The headline from 7 years ago has not borne out; Africa is a minor supplier.
The point is quite different: that everyone is harmed by advanced open weight agentic models even as we are benefited by them. This holds of African states as of any other. Many of them, by the way, are under attack from well-funded moderately insane groups like JNIM. Also by the way, groups like this are famously really good at recruiting the specific psychical type of the engineer.
It is not possible that releasing open weights will continue forever, or probably even to the end of the year.
China has no end of money to support these companies. The reason this equilibrium is unstable is that the autonomous agentic coding aspect of the models has been so successfully improved that it will soon be a threat to China state security.
Everyone just keeps assuming, as if it were the law of gravity, that China will continue in perpetuity to deliver the weights of its 'frontier' models to Hugging Face. Its Mythos moment is a few months away and there is plenty of reporting suggesting their response will be similar, which is anyway obvious.
It baffles me that anyone can seriously believe that China is going to put its Mythos successor on Hugging Face and let us all strip the guardrails etc.
Things just didn't turn out the way OpenAI and Anthropic thought; nor did they turn out the way China thought.
Everyone keeps assuming that China will forever be publishing these models, but it is clear as a bell that that day is over. In the end every model will be buried under state control for the simple reason that they will soon be really dangerous, as China understands better than anyone. They will give Kimi K3 to the Sinaloa cartel and the US military, but that's about the end of it.
Perfecting these autonomous agential coding AIs just moved way way faster than anyone expected. People are desperately trying to save their principles from six months ago but it's a different world. They can work out their issues with Trump, Dario and Sam, but it is Xi