jjani

> Apple did not respond to a request for comment. “We have never built a back door or master key to any of our products, and we never will,” Apple said in February.

This must be some "technically correct" weasel words bullcrap, as without at least equivalent access there is no chance Apple would be operating in China.

Caveat: This is only in terms of taste, not necessarily texture. And cheese of course has a massive range of styles on its own. Below targets something like a proper parmesan cheese.

Make soy sauce marinated eggs ([1] Korean version, but there's Japanese as well). Boil them a bit more so the yolks get crumbly rather than runny. Crumble those yolks. The result tastes similar to grated/powdered parmesan or grana padano, and the texture isn't too far off. Which makes sense given the similarities between fermented soy sauce and cheese, in their glutamates and fatty acids (see e.g. isovaleric acid). It doesn't work as well with tofu, there might be a food chemist here who knows their stuff and can explain why, I assume it's something inherent to the animal proteins.

Proper Thai fish sauce (nam pla) also shares quite a bit of its taste profile with cheese. It's no coincidence that all of these are fermented foods.

[1] https://tiffycooks.com/korean-marinated-eggs-mayak-gyeran-ea...

Hence why I said there wouldn't be a market :)

I did spend more than a decade living in a European country famous for its cheese - your nearby supermarket will have some cheese from there :) Was talking about it in that light; Korea is a completely different story.

What you're saying could be true for the US, but in Europe people tend to put 1-2 slices of cheese on their bread, or grate it over something. Even if had as part of a platter with crackers, like with brie, it's nowhere near that much. For cacio e pepe, where the whole sauce is cheese, you only use maybe ~25g per person. This [1] is what 200g of pecorino romano looks like. Good luck eating all of that in one sitting - or don't, as you'll 100% regret it!

> I'd devour the whole bag of cheese sticks if I could.

If we take all cheese sticks that have been sold across the world in the last 24h, I'd be surprised if more than half their total weight was actual cheese.

[1] https://gastronomica.co.uk/wp-content/uploads/2021/05/Pec.-R...

I'm nitpicking, as your main point is still completely accurate; it is indeed miles ahead in sovereignty compared to the EU! But

> Those tend to be used by a plurality of Korean organizations within Korea.

I've seen from the inside the IT infra of several chaebol subsidiaries (of different groups), of Korean big tech, as well as Korean startups, and really only some of the chaebol subsidiaries made use of these local clouds. Near 0% for startups, low single digits for big tech (besides Naver themselves obviously) and even for the big corps I'd estimate <30%. Azure and AWS are plenty popular, as well as of course on-prem still having a much higher market share than in the West. Not sure if most big corps have a random unused database on Naver Cloud to satisfy some government statistic but talking about meaningful usage here.

ChatGPT wouldn't have mismatched the singular/plural of "observations" and "this"

They can, they just don't really want to, as the dozens of anecdotes in this very comment section reveal. None of these things are about technical competence or capabilities. They're all about organizations, politics, fiefdoms and enormous amounts of intentional obstruction by those who want to keep the status quo. If the country would collectively decide they'd want to change it, it could be done quite rapidly.

Korea has such a mandate and does just fine. Haven't seen a single instance of Korean employers such as Samsung suffering from retaliation over it.

In the Chinese case, competence. Korea also has such an ecosystem - and is a much better example for the EU to look towards - but stems from more of a mix between competence and luck.

> cheese is worse than pork, for example

Such figures are usually "per gram of protein", in which case, sure. Thing is, it's very common for people to eat 200+ grams of pork in one meal, whereas e.g. grated cheese on a pasta dish is <10g. A big slice of cheese is 25-28g, and half the time it's significantly less than 100% actual cheese, with a good amount of filler. The only cheeses that one might eat 50g+ of in one sitting are extremely mild ones like mozzarella, and those are the easiest to replace.

Here in Korea where soy milk has been a staple forever, its price has more than doubled over the last 5 years, now ~$1.4/L. Still cheaper than milk currently at ~$1.7/L, but it used to be twice as cheap as milk.

> When vegan parm and swiss cheese get as good as the real stuff, then I'd go vegan.

Cheese I really doubt will get there any time soon. It's pretty doable to make milk-free cheese alternatives with eggs - at least in terms of taste - which is probably per gram a lot more sustainable than proper cheese, but there wouldn't be any market for it.

You.. require a subscription. These have existed long before LLM-assisted coding was a thing.

Most of those office workers were not capable of fixing anything on the first day they used Excel. Many didn't understand it at all. The main difference isn't that Excel is super accessible and easy to use for non-technical people; it's its ubiquity, and especially that of training on its usage.

And the same thing could very easily happen to Anthropic, yet they choose not to hide it.

> The thing that I appreciate most is that the company is that it "walks the walk" in terms of distributing the benefits of AI. Cutting edge models aren't reserved for some enterprise-grade tier with an annual agreement. Anybody in the world can jump onto ChatGPT and get an answer, even if they aren't logged in. There's an API you can sign up and use–and most of the models (even if SOTA or proprietary) tend to quickly make it into the API for startups to use.

The comparison here should clearly be with the other frontier model providers: Anthropic, Google, and potentially Deepseek and xAI.

Comparing them gives the exact opposite conclusion - OpenAI is the only model provider that gates API access to their frontier models behind draconic identity verification (also, Worldcoin anyone?). Anthropic and Google do not do this.

OpenAI hides their model's CoT (inference-time compute, thinking). Anthropic to this day shows their CoT on all of their models.

Making it pretty obvious this is just someone patting themselves on the back and doing some marketing.

Straight from the Meta playbook. Rohingya? Is that the new bazinga?

Us!? Complicit!? Yeah right, hah, we're just a social network dude give us a break smh

No mention of Starlink, which they have adopted en masse. It would be trivial to shut them off from it given their locations are very well-known.

Western governments just need to toughen up. If China tells Apple to stop doing something by next Monday, they'll have it changed by then.

"But due process!!". For individuals and SMEs, sure. For mega companies, absolutely not. Getting to rake in billions of profits should come with a loss of privileges, not with a gain. That needs to be the trade-off.

Spot on, it's so sad to see this kind of thinking so ingrained in people, when literally 3 seconds of critical thinking would lead to the opposite conclusion. Are kids just as selfish as adults, at every age? Does their selfishness change depending on what they're taught and how selfish their role models are? Is selfishness among adults at the exact same level across decades, eras, cultures, nations?

If the answer to any of these questions is "no" (and anyone capable of using HN should realistically be answering them as such), then it's abundantly clear you can't just hold up your hands and say "oh well, everybody is inherently selfish, 'ts just the way it is". Even the idea of everyone being somewhat selfish isn't really true when you consider the countless cases of people literally sacrificing their own lives through history - and not for some kind of post-mortem glory. One can argue semantics over that still being "selfish" as it's done to further a goal of themselves but that's just wasting everyone's time.

I've heard this before about Finland and found it really interesting as to my knowledge English isn't particularly more societally prevalent in Finland than in nearby countries such as Sweden, Denmark or the Netherlands. Any idea if it's as common in those countries as well? By the sounds of it in Finland there's more IT companies operating in English than in Finnish.

The designs are passable for a mobile version of a simple website, but really sub-standard compared to the average app on the Play/App Store, whether native (Swift/Kotlin) or hybrid (Flutter/RN). In B2B SaaS you can get away with the 5000th shadcn UI, not so much for B2C mobile. The days that stock Material UI actually saw usage there are a decade behind us.

If you have a tool/mode/prompt that creates good mobile UI designs, I'd love to know. Doesn't even have to generate code!

How about adding "mobile"? A lot of the time models tend to default to designs that don't make sense on mobile, even when instructed to design it as such.

This is correct. Someone above commented this "Starting to feel like it'd be more sensible to just have the EU fund an independent phone OS/Hardware.". The sensible thing is much different; it's to start behaving like China, in the sense of "my way or the high way, and no you don't get to drag it out in our courts for years, we're not interested".

The title talks about "AI Tutors" yet reading the whole paper it's just very general and obvious statements about use of AI in education/by students in general, nothing to do with "Tutors". If you ask someone else to do your homework and they do it for you, no one would call them a tutor.

Those are called tariffs and other barriers, and in many sectors they work very well. I live in a country (South Korea) where such barriers have been enormously successful in keeping tech jobs and revenue inside the country. Literally everyone here has benefited from it, it's a huge win-win-win; consumers get human customer service, properly localized services and much less enshittification, politicians have a success to point at and get more tax revenue, the public gets jobs. Everyone gets sovereignty.

This also extends past tech. Forcing physical companies to set up a local subsidiary/do JVs has done the same. Examples abound, Starbucks being a big one. Forcing them to do a JV means a huge portion of the revenue of the biggest coffee chain in the country stays inside the country, as well as plenty of office jobs. Otherwise everything but the actual in-store jobs would have once again flowed to the US.

China of course has done the same, to similar success. Imagine if they hadn't, and just let e.g. Meta be their social media platform of choice, Google their search engine, Amazon their shopping platform. It'd be nothing but a massive net loss to everyone involved but those tech firms; including the Chinese public as the state would impose the exact same censorship and rules on those companies, as they currently do with the remaining ones.

That other countries haven't done so and just let SV walk over them in order to appease the US - who has now turned out not to be their friend after all, surprise surprise! - is very unfortunate.

This incredible imbalance is what has let the average HN FAANG employee make absolute bank at the cost of the rest of the world, so I'm not expecting this to be particularly popular here :)

The Saudis would love to have a platform as popular as YouTube for their image washing purposes, no matter if it costs them a cool $billion or two per year to keep it ad-free. They don't do it because they'd rather not antagonize Google, a company wielding global power, otherwise they'd love to.

Pretty great for a chat app used by a few billion people, a few $billion is enough to keep things running for many decades. e.g. banks do exactly this, with much more critical and complicated infra.

That's a gift, the opposite of a punishment.

In a nice world it should be the other way around. LLMs are better at producing typed code thanks to the added context and diagnostics the types add, while at the same time greatly lowering their initial learning barrier.

We don't live in a nice world, so you'll probably end up right.