HNHacker News
TopNewBestAskShowJobs

klausa

3,943 karma · joined August 15, 2011

contact: jan@klausa.pl
submissionscomments
klausa··on Apple Foundation Models
This is support for a new framework that ships with reality/mac/iPad/watch/tv/iOS 27 (and that they've promised to open-source later in the year, so presumably you'll also be able to lean on this if you ship Swift on your backend).

The framework's whole deal is that it lets you use the same API to target either the device built-in models, the Apple-hosted online models (Private Cloud Computer), or write your own shims to call out to arbitrarily hosted online models.

You can then dynamically route your calls to a different kind of model/provider, using system APIs, without having to write your own abstraction layer over "I want to use local model for this, but I want to use Claude for that", or having to integrate your own API integration with Anthropic/OpenAI APIs.

It abstracts things like tool calling in one place; and has a bunch of other niceties/oddities (it keeps the same "transcript" going, even if you dynamically switch providers/models during a session) and some other things.

klausa··on Apple Foundation Models
The apps can use the system provided on-device model using the same framework and APIs; but there's no affordances to deduplicate custom models between apps.
klausa··on MiMo Code is now released and open-source
This is not exactly what you’re asking about, but I started learning Japanese when I was in the middle of playing Cyberpunk 2077 (for unrelated reasons); and I gotta tell you; realizing that 98% of the Japanese text everywhere in the game was just “hotel” or “karaoke” definitely took away some charm from it.
klausa··on Sony Launches Bravia 9 II and Bravia 7 II with 'True RGB'
You're confusing "benefits of playing 4K-mastered media" and "benefits of having more pixels on the panel"; which is what the original point was about.
klausa··on Sony Launches Bravia 9 II and Bravia 7 II with 'True RGB'
Yes, unless you wanna go down the line of "well actually Apple screens are not that good".

This is a semi-sponsored video from a scammy company that makes glossy displays, so you might want to take it with a grain of salt, but it talks about a real effect with (most, again, Apple's etching is _much_ better, but still _noticeable to me_) matte coatings: https://youtu.be/3mTV1TOblbA?t=124

There are people taking super-zoomed-in photos of their monitors all over reddit too, so you can judge how much haze/blurriness they add to the image:

https://www.reddit.com/r/ultrawidemasterrace/comments/18sb63... https://www.reddit.com/r/OLED_Gaming/comments/1jf2u54/the_di...

klausa··on Sony Launches Bravia 9 II and Bravia 7 II with 'True RGB'
I’m at 65” and less than 2m away, and I absolutely can tell when text is not rendered at native resolution, which is why I’m also confident I’d be able to notice the matte coating too.

(I am also probably like three standard deviations _more annoyed_ by the blurriness than an average person, I’m more than willing to believe that an average person wouldn’t be able to tell, or at least wouldn’t be bothered by it anywhere close to the degree that I am.)

klausa··on Sony Launches Bravia 9 II and Bravia 7 II with 'True RGB'
> Because matte vs glossy is a separate discussion from cheap vs premium.

My whole complaint is that it increasingly isn’t!

If I wanted a “premium” variant of Steam Deck, I had no choice to opt out of matte coating.

If I want the best TV panel possible from Sony, it only comes with the matte coating.

I’m fine with the option for matte being there for folks who like this! I just want to be able to opt-out of it, if I still want the “best” among other axes.

klausa··on Sony Launches Bravia 9 II and Bravia 7 II with 'True RGB'
I’d now be actually curious to do a Pepsi-style challenge to test this on myself.

I give myself a ~low 40% chance of being able to tell within a minute when watching a TV show/movie _without subtitles_, maybe 60%ish with subtitles, and maybe low 80% if I was playing a game with any amount of text rendered on the screen.

klausa··on Sony Launches Bravia 9 II and Bravia 7 II with 'True RGB'
> Not every TV is used by two people in a room devoid of lighting. Friends will come over, other things will be watched. Some people are very bothered when the ceiling lighting or sunlight from the window "alters" the image.

> Enthusiasts are an interesting bunch. Some features loved by others are grave flaws for others. One can't make everyone happy, so matte screens are an acceptable compromise for it, AFAIK.

You know what, fair enough. I don't mind these; but I do _hate_ the blurriness that the matte coatings introduce.

Now, I wouldn't call my wife a TV tech enthusiast, but I do know that she's bothered by the reflections occasionally when I don't mind, and I know with her eyesight she straight up physically cannot notice the extra blurriness that the coatings bring.

klausa··on Sony Launches Bravia 9 II and Bravia 7 II with 'True RGB'
Is it fully matte, like on a desktop monitor, or is it some "reflection reduction" semi-gloss/semi-matter whatever magical coating?
klausa··on Sony Launches Bravia 9 II and Bravia 7 II with 'True RGB'
What's with the recent trend of putting matte coatings on the most "premium" devices?

This TV does it, the Steam Deck's done it, and it almost always looks terrible (the nano-etching on recent MBPs is _fine_; but still makes the text noticeably fuzzier).

The market for top-end TVs is the people who _really_ care about image quality, why would you jeopardize that with a coating that makes it _worse_?

I'm fine with this being an option for those who want it, but associating "top-end" with matte is bizarre to me. I regularly regret buying the OG 512 GB Steam Deck because the matte coating on it is just so bad.

klausa··on I think Anthropic and OpenAI have found product-market fit
You can point both Codex and Claude Code at a local model and they'll work just fine. Codex even explicitly supports that as a feature! [1]

With a nice UI on top, for the desktop app too: [2]

[1]: https://developers.openai.com/codex/config-advanced#custom-m...

[2]: https://docs.ollama.com/integrations/codex-app

klausa··on I think Anthropic and OpenAI have found product-market fit
The purchasing department _really_ doesn't like it, but it's the corpsec department that will really want to murder you for using "personal" accounts to interact with corporate codebases.
klausa··on Toshifumi Suzuki, founder of Seven-Eleven Japan, has died
I think you could make a relatively uncontroversial argument that pizzerias in Germany or Poland on average are pretty bad, but if you made that claim about Berlin or Warsaw _specifically_, I wouldn't take you seriously.
klausa··on Toshifumi Suzuki, founder of Seven-Eleven Japan, has died
This hot take, along with the "oh you can get better chicken than FamiChiki almost anywhere in Japan", drives me _nuts_, even as I live in Tokyo.

Yes! There are better options available if I want to sit down for a meal, or even just wait for a couple of minutes for someone to fry me a piece of chicken to order.

That's _not the point_! Both the SEJ meals, and FamiChiki, are _fantastic_ for what they are — available in literally tens of thousands of locations across the country, and available _instantly_, 24/7.

They're both _not that special_ if you compare them to a "real" restaurant (though, and I will die on this hill, FamiChiki is hands-down better than a good ~80% of chicken I would get in a restaurant in my home country; but that's a somewhat different conversation).

But if you compare them with convenience store meals available elsewhere in the world (especially in the broadly understood West), they still _are_ pretty damn special.

(And don't get me started on the 7-11 frozen pizzas from Da Isa. Those, reheated in a Balmuda also clear a good 75% of "real" pizzerias back home, and not because pizzerias in Berlin or Warsaw are particularly bad!)

klausa··on Nobody cracks open a programming book anymore
Latin can't be used to express Polish; at least not without extra letters, at which point I think it ceases to be "just" Latin script?

But maybe that was also your point.

klausa··on Ferrari Luce
They basically stopped _making_ any cars; it's kinda hard to not have a drop in sales after that.
klausa··on Iliad fragment found in Roman-era mummy
Maybe we run in different circles, but I feel like “appreciates scientific accuracy” and “appreciates historical accuracy” are pretty disjoint sets?

Not entirely without overlap, but pretty distinct.

klausa··on Iliad fragment found in Roman-era mummy
>I think that Nolan sells himself (The online worship can hardly all be organic) as an authentic, technical director interested in accurate physical props.

