The other day I used agents to convert a dollar field to a cents field. I’ve wanted to for a literal decade. But this is dangerous territory. You don’t want to 100x your users’ bills.

Using agents I felt safer than I ever would without.

Agents…

1. Investigated the necessary changes and their consequences
2. Confirmed that this would generally be a good idea – to standardize and to remove the risk if hitting float math troubles.
3. Made the necessary changes to code and unit tests.
4. Wrote the data and schema migrations.
5. Did manual testing in Chrome.
6. Watched the deploy.
7. Ran spot and full diagnostic checks on db backups of before/after to verify.

Not only can I work faster, I can also be extra thorough, be my best developer self.

What if AI is the best thing ever to happen to junior devs?

When I was coming up what held me back was not knowing and having to find the right apis, googling for errors, wasting time on the wrong abstractions, missing blind spots. All of this can be done at 10x the speed with AI now.

To be useful (ie. get a job) practice unblocking yourself. Be like an AI agent++. The further you can go without needing help or guidance the more useful you are. Has to be in the right direction of course or close to. So also practice business sense, actually listening to your lead, understanding the company direction, dabble in design, SEO, marketing. Just enough of everything and don’t tell yourself anything is “not for you”.

You can be more than twice the speed of an anti-AI-maxxing senior dev at a fraction of the price. That’s a great bargain.

Maybe it’s the best time ever to be a junior?

I wish I could justify using only Amp. With Opus 4.5 it is incredible.

I have it set to buy tokens for the same price as a Codex/Claude subscription every month but it only lasts a fraction of that 😭

Sponsorship deal if I put your logo on my wedding suit maybe?

Sadly I married before I knew we had options like that.

My mind is also starting to shift towards this.

Both because of increased belief in human ingenuity and because of decreased belief in human ability to downgrade comfort expectations.

Degradable plastic, lab meat, carbon removal tech. It’ll work out. I hope it will. It has to.

Started a thing 5 years ago. Spent a few days, realized the scope made it unrealistic, walked away.

Coding agents have made it feel within reach.

In the evenings, on the side, I am making 𝐩𝐫𝐨𝐠𝐫𝐞𝐬𝐬.

I bet you didn’t consider the wurst case

The only bad thing about Sketch is how impossible it is to google for shit and not hit tons of generic “sketch” stuff.

New Aerospace 0.20 can print the current window’s layout mode. Means we can add icons to Sketchybar.

I wonder if, in this new agentic coding world, the old we can’t justify both iOS and Android apps is obsolete? What if you now make the one you prefer first, but faster with AI, then when demand is there, make the agent of the moment clone it to the other platform – or maybe 85% there – and you just bring home the details?

They call these feeds “For You” but it’s not for you. It’s for them.

Powerful stuff as always from Jack Conte, CEO of Patreon.

Some open questions of course – like, how would Patreon try and filter stuff if it had as much as say X or YouTube?

Also, ads are actually kind of great for funding internet distributed digital stuff.

They aren’t the enemy, they’re just ads.

The reality is that creators want as many eyeballs and earholes as possible. And only so many want to - or are able to - pay. Bless their hearts, they’re the best. But because of the scaling of digital media, you don’t actually need everyone to pay.

My take is Patreon (and my own 10er!) are super important, wonderful services that work great on their own or alongside the ad-bound platforms.

Planning and finishing the project becomes the biggest task.
You wrote 100,000 lines of whatever? Done in an afternoon.

You came up with an original idea, carefully managed scope, beat edge cases and presented it with a clear description in a tidy package?

How did you do that?!

Incredible video on Reinforcement Learning.

• AI is a tool. Best results come from Human plus AI (still).
• Games make great visualizations of RL improvement.
• Humans are incredible.

Maybe if I get an MCP that’ll tell me to drop it: it’s marketing week

🔥 Agent quick tip: Review changes in vim.

1. Pipe your git diff into vim with vimreview.
2. Add your review comments in the comfort of your favorite editor as basic # comments.
3. :wq and paste into your agent session.

