Jumping on this... and pretty happy with my 0.0041

I found that looking away and then looking back at the screen and rapidly answering helped a lot. 🤷‍♀️

Did you intentionally pick a pastry that looks like a bunch of endless loops too?

Used Cursor in an interview for the first time.

It was SO refreshing to be able to talk through my thought process rather than getting stuck on syntax due to interview nerves.

And it helped me get to a working solution in the time alotted, which is always satisfying.

I was inspired by this and finally connected Claude Code to my notes. It's been handy for interview prep and collating all my past experiences into talking points.

I'm currently using Claude Code to turn my disparate, fragmented interview prep notes from the past 3 months into a story bank I can pull from for future interviews.

I'm am chuffed with how well it's working.

The 2026 F1 season is here 🏁

So I built another experimental custom <select> — this time featuring F1 drivers.

See the demo on @codepen.io
codepen.io/editor/cbols...

What has been your favorite optimization so far?

I've been thinking about doing the same.

WebHaptics – Haptic feedback for the mobile web. Haptic feedback for the mobile web.

this is so cool
haptics.lochie.me

Congrats!! That's huge!

Are you sending or receiving applications? I've been applying and it is rough. It makes me curious if hiring managers are just getting spammed with resumes.

My favorite part of @nerdy.dev's customizable select is the sticky section headers 😍

nerdy.dev/nice-select

Whew. It's amazing how quickly skills atrophy.

Just took a coding challenge where I had to use a useEffect for data-fetching and I am now so used to server-side data fetching in NextJS, it was more of a struggle than I'd like to admit.

You can vibe code your way to a working prototype. You cannot vibe code or one-shot your way to a competitive product that works at scale. The hard part isn't writing code; it's the architectural supervision.

Vibe prototyping isn't solving problems. It's creating new ones. It's easy to prototype and so everyone is prototyping, without really knowing why they are doing it. The resultant noise drowns out thoughtful work.

Every non-hype defense of LLMs starts with "you must already understand your work really well."

Vibe coding isn't being done by those people. It's being done by execs who are making a prototype because it is now easy to do.

This kills the systems of intention and maintenance necessary for good UX.

a reminder that if you're feeling like everyone is mad at you and you're screwing everything up, the fastest fix is to get on the phone or meet in person. chat/email loses too many cues (tone, body language, etc.). if you're already worried you're in trouble, it's easy to imagine anger when reading

I'm learning Elixir for a job interview, and the syntax is just close enough, and just different enough, to Javascript that I'm definitely going to require Post-It reminders up during the call. 🙃

You might have seen it in the past, but in case you haven't looked recently, GitHub has a couple really nice font sets:

Mona Sans & Hubot Sans:
github.com/mona-sans

Monaspace:
monaspace.githubnext.com

💯

I also feel this way when writing up any sort of idea proposal document. I clarify "Yes, AI helped format and polish this, but I stand by every line in it." Two years ago I never would have had to caveat my work like that.

I'm so bothered that by using the barn door, you cannot use any of the drawers.

The room configuration where this layout would be preferable over a regular cupboard door is niche, to say the least.

The only system that has been consistent throughout my productivity journey is a notebook of some sort.

Working a couple days during this void week because I know it will quiet enough that I can just code, uninterrupted. 😍

I've also struggled with people generating tickets, content, strategies without reading them first, then sharing them out for feedback. It's exhausting to read through a bunch of nonsense and decide whether to give actual feedback on it.

💯

"Almost anyone can prompt an LLM to generate a thousand-line patch and submit it for code review. That’s no longer valuable. What’s valuable is contributing code that is proven to work."

simonwillison.net/2025/Dec/18/...

Another update immediately code red situation

Completely agree. And that has been my experience so far.

It really can turbo-boost me as an eng, but now there are talks that one eng can do the work of what a whole team could do before...

Hotels with Bathroom Doors I’ve searched high and low and managed to find these hotels that have confirmed bathroom doors in all rooms. While every hotel I found is guaranteed to be a great stay, that isn’t why I…

This may become my primary hotel booking site from now on: bringbackdoors.com/hotels-with-...

100% It feels very different to how I worked before. It's more akin to when I managed people when working in marketing. Others would make the first draft, but I was ultimately responsible for the output quality.

“AI Delegated Development” feels accurate to me - that’s how we’re working with it at my company. We’re assigning tickets to AI for the first draft, then involving humans for the edits and reviews.

I fixed my issue with an answer from StackOverflow today, before I asked Claude Code. Please clap.

Loving the colors on that title slide