I'm not a huge Nolan-discourse-insider, but that seems like a pretty bizarre reputation to have for someone who's famous for directing three Batman movies, Inception, Interstellar and Tenet?

Is this reputation just because of Oppenheimer?

I haven't seen Dunkirk (and I'm not a WWII buff so couldn't tell if they used right planes/boats/guns/uniforms/whatever even if I tried), but even a short blurb on Wikipedia talks about a "balance historical accuracy with aesthetics that would favour the film stock".

klausa··on Archaeologists find Egyptian mummy buried with the 'Iliad'
> "Oh its a fantasy film" its set in a historical time period, I wouldnt watch a WW2 Zombie movie if the nazi zombies were wearing viking armor driving an Abrams tank either.

WW2 Zombie movie with Nazis in Viking armors and random tanks sounds so much more _fun_ than a "historically acurate Nazi Zombie" movie!

klausa··on Not alive, but not dead: disembodied human brains used for drug testing
I would save my cat before a distant cousin I met like four times in my life, absolutely.
klausa··on Apple unveils new accessibility features
It's licensing and/or product decision, not union labor.

I guarantee you for most (if not literally all) titles you see, they actually have more subtitles, they just don't / won't show them to _you_.

Netflix has gotten pretty good at VPN detection, but you used to be able to see this _very_ clearly by VPNing to a different country in the EU, and the available audio and subtitle languages changed dramatically.

klausa··on Click (2016)
>Should you have to prove legal residency to park?

...what? How is residency tied to parking now?

klausa··on Bun's experimental Rust rewrite hits 99.8% test compatibility on Linux x64 glibc
> Also their feelings are not hurt by what you or I think, they don’t care.

I would have agreed with this like 15 years ago, but the very existence of Twitter (and the acquisition saga) proves this to not be true.

klausa··on EU Age Control: The trojan horse for digital IDs
If you want to argue for a theoretical system that is self-contained, only relies on the data that is present on either the physical (or the theoretical cryptographically signed digital) passport, you're free to do that.

But in the real world, the systems that deal with processing people's entries already cross-reference multiple other existing databases, require internet connectivity to do so, and I think you'll have hard time convincing anyone to stop doing that.

klausa··on EU Age Control: The trojan horse for digital IDs
To check against $your-local-law-enforcement-agency database, $your-local-immigration-agency for history of entry, etc.

The internet requirement is not there for the person presenting the document, it's for the person/system checking it.

klausa··on EU Age Control: The trojan horse for digital IDs
This is not how the world already works.

If CBP's systems go down, they will not process (foreign, they'll process US citizens still) arrivals [1], even with physical passports in front of them. I assume the EU ESS works the same.

"If the internet goes down, your border checkpoint is down" is not some terrifying future we need to protect against, it's the reality of the world as you live in right now.

[1]: I've had to wait for an hour, at SFO of all places, because of exactly that happening.

klausa··on Used La Marzocco machines are coveted by cafe owners and collectors
Gaggia.
klausa··on Our newsroom AI policy
I don't want to come off as an AI-maximalist or whatever, but, I mean, at some point, skill issue, right?

You can use Google to find you results reinforcing your belief that the earth is flat too; but we don't condemn Google as a helpful tool during research.

If you trust whatever the LLM spits out unconditionally, that's sorta on you. But they _can_ be helpful when treated as research assistants, not as oracles.

klausa··on Our newsroom AI policy
The LLM can find material that it would be hard or time-consuming for you to do.

You still need to verify it, but "find the right things to read in the first place" is often a time intensive process in itself.

(You might, at that point, argue that "what if LLM fails to find a key article/paper/whatever", which I think is both a reasonable worry, and an unreasonable standard to apply. "What if your google search doesn't return it" is an obvious counterpoint, and I don't think you can make a reasonable argument that you journalists should be forced to cross-compare SERPs from Google/Bing/DuckDuckGo/AltaVista or whatever.)

← PreviousPage 6 of 30Next →