Posts by sh03
IT'S HAPPENING. GITHUB, THE FIRST ENTERPRISE CLOUD SOLUTION TO REACH ZERO NINES RELIABILITY. Screenshot shows 95 incidents in 90 days, with 89.91% uptime.
this is your website on vibe coding.
Gemini hallucinating today as being April 2, 2025 "Liberation Day" where trump announced reciprocal tariffs.
๐
I think it also considers difficulty (in the top right as you guess right the difficulty increases).
I'm a dev and I do NOT approve this message.
What do you use LLMs for?
The intersection of people that love vibe coding and people that don't understand that each line of code is a liability is a perfect circle.
Yeah, LLMs and calculators are... basically the same.
That's why calculators run on a finite resource controlled by a bunch of companies that unapologetically stole copyrighted material and will grift us all into a global recession and will have to be bailed out by taxpayer money.
Wait...
5.4
is there a polymarket for Github reaching zero nines?
Pleeeeease Apple let me add web apps to my Home Screen in a way that lets them live in the App Library.
Friendly reminder that downloads are a useless metric.
upcoming android update solves sideloading and age verification in one fell swoop with new "wait 18 years" button
Code quality seems to matter in the same way that good writing in tv shows and movies matters: it's a nice to have only people passionate about the craft care about.
Slop marvel movies can make millions but well crafted stories like Andor struggle.
Maybe the world is set up for slop to win.
โNewer and larger models are smarter and more reliable. But they fail unexpectedly. There is no relationship between size and how failures change over time. AI-generated code creates 70% more issues than human code.โ tomtunguz.com/you-are-resp...
I don't think it's comparable to a junior developer. When it's good it seems more knowledgeable than a senior developer without the lessons a senior developer learned (e.g. better to remove than to add). When it's bad it's worse than a junior developer: it's closer to an adversarial entity.
Another example is I asked it to update to the latest tsdown version (that generates .mjs files instead of js ). I had to spoon-feed it the other two places where it needed to update references (package.json and dockerfile).
Somewhere in the middle of my prompts it also created a git repo in one of my subprojects (monorepo setup) causing me to spend 10 minutes trying to understand why the git diff in VSCode and locally was different.
A git diff showing codex replaced the entire vscode settings with some tailwind configs.
Another example: told it to fix a conflicting warning in VSCode (medium reasoning) it simply replaced the entire VSCode configuration file (tbf while also fixing the issue)
See skeets description
This is codex with gpt 5.4 (paid plan) on high reasoning: it just needed to remove the useViewer reference (it's just a react query) instead it left a useless useViewer call and removed unrelated code.
You guys are joking right? This is the best LLMs have to offer?
I've just bought Codex (the โฌ20/month plan) and tried it. It's now clear to me that pro AI folks are trolling me.
spent two weeks last month debugging an AI-generated PR that passed every test but silently broke edge cases a human would've caught just reading the requirements. clear sentence in an issue > 500 lines of generated code, every time
Probably bun
Just a quick follow-up on this, this is incredibly annoying
an LLM is a compiler in the same way that a slot machine is an ATM
Could you link to that Quora post? I can't find it.
A calming playlist for people that get ghosted at work
JavaScript's date object has been tricky for years, but that is changing.
The NEW Temporal API brings,
๐ Easily handle time zones
๐ Precise date math
๐ Parse ISO strings without errors
โ Durations, date ranges, and more.
Start experimenting ๐
developer.mozilla.org/en-US/docs/...
let me introduce you to "maneuverability", my Achilles' heel