bjord
bjord on Hacker News
didn't realize that, though I wouldn't be surprised then if it also changes on sony's end
from your perspective this feels like a positive, then, no? cheaper price per hour, no requirement that you use those hours all at once?
yep, actually in my area this will be cheaper than the gaming cafes people frequent—I think the lack of reasonably-local servers is the main thing that will block adoption.
I'm not so certain the average gamer is well-informed enough to realize what they're giving up, or that they feel strongly enough to take a stand for it if they do. I'm hopeful they will, we'll see.
frankly, if I'm playing a game I own on a remote server, I expect to pay something for that
and I don't actually think that's substantially different from other competitors, like geforce now
whether I will choose to play it on a different server in the first place is a different question (I won't)
what's stopping you from using the same extension (ublock origin lite) in safari?
am I missing something or is there an incredible amount of title editorialization here? on the page itself (and within the slug), the title is:
"Benchmarking AI decision models against traditional guardrails"
neither, actually, just that unattended "oneshot" sessions are incredibly token inefficient
if you talk about them (in which you lean on an LLM as a sort-of independent employee) and conservative, chunk-based usage (in which you use the LLM as more of an extension of yourself), you're comparing apples to oranges
a predefined spec obviously reduces that gap but how much is highly dependent on the level of detail
> unattended overnight claude opus session
yes, exactly
Honestly, I'm not clear where the line is as far as what is reasonable, and it seems to me that no one is really addressing that question. It's easy to wave your hands and say "vibecoding is bad", without really defining what vibecoding is.
I suppose it'll be a few years before we see the true volume of technical debt catch up with the worst offenders, but even then, without the original LLM conversations associated with it, it'll be difficult to assess that in a structured way. Mind you, that doesn't even address the ever-improving models.
I think the issue here is the nebulous definition of vibecoding.
You, as an experienced engineer, are doing a lot of hand holding and review of LLM-generated output, maybe even(?) using it as purely a check on your own work. There are others, though, that are essentially outsourcing the entire process to a basic, underspecified chat prompt.
absolutely! no shade towards contributors, my impression was that compiling for iOS is just a bit of a moving target. everyone was willing and ready to help where they could, I just had a lot of trouble.
kivy, which does the same thing, has been around for over 10 years. I built my thesis in it a few years ago. not recommended—too many limitations and build issues in practice. given that this one is less mature...I'm not optimistic.
sorry, what? servo existing as a mozilla side project did not give it broad adoption
waterfox is downstream of firefox and can't meaningfully match mozilla's maintenance of the core browser
librewolf devs are just more honest about it
I do believe it makes sense in this particular case. xPaw made it clear in his other half of the announcement that he's been burnt out for years, and a big portion of burnout is being asked to do things, whether you love them or not, for free.
https://steamdb.info/blog/steamdb-nexus-mods/
I think that's perfectly reasonable if it means a more sustainable future
need this for android. on average, it takes genuinely about five minutes to open—I think it reindexes my entire micro sd card every time.
who's ready for gpu piracy?
enrico, at this point you're knowingly spreading misinformation. I corrected you in my last comment: the yandex search index is owned by the russian entity. there is no non-russian "yandex" anymore, period.
> After a short transition period following completion, YNV will cease to use the Yandex brand, which will be retained by the Target group. We expect that our International Businesses will develop their own branding going forward. We intend to seek shareholder approval to change the legal name of YNV in due course.
https://www.sec.gov/Archives/edgar/data/1513845/000151384524...
this is absolutely not true. yandex IS the russia-related business that got sold in 2024, the dutch entity (now known as "nebius group") is basically just data centers now.
I'm not saying I agree with it, but I think this comment
> Yandex does not add any special feature or functionality that is so great they should make it obligatory to include as a default option.
kind of captures kagi's argument: that its removal would degrade search results in a not-immediately-noticeable way, particularly for certain areas of the world, which might actually be worse than a loss of a distinct add-on feature or something.
...but yeah, from my perspective, it'd be great if you could turn it off.
they're conflating yandex with the international entity that sold it off to russian investors (now known as "nebius group"). the entire basis of the argument is wrong.
no it didn't. the international entity sold off the vast majority of yandex businesses and assets to russian investors, then rebranded to "nebius group". that is to say that the yandex index/api used by kagi is wholly russian-owned and has essentially doubled down on russia/its allies.
if you're curious, the dominant business retained by the international entity was their data center stuff (though yandex still operates its own cloud business in russia/elsewhere).
I'm aware. regardless, when it comes to models, the advice to "not use Google in Production" falls short.
there was a community of people who used (famously sycophantic) gpt 4o as—for lack of a better word—a friend, who were devastated when it was shut down
I suppose at least in this case the loss is not an emotional one?
have you experimented at all with the deepseek flash models?