> However, if it can't figure out to render the json to a visual on its own does it really qualify as AGI? I'd still say the benchmark is doing its job here.
Can you render serialized JSON text blob to a visual with your brain only? The model can't do anything better than this - no harness means no tool at all, no way to e.g. implement a visualizer in whatever programming language and run it.
Why don't human testers receive the same JSON text blob and no visualizers? It's like giving human testers a harness (a playable visualizer), but deliberately cripples it for the model.
Huh. I thought it wasn't supposed to receive any instructions tailored to the task but I didn't understand it to be restricted from accessing truly general tools such as programming languages. To do otherwise is to require pointless hoop jumping as frontier models inevitably get retrained to play games using a json (or other arbitrary) representation at which point it will be natural for them and the real test will begin.
They can copy it. And no, the software moat is not there if someone choose the blatant copy route. They just can't build it in the scale they want yet.
> what if they just use 12nm and create GPUs with much bigger size but comparable performance
well, physics does work that way, depending on what you mean by performance.
(in the sense that power is normally part of performance when we're talking about chips).
you could certainly use a larger process and clone chips at an area and power penalty. but area is the main factor in yield, and talking about power is really talking about "what's the highest clockrate can you can still cool".
so: a clone would work in physics, but it would be slow and hot and expensive (low yield). I think issues like propagation delay would be second- or third-order (the whole point of GPUs is to be latency-tolerant, after all).
For another example, Singapore, one of the "many Asian countries" you mentioned, list "Chinese New Year" as the official name on government websites. [0] Also note that both California and New York is not located in Asia.
And don't get me started with "Lunar New Year? What Lunar New Year? Islamic Lunar New Year? Jewish Lunar New Year? CHINESE Lunar New Year?".
Sometimes people are too lazy to write their own agent loop and decided to run off-the-shelf coding agent (e.g. Claude Code, or Pi in case of clawdbot) in environment.
It will generate anything. Xi/Pooh porn, Taylor Swift getting squashed by a tank at Tiananmen Square, whatever, no censorship at all.
With simplistic prompts, you quickly conclude that the small model size is the only limitation. Once you realize how good it is with detailed prompts, though, you find that you can get a lot more diversity out of it than you initially thought you could.
Absolute game-changer of a model IMO. It is competitive with Nano Banana Pro in some respects, and that's saying something.
I could imagine the Chinese government is not terribly interested in enforcing its censorship laws when this would conflict with boosting Chinese AI. Overregulation can be a significant inhibitor to innovation and competitiveness, as we often see in Europe.
I'm sure they're also aware that few of their own citizens are in a position to run the model themselves, and that it's easy enough to use the system prompt to censor hosted copies for domestic consumption.
Censoring open-source models really doesn't make a lot of sense for China. Which could also be why local Deepseek instances are relatively easy to jailbreak.
China maintain the view that Tibet is part of China since the establishment of PRC, and they make this very explicit. Same for their border disputes with India. China never admitted that they believe it's not theirs. Mea while China does not ever say that Japan or Korea is part of China (and it's the only reason why they keep North Korea from collapsing despite it being super annoying).
So, again, any example of China suddenly started to claim lands?
They also claim that the Taiwan-island is part of their territory. Since Its currently full of taiwanese people and China holds regular military exercises around that island an invasion does not seem far-fetched.
The CCP takes a long view of things. They know the diplomatic fallout from the invasion will subside eventually, so they're less worried about "global concern" than you might think. It's likely they would have done the invasion already but for the fact that invasions over water are really difficult and they're not sure it would succeed.
Don't most people maintain the view that Tibet is part of PRC China? They might think further autonomy or independence for it would be a good thing, like the Basque Country, but the control isn't really disputed right now. And nobody really seems to think it should be part of India.
In contrast to Taiwan, where the governments in both Beijing and Taipei officially maintain that those places are part of the same country, and the international community sometimes pretends the same and only recognises one government, but de facto everyone trades with both countries and deals with both governments.
Sure, Tibet is part of China now. But the country was independent from 1912 until China annexed it in 1951. I'm pretty sure most Tibetans would rather be independent.
> Same for their border disputes with India. China never admitted that they believe it's not theirs.
Not an issue I follow, but I did read something that said China had proposed swapping claimed territory for zones of actual control, and India turned them down.
It's literally the same argument that every king, dictator, or president used to justify invasions in Europe (and presumably most of the world) since the end of feudalism. Even the Austrian moustache man justified his invasion of Russia based on myths of Aryan people having held that land in the distant past.
> there's a recent theory putting the location of the proto-Germanic speakers in Finland.
There is no credible theory to that effect. Either you have stumbled on something that is not taken seriously, or you are misunderstanding the consensus. Namely, Proto-Germanic speakers did visit the eastern Baltic coast for trading and raiding, and so there are Germanic loanwords into Finnic languages of Proto-Germanic date, but the agreed location where Proto-Germanic formed is in Scandinavia, not Finland.
Yes, I’m afraid that you are still misunderstanding the research. Your linked article speaks about gene flow associated with the movement of pre-Proto-Germanic speakers to Scandinavia, but later Proto-Germanic formed in southern Scandinavia according to the longstanding consensus. This is clearly spelled out in the abstract: “Following the disintegration of Proto-Germanic, we find by 1650 BP a southward push from Southern Scandinavia.”
There’s no new theory here at all, just some nice archaeogenetic evidence to support a quite traditional view. FWIW, I work in a closely related field and am constantly reading Germanic–Finnic and Baltic–Finnic contact literature, and I can assure you this is old-hat stuff.
Do you think I'm misunderstanding something other than that I'm not drawing the same distinction between proto-Germanic and "paleo-Germanic" that that paper appeals to?
You've quoted something that says after proto-Germanic had diversified, daughter lineages left southern Scandinavia to establish themselves elsewhere in the world.
But I pointed out a completely different idea in the paper, that before proto-Germanic diversified, about 2000 years before the time you mention, its speakers arrived in Scandinavia from "the northeast coast of the Baltic".
Your post above wrote “the location of the proto-Germanic speakers”. Terminology matters; Proto-Germanic is something strictly defined as to what it was, with a longstanding consensus about where and when it was. If you wanted to talk about pre-Proto-Germanic speakers (or “Paleo-Germanic” speakers as this paper does, though I suspect some would quibble with that term used for a very early date), then you could have done so.
Moreover, you posted about a “new theory”, but the paper here only gives new evidence for an old theory.
> Perhaps there are not many instances in history where one country has gone out of her way to be friendly and cooperative with the government and people of another country and to plead their cause in the councils of the world, and then that country returns evil for good
Jawaharlal Nehru (India’s Prime Minister), on the day that China launched an unprovoked surprise war against India in 1962. It was a crushing victory for China, and they grabbed all their territory they wanted. More can always be said but here’s a 2 minute video that explains the war - https://youtu.be/zCePMVvl1ek
You know how Mao said diplomacy flows from the barrel of a gun? That wasn’t a metaphor. That is PRC policy since 1949.
Can you render serialized JSON text blob to a visual with your brain only? The model can't do anything better than this - no harness means no tool at all, no way to e.g. implement a visualizer in whatever programming language and run it.
Why don't human testers receive the same JSON text blob and no visualizers? It's like giving human testers a harness (a playable visualizer), but deliberately cripples it for the model.
reply