Get it:

$ brew install moreutils

~/.zshrc:

vimreview() {
  {
    echo "# [NOTE]\n# DIFF REVIEW:\n# The following is our current changes as a git diff with my comments added. Comments are denoted by a leading #\n# Please review and discuss or process my comments.\n# [/NOTE]\n"; \
    git diff
  } | \
  vipe --suffix diff | \
  pbcopy
}

🔥 Agent quick tip: Make an agent summarize a change from the outside in so the diff tells a story.

Large diffs can be hard to understand when they’re just presented from top to bottom, alphabetically by filename. If you know the big picture, it’s easier to understand the parts.

Amp’s -x param is great for this.
Pipe it to glow for pretty terminal markdown tui.

After using iPhone minis for 4-5 years, the battery in this new phone feels infinite.

Petition to change the spelling of default to deafult (because my fingers will not learn)

Sign here
​________________

Remember how much time we spent googling just a year or two ago and now we can just type “make my tmux.conf support italic and squiggly underlines” and it works

I’m trying to figure out how to use generative AI as a designer without feeling like shit. I am fascinated with what it can do, impressed and repulsed by what it makes, and distrustful of its owners. I am deeply ambivalent about it all. The believers demand devotion, the critics demand abstinence, and to see AI as just another technology is to be a heretic twice over.

Today, I’d like to try to open things up a bit. I want to frame the technology more like an instrument, and get away from GenAI as an intelligence, an ideology, a tool, a crutch, or a weapon. I find the instrument framing more appealing as a person who has spent decades honing a set of skills. I want a way of working that relies on my capabilities and discernment rather than something so amorphous and transient as taste. (If taste exists in technology, it needs to be smuggled in.)

Frank Chimero just dropped the first essay I read top-to-bottom in what feels like years.

Really good shit about how to feel, as an artist, a creator, about “AI”.

I wouldn’t want an irregular AI in my bank app, but in a creative workflow, hallucinating feels like the point of it all.

I’m seeing a trend: Some poor, misguided person passes on an AI generated answer without adding any direct human effort.

The intention is to be helpful, but we all have access to the same tools, more or less, so merely passing on a slopjob isn’t helpful. It’s disrespectful.

Passing on AI content without judgment signals both:

1. I think this issue/problem/feature is simple enough that it can be done just by asking a perpetual internet machine, and…
2. I don’t think you were smart enough to consider asking AI for guidance on your own.

The same wall of text, human written, would at least show some real effort in understanding the whole problem. Even if misguided, it is a genuine sign that someone spent time on understanding and formulating a well-structured, researched answer. Proof of work, one might say.

[Requiring AI use disclosure] is to help maintainers assess how much attention to give a PR. While we aren’t obligated to in any way, I try to assist inexperienced contributors and coach them to the finish line, because getting a PR accepted is an achievement to be proud of. But if it’s just an AI on the other side, I don’t need to put in this effort, and it’s rude to trick me into doing so.

@mitchellh requiring disclosure for AI assisted PRs to Ghostty.

Copy/pasting code into an AI chat platform then copy/pasting the response to this issue thread is, in my opinion, not at all helpful. […] If you provide some professional insight based on your own experience and experimentation in addition to hitting Ctrl+C/Ctrl+V, that may in fact be useful, but otherwise I would refrain from doing that again on GitHub.

@viscount-monty providing responding to a user throwing a complex engineering to ChatGPT then blindly commenting the answer.

“Please, just stop sending me AI videos of Dad [late comedian Robin Williams] ,” Zelda wrote. “Stop believing I wanna see it or that I’ll understand, I don’t and I won’t. […] It’s dumb, it’s a waste of time and energy, and believe me, it’s NOT what he’d want. […] To watch the legacies of real people be condensed down to ‘this vaguely looks and sounds like them so that’s enough’, just so other people can churn out horrible TikTok slop puppeteering them is maddening,” she continued. “You’re not making art, you’re making disgusting, over-processed hotdogs out of the lives of human beings, out of the history of art and music, and then shoving them down someone else’s throat hoping they’ll give you a little thumbs up and like it. Gross.”

Zelda Williams on people sending her AI clips of her dad, the late Robin Williams. If she wanted those, she could make them herself.