The US revealing its true face (the fact it was no allies, only vassals) and trying to bully Taiwan into giving up its most precious economic could be something that help China in the long run. I mean, given two bullies, why try to appease the distant and foreign one, instead of the one with cultural and linguistic ties? Seems an incredibly short-sighted move from Taiwan, but it's good that more people see its true colors. Taiwan should try to gain more protection from Japan.
It's true tho that Japanese universities have a problem with labs communication. In the former university I was a MEXT student in, there was at least three labs that could have collaborated on topics that my lab was focused one. Granted, one professor welcomed me in his lab to do some stuffs with his students. But otherwise it was like little fiefdoms taking pride in doing things alone. This is contrasting with what's going on (at least in intention) in French research.
I think it's a tech advancement issue, just like touch screens wasn't a thing for a very long time while it they were resistive. Then the iPhone came with capacitive touch screen and now they're everywhere. When (if ever) XR won't suck, which could take a decade or two, it will be adopted massively.
The fundamental difference is in the public: extreme-left for bluesky, trumpist for truth social and twitter for whoever isn't making a platform choice part of its identity.
> Actually, the alphabet invention was precisely that the same alphabet could be used for different languages.
No. Multiple languages were written with the same alphabet because of the borrowing of that technology by speakers of other languages, not because it was designed as such.
Actually the process of borrowing is fondamental to the emergence of alphabets: the three biggest (by corpus size) logographic scripts all gave birth to more phonetic scripts across language boundaries. That happened with hieroglyphics (the case we are discussing), but also with Sumerian cuneiform > Akkadian writing and Chinese characters > Japanese kanas.
This. I'm making a game and I have an Android and iOS experimental branches. Even with a multiplatform framework (Monogame), compilation and basic run isn't a given. Then thing like small buttons needs biggee touch target, text input requires adaptation. All those 3 days + 1 day + a week + ... quickly adds up.
Those people are doing real work, it's normal they're paid too. If musicians want more for themselves they could cut middlemen and produce and commercialize themselves their music.
This was also a thing in Alsace (France), under the same name. However I only ever heard of it from my mother, as it is was a thing people did when she was young, but isn't a thing anymore since decades. Maybe in a few villages?
No, GPT-4 would have been classified as it is today: a (good) generator of natural language. While this is a hard classical NLP task, it's a far cry from intelligence.
That's not always a bad thing, especially when it's about things like "let's do this or read about that" instead of sleeping at a fixed hours. Might be a part of why night owls are more creative.
Enshitification is slower on Wikipedia than elsewhere, but sadly still happening. I'm still upset about the language button that cause me one additional click instead of the direct links on the left menu.
It was decided years ago, but some politicians and shareholders decided it was a good opportunity to exploit. Also the Russian economy was supposed to collapse few years ago because of UE sanctions. Still waiting for that to happen.
> I'd argue that if you accept that the results were "likely influenced by a likely state-actor campaign" then the means by which they achieved their objective are not above scrutiny.
No influence doesn't exist in any country. In France during the first elections Macron was elected, he was pushed heavily by the media. In 2016, almost every news outlet in the country were vocally against Trump. Those are example of "state-actor" level influence on elections, but somehow it's supposed to be fine because it was supporting the "good" candidate. Democracy in unironically in danger, but not for the reasons often voiced. So people and factions try more and more to impose what is the supposed correct outcome of elections or the definition of democracy.
We need to find a Plato cave analogy for people believing LLM output is anything more than syntactically correct and somewhat semantically correct text.
That's not a problem. OpenAI need to get some cash from its product because the competition is intense from free models. Moreover, since they supposedly used most of the web content and pirated whatever else they could, improvements in training will likely be only incremental.
All the while, after the wow effect passed, more people start to realize the flaw in generative AI. So current hype, like all hype, as a limited shelf life and companies need to cash out now because it could be never.
Research disagrees with you. Memorizing vocabulary list is one of the most efficient use of time for learning a language. Sure it's not sufficient alone, and not fun but it works. It's especially useful at the start when learning even a few hundred words makes a huge difference.
It is a good branding, like neural networks, and even artificial intelligence was. The good point is it makes really easy to detect who is a bullshiter and who understand at least very remotely what a LLM is supposed to produce.
> I have some strong FOMO about not being able to learn much technical stuff from my peers or seniors.
You don't need to FOMO that, really. It works like that in theory, in practice... In about 2 years of work in a midsize company, I learn about one or two little details from the tech leads in the company. Two third of them were spending more time asking why one would think there are worth asking them a question than replying the question itself. The level of rockstar insufferableness was off the charts.
A well working CPU is probably beside the point. What's important now is for researchers to publish papers using or speaking about AI. Then executives and managers to deploy AI in their companies. Then selling AI PC (somehow, we are already at this step). Whatever the results are. Customers issues will be solved by using more AI (think chatbots) until morale improves.
> the only avenues I see for curiosity-driven researchers are becoming independently wealthy, living like a monk
I came to the same conclusion. This is the path I'm following (trying to set up a company and lean FIRE). It's sad in a way because those efforts and years could have been directed to research but we have to adapt.
People who should use Sumerian characters don't even use them, sadly. First probably because of habit with their transcription, but also because missing variants of characters mean lot of text couldn't be accurately represented. Also I'm downvoting you for discriminating me.