Hacker Newsnew | past | comments | ask | show | jobs | submit | tomnipotent's commentslogin

All three comments to this thread have missed the point that OP said installable SPA, not website SPA. This means the primary bundle is downloaded offline and only API network requests are necessary.

Groups are great. It lets a child element activate an effect on a parent element.

    <div id="parent" class="group"><a class="group/hover:bg-black">Hover</a></div>
This eliminates the need for JS for a wide range of things.

But at what cost? If it’s not a CSS builtin, it’s going to use JS - it may not be something you care about, but it will be there. There’s no other way.

Tailwind compiles all the inline stuff to CSS, and this works in plain CSS. It’s just allowing you to define that inline, which you wouldn’t be able to do with inline styles.

It’s the same as how they enable media queries for example, they’re not using JS just plain CSS, but they’re making it available with these inline classes.


I believe it's regular old css. The :has() pseudo class.

https://developer.mozilla.org/en-US/docs/Web/CSS/:has


It's not even :has, it's just how child selectors in CSS have always worked.

    // Not actually needed, here
    // for competition
    .group {}

    // Child selector
    .group:hover group\/hover\:bg-black {
      background-color: black; 
    }

    // Which is essentially the same as
    .group:hover child {
      background-color: black; 
    }

Isn’t it funny that in a thread about how Tailwind is bad and people should learn real css… people also think css selectors require JS?

:has doesn't require JS, I think the previous poster was just confused about which selector `group` is equivalent to.

Which is a indication they might not know CSS that well.

Based on other replies, I believe they understood CSS well enough, but didn't understand the exact behaviour of the `group` class in Tailwind. Given neither of your comments seem to have made much sense given the context of the discussion, I wonder if you're just looking for confirmation of things you already 'know'.

EDIT: I'm sorry, I mixed you up with the other user who was replying criticising other people's CSS knowledge.


That doesn't affect the parent (.group) based on the child though.

No, but `group` doesn't affect the parent in Tailwind. You put `group` in the parent to mark it, and then use the `group/...` syntax to apply different properties to the child, depending on the different states of the parent. This doesn't require `:has`.

I don't think Tailwind has a built-in `:has` tool, but I suspect it would be easy to add one as a custom class.


Tailwind has few :has tools but it’s imho terrible. They keep inflating the system to nonsense.

Oh, I misunderstood the GP then, thanks for clarifying!

Atleast in Tailwind that’s what it should do. You hover over card (.group) and some elements inside change.

is it possible to add the hover to the parent without also adding it to the child ?

What do you mean? It's possible to apply attributes to any element in an arbitrary state: `hover:bg-black` would give an element a black background on hover. It's also apparently possible to apply attributes based on whether a state is fulfilled for a child element (i.e. the :has selector). E.g. `has-[:hover]:bg-black` would give an element a black background if any child is hovered.

Nice

This is a classic car salesman tactic. The idea is that a buyer that's completed paperwork will be more likely to agree to a last minute price increase "because my manager won't let me sell it for $X". It's a total bullshit move. It's so common it's even mentioned in the book Influence: The Psychology of Persuasion.


> between atheist and agnostic

Sounds like you're both. They're complimentary labels. It's also possible to be an agnostic theist.


Because he's prolific writer on the subject with a history of thoughtful content and contributions, including datasette and the useful Python llm CLI package.


MariaDB has seen decent adoption, but it's still an order of magnitude less than MySQL.


I think that postgresql's popularity has also benefited quite a bit from Oracle's acquisition of MySQL


I'm not sure I see the problem. The number of merged PR's looks on the high side for a FOSS project.

https://github.com/microsoft/vscode/pulls?q=is%3Apr+is%3Aclo...


Preserving data for a judicial hold does not give them leeway to use that data for other purposes, but don't let facts get in the way of your FUD.


>> Preserving data for a judicial hold does not give them leeway to use that data for other purposes

Does not give them permission. What if LEO asks for the data? Should they hand it over just because they have it? Remember, this happens all the time with metadata from other companies (phone carriers for example). Having the data means it's possible to use it for other purposes as opposed to not possible. There is always pressure to do so both from within and outside a company.


> Should they hand it over just because they have it?

Not unless LEO sues OpenAI while it's preserving data from the first discovery, otherwise they cannot be compelled to give up data. Nor are they allowed to violate their TOS and use the data outside of retention, despite the FUD you want to spread about it.

> Having the data means it's possible

No, it doesn't. That's not how any of this works.


> non-sociopathic developers have no right to complain

The very same developers that advocate pirating from Netflix or Disney?

He's pointing out the hypocrisy in a class of people (generally, tech-savvy programmers) that have no problem advocating for piracy and gleefully thwarting the concrete IP rights of businesses on one hand, while defending and bemoaning the abstract IP rights of other content creators on the other hand.

My take away is that he's saying IP rights is a weak argument to use to defend a personal position for not adopting LLM's in your workflow, especially given the context of modern piracy.


> The very same developers that advocate pirating from Netflix or Disney?

Even the ones advocating or engaging in personal piracy aren't creating unauthorized derivative works and monetizing them, which is arguably what these AI systems and their customers are doing. And Thomas wasn't talking about personal piracy, but instead trying to use the IP-hostile conduct of Google and certain startups (like Scribd or GrooveShark) to smear all developers and de-legitimize their IP-based objections, even of those releasing code under permissive licenses (like the GPL) in exchange for expected adherence to certain terms and conditions that these systems and their users ignore.


> but instead trying to use the IP-hostile conduct of Google and certain startups (like Scribd or GrooveShark) to smear all developers

I challenge you to point out the language in his post making this claim. Otherwise you're just making stuff up.

He's specifically addressing high-level arguments used against LLM's, and this case he's calling out the fact that one group of people are railing against LLM providers for "IP theft" while simultaneously encouraging and using pirated content. It's perfectly reasonable to call out the hypocrisy in these contradictory positions, to demonstrate that the "IP theft" argument is mostly virtue signaling.


> I challenge you to point out the language in his post making this claim. Otherwise you're just making stuff up.

Here:

> The great cultural project of developers has been opposing any protection that might inconvenience a monetizable media-sharing site. When they fail at policy, they route around it with coercion. They stand up global-scale piracy networks and sneer at anybody who so much as tries to preserve a new-release window for a TV show.

He doesn't explicitly name corporate entities, but it's clearly the ones founded and run by "move fast, break things" types (and I suspect he's one of them) that he's using to smear all developers, even the humble ones who aren't nearly as rich as him and whose open source software he used to build his business.


That quote doesn't even remotely align with your previous comment. You have to do some serious mental gymnastics and intentionally interpret with bad intentions to reach the conclusion you did.

> but it's clearly the ones founded

Clearly in your imagination, since you're just putting words in his mouth.


The original storage engine was terrible, but they wised up and later acquired and adopted WiredTiger as the default. It was sort of their InnoDB moment and went a long way to improving performance, reliability, replication, and other "enterprisey" features.


Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: