Your manager is in serious need of lightening up π
Your manager is in serious need of lightening up π
This is the kind of content Iβm here for π
New blog post: External import maps, today!
lea.verou.me/blog/2026/ex...
I suspect there is a more succinct way to achieve the same result (I never claimed to be a prompt engineering expert!), but this works.
[2/2]
This applies even to direct requests to revert or simplify. Still evaluate whether your original approach was better. The user may be missing important context. If there was a solid reasoning you suggested that approach, push back with reasoning instead of silently complying.β
2. For pushback:
βUser should not have to ask you for your opinion explicitly. Always evaluate what the user is asking you to do, and voice your concerns before proceeding if you donβt think it's a good idea. If possible, propose a better solution, but you can voice concerns even without one.[1/2]
Since many asked, the instructions from my CLAUDE.md were:
1. For modern web tech:
Always use the most modern HTML, CSS, JS you can. If supported in every browser, even if it's only the most recent version, use it.
Overshoot is intentional: βMost recentβ as far as Claude is concerned is May 2025!
Claude has a lot more context than me since theyβve just looked at the whole codebase. Even if Iβve written all the code, it may have been years. Itβs quite common that Iβm missing important context when I think theyβre wrong. Anecdotally, about 50% of the time Claude pushes back, I agree with them.
Note this may be unnecessarily verbose, Iβm not a prompt engineering expert π
Also, itβs part of my Claude .md, not the prompt.
[2/2] This applies even to direct requests. Still evaluate whether your original approach was better. The user may be missing important context. If there was a solid reason you suggested that approach (e.g. avoiding code duplication), push back with reasoning instead of silently complying.β
βUser should not have to ask you for your opinion explicitly. Always evaluate what the user is asking you to do, and voice your concerns before proceeding if you donβt think it's a good idea. If possible, propose a better solution, but you can voice concerns even without one. [1/2]
Claude .md!
Where is this checkbox from?
"Always use the most modern HTML, CSS, and JS you can. If it's supported in every browser, even if it's only the most recent version, use it."
The overshoot is intentional β the "most recent version" as far as Claude is concerned is what was most recent in May 2025.
I generally do, sorry for the oversight! Here you go (unfortunately I can't edit the OP): bsky.app/profile/lea....
Me: ok yeah I think we'll need VTs after all Claude: Before going with VTs β have you considered @property? Registering --tilt as <angle> would make it natively transitionable, and since both the parent and children read the same variable, they'd stay in sync (no double-animation issue like with raw transition: transform): @property --tilt { syntax: "<angle>"; initial-value: deg; inherits: true; Then just transition: --tilt .4s ease-out on .equation-display. Much simpler than VTs for this specific case. Or do you want VTs because other things should animate too (factors moving, etc.)?
With alt text (posted too soon!):
Iβve instructed Claude to question anything I ask it to do and to always use modern web technologies, and itβs glorious. π
- Some way to still have both ends on the pointer but clip the one we don't need
- Have two pointers and find a way to hide one of them (suboptimal because it would take up two pseudos, but not the end of the world)
Some ideas that led me nowhere but may lead you somewhere:
- drawing the arrow with backgrounds and then using some clever combination of background-attachment + positions to move the background.
- some way to trigger a *different* CQ on the parent when it flips
Pretty much, yes.
If youβre distributing a component as part of a design system, it's suboptimal to have it break if consumers style it.
This is a great writeup as well (but no solution): frontendmasters.com/blog/perfect...
(Yes I know that anchored()ΒΉ will fix this, but that's no help right now if you need Safari support)
ΒΉ caniuse.com/wf-container...
Has anyone figured a bulletproof way to use anchor positioning w/ a tooltip pointer? @css-only.dev got farΒΉΒ²Β³ but these all depend on the tooltip being opaque.
Just got nerd sniped trying to figure this out!
ΒΉ css-tip.com/tooltip-anch...
Β² css-tip.com/tooltip-anch...
Β³ css-tip.com/tooltip-anch...
A big advantage of using LLM chatbots I have not seen mentioned anywhere else: you develop an βAI-darβ: you can detect when someone is using LLMs to talk to you.
Yes β for now.
I feel that to my core. My condolences to you too!
I do the same with booze that goes into food! Eg red wine in tomato sauces, Metaxa in meat sauces etc
π€£π€£π€£
You sir, just won the internets. π
Vibe cooking!
Happens to me all the time.
nondeterministic, nonreproducible, but oh so enjoyable, and sometimes the end result is better than expected π€
A tragedy in two acts:
I just cooked the best broccoli π₯¦ of my life.
It had all the good stuff: guanciale, π§, eggs, parmesan, anchovy sauce, MSGβ¦
It was SO GOOD even my 6 year old loved it! Thatβs HUGE.
But it was all by eye β I have NO idea what proportions I used so I can never recreate it πππ