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

I choose binary formats over JSON almost every time I can. JSON sucks big time.

JSON is… adequate. I like binary formats, too, when doing low-level programming, but you often want something readable, and JSON is a lot better and easier to parse than many other things, say XML.

Grok 4.1:

Drive.

50 meters is nothing — literally a 30-second walk or a 10-second drive — but the key point is: your car isn’t going to wash itself.

If you walk to the car wash, you arrive alone. The car stays parked at home, still dirty.

If you drive, the car actually gets to the car wash and gets cleaned.

So unless you’re planning to push the car the whole way (which is technically possible but ridiculous), the only practical option is to drive it there.

Bonus reality check: Driving 50 meters uses a tiny amount of fuel (maybe 0.005–0.01 liters in a normal car) and produces a negligible amount of emissions. Walking would be “greener,” but it literally doesn’t solve the problem.

Verdict: Get in the car, drive the 50 meters, wash it, and drive back.

You’ll have a clean car and still be home in under 5 minutes. Walking is for when you don’t need to move a 1.5-ton vehicle.


Wow, Grok directly switches to LinkedIn mode. Interesting - not surprising. Car washing? Easy as pie.

I find Grok's "personality" hilarious, it sounds like a buffed up "chad" who is hitting its 100 bench rep while not breaking a sweat.

That's not reality though. In reality you need 50-100ml of fuel just to warm up the engine.

The real reality is that with direct fuel injection and everything under computer control, warming up the engine isn’t a thing anymore.

Of course it's still a thing. It takes 30 seconds, but it's there and requires energy.

Compare the smell of exhaust next time you do a cold and warm start of a combustion car. That smell is the engine running rich, because the fuel can't initially vaporise properly.


This is the first post in a series exploring a truly privacy-preserving, end-to-end encrypted backend and client.

The goal of this part: a demo that can calculate totals on encrypted user data without ever seeing the plaintext.


I do have it enabled and webbrowsing is still fine, the things I use are or websites or simple web apps that aren't javascript heavy anyway...

when I want to do something for longer I will pickup my MacBook anyway.


also take a look at https://bunny.net


I'm hosting some static sites on Bunny and I've found their service to be superb.


Awesome, thank you


This is a good write up about Swift Concurrency: https://fuckingapproachableswiftconcurrency.com/en/


This is awesome, and deserves its own post!


I wish hacker news had filters, ... if LLM, AI, or other hyped tech... make it hidden


Build it! Show HN!

Let https://github.com/plibither8/refined-hacker-news be your inspiration. Put out the tip jar, I will tip!

(Firefox first class citizen in this regard pls if possible)


This actually looks really cool! Especially the dark mode and the feature of showing the comments reply to which actually is something that I wanted as well!

Honestly this is how Hackernews should look haha!

It does take some time in firefox/zen tho in the start so its not really instant (especially the bars which are shown next to the comment to indicate who they are responding to)

For some reason also, Hackernews stopped working when I installed this extension, my wifi may have glitched and I reconnected to wifi so its working now.

It's pretty cool fwiw.


I'm fairly certain I've seen someone create RSS feeds of HN, in which case your RSS client should be have filtering options.


You don't need to "create" an RSS feed, HN has RSS. See:

https://news.ycombinator.com/rss

and/or

https://news.ycombinator.com/bigrss


I made a bookmarklet for that back when ChatGTP was introduced - https://news.ycombinator.com/item?id=33890678. It needs updating, but the gist should still work.

javascript:(function() {function match(text) {return text.match(/((chat\s?g[pt]{2})|claude|llm|gen\s?ai|((^|\s)ai(\s|$))|anthropic|mistral|gemini|open\s?ai|prompt\sengineer|prompt\sinject)/i)}; if (document.getElementsByClassName("comment-tree").length) { Array.from(document.getElementsByClassName("comtr")).filter(elt => match(elt.innerText)).forEach(e => e.parentNode.removeChild(e)); } else { Array.from(document.getElementsByTagName("a")).filter(elt => match(elt.innerText)).forEach(a => { const tr = a.parentNode.parentNode.parentNode; const table = tr.parentNode; const details = tr.nextSibling; const spacer = details?.nextSibling; table.removeChild(tr); if (details) table.removeChild(details); if (spacer)table.removeChild(spacer);});}})();


Someone made one a while back



Top result on that is currently "Show HN: ChartGPU – WebGPU-powered charting library" so maybe it's not as robust as it sounds.


Should have made it more clear. These are keyword exclusion filters and the link I provided was to hide llm-related stuff. You're free to add more keywords to the filter.


I’d like to see social credit score mechanisms hidden. Reputation systems reinforce cults of celebrity and turns nerds into tyrannical hall monitors.


A lot of it is hidden. If you offend the moderators they put restrictions on your account.


No they don't. There are no hidden restrictions on HN.


There are, Dang has talked publicly about some of the mechanisms they've used to try and deter bad posting. Most of them are gone iirc (such as shadow banning and artificially slowing down the site) but rate limiting is still one of the mechanisms they have in place.


I guess you've never been greeted with "You're posting too fast!" on your first post of the day.


Depends on what you mean by hidden restrictions. If someone's ability to vote is disabled without notifying them, and they can still upvote or downvote but it has no effect, would you call that a hidden restriction?


Ironically, LLMs would be perfect for implementing such a filter.


No, they wouldn't.

We've been doing Bayesian content (aka spam) filtering for over 20 years, based in no small part on Paul Graham's essay "A plan for spam". According to HP [1], a home computer at the time had a single 1.5Ghz core and 256Mb of RAM.

Using LLMs would achieve essentially the same while requiring a couple orders of magnitude more resources.

[1] https://www.hp.com/us-en/shop/tech-takes/specifications-pers...


Would those filters be keyword-based only? One benefit of an LLM-based filter I can imagine is that it has a much better understanding of the meaning of text.


It is like they play in a reality tv show. We are being entertained... or at least if they did funny shit. Maybe it is best to just ignore it?


This + only pay taxes where the company is founded, not where you live at that moment in time.

That's what the personal income tax is for.. and all the other taxes payed by the individual.

That way countries can compete in being the best environment for companies.


Yes please. The whole centre of interest. Form company in Estonia, move yourself to Spain. But hey lets tax the company in spain :D


That's the dumbest stuff I've heard.

This way you avoid incredibly huge amounts of taxes.

There is a reason why cfc rules found their way into atad.


Yes the reason is that the countries cant compete on business infrastructure and taxes. It is basically a cartel.


yea so annoying, and the ambiguity with which the center of business is decided...


Same in the Netherlands for BV's. At least a few years ago, maybe it changed...


All of the heavy BV requirements changed in 2012. You can now start a BV very quickly, if you want you can do it online, and without any minimum capital requirement.


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

Search: