Hacker News
Ask HN: How is AI-assisted coding going for you professionally?
Ask HN: How is AI-assisted coding going for you professionally?
Show HN: Hacker News archive (47M+ items, 11.6GB) as Parquet, updated every 5m
Show HN: Hacker News archive (47M+ items, 11.6GB) as Parquet, updated every 5m
"This is not the computer for you"
“This is not the computer for you”
Lies I was told about collaborative editing, Part 2: Why we don’t use Yjs
Levels of Agentic Engineering
Levels of Agentic Engineering
An opinionated take on how to do important research that matters
Okmain: How to pick an OK main colour of an image
An opinionated take on how to do important research that matters
Willingness to look stupid
Willingness to look stupid
X / Twitter
In Oct. 2025, I created https://t.co/aiZgosvMTR, a browser agent extension for Chrome.
Here it is beating the crap out of OpenAI Atlas by cheating.
It is now OSS. Go forth, and fork/remix/make it your own.
https://t.co/7YNmK3GfDZ https://t.co/Lo2vybzQPo
@denicmarko https://t.co/S9wB4Sr8nG is another really good free alternative
From 774 messy long notes to 5,805 perfectly organized single-concept ones in 51 hours. GLM-5 did the job.
https://t.co/nH4dbIbi18
A free, open-source Screen Studio alternative.
Supports MacOS, Windows and Linux. https://t.co/XP9YiZqg2X
Introducing element-source
Get the source file location of any DOM (or native!) element
Works w/ React, Vue, Svelte, and Solid https://t.co/fMisgePGU7
And the most important part: we open sourced the /autoresearch plugin for pi. Just tell it what you want, it will do the rest.
https://t.co/C2H6kxR7jG
@ctatedev Cool! There's a workgroup (https://t.co/2b51KojsEY) dedicated to establish interop between diff infinite canvas tools (tldraw, obsidian canvas, excalidraw, etc). Can easily imagine bringing 2D canvas to 3D via json-render
Expanding our code to canvas Figma MCP server capabilities with 5 new partners: Cursor, Warp, Factory, Augment Code, and Firebender. https://t.co/SYXFIm8UPR
We built a better alternative to json-render & A2UI
Up to 3x faster rendering & 67% fewer tokens. Far fewer broken outputs.
We've been shipping Generative UI interfaces to 10,000+ developers for the past year. We know where JSON-based rendering breaks in production.
So we built a new format - one that works with how models actually generate, not against it.
And today we're open sourcing it: openui [dot] com
Some tips from the https://t.co/RP4FIdUrS8 skill file that you can feed to your favorite coding agents. https://t.co/sHa9i3cRzx
Developers are now shipping way more by just prompting the feature they want
Expecting design to come in and "just style" things
For simple things this is fine, but for most things that'll have an impact, this doesn't work
You are taking away the hard part, which is exploration and understanding what users really want
Instead you just ship it! Cos hey, look at us, we are shipping so much more with AI
Your users of course get used to your changes and build habits
Then you realize that, actually, we've implemented this wrong
Now you have two choices:
- Rethink the flow, reorganize the UI, maybe deprioritize/remove things
- Or more likely, just leave it
And this is where dealing with design slop is so much worse than dev slop
Because if your product has reasonable usage, and you do this a few times, you'll burn whatever trust your users have in you
Or if you just leave it, your product grows into a bloated mess
And it's just sitting there, waiting to be replaced by something newer and simpler
and if you need an OSS variant of the same thing - i've built hodor. Uses the excellent @badlogicgames 's pi-sdk, supports gitlab/github, and you can use any model (even bedrock with custom profiles to track costs)!
You can also load custom "skills" per repo/CI job - enforce your own conventions, security rules, or review checklists.
This is what we @zerodha have been using for the past few weeks and it has caught some really non-obvious subtle edge cases/bugs.
https://t.co/EInC9k7JL7