hesitantly for MVPs or quick fixes. Today, I use Claude Code for almost all non-trivial programming tasks and have spent $500+ on it just last December.

AI talks online revolve much around the hard skils. Initially it was prompt tricks to accomplish X, then the best MCPs for Y, and so on. But with Opus 4.5, using vanilla Claude Code gets you 80% there. Even in the age of AI, the 80/20 rule still applies. So, what should engineers focus on?

One thing with coding agents is that the better the spec, the more in line they will be with the technical and business requirements. But getting a good spec is hard.

In real life, tickets rarely contain all the requirements. To do so, you might need to:

• Ask questions that reveal assumptions people didn't know they had
• Facilite trade-off discussions
• Push back on scope without burning bridges
• Make calls on things nobody thought to specify

Doing these things well used to be optional for individual contributors. Certain teams would enable engineers to thrive being an average communicator but excellent coder. Now, the non-coding parts are becoming a non-negotiable.

Software engineers are problem solvers. We believe that every problem has a solution, a "best practice". But working with people is messy.

Unfortunately, we won't be able to AI our way into better communication skills. Good communication requires empathy, and we can all use a little more of that in today's landscape.

Introduction

When I run Claude Code on a larger task, I often end up waiting on it while wanting to step away: coffee, a meeting, or a walk. I didn’t want to keep checking my laptop just to see whether it needed input or had finished.

Over the past month, I’ve been using a simple setup that lets me:

• keep Claude Code running on my Mac,
• attach to the same session from my phone,
• and get notified when it needs attention.

tmux new -s session_name

claude

tmux attach -t session_name

{
  "hooks": {
    "Notification": [
      {
        "matcher": "",
        "hooks": [
          {
            "type": "command",
            "command": "terminal-notifier -title \"🔔 Claude Code\" -message \"I need your input\" -sound default -activate 'com.googlecode.iterm2' && curl -d \"🔔 Claude Code: I need your input\" ntfy.sh/my-topic"
          }
        ]
      }
    ],
    "Stop": [
      {
        "matcher": "",
        "hooks": [
          {
            "type": "command",
            "command": "terminal-notifier -title \"✅ Claude Code\" -message \"I'm done!\" -sound default -activate 'com.googlecode.iterm2' && curl -d \"✅ Claude Code: I'm done!\" ntfy.sh/my-topic"
          }
        ]
      }
    ]
  },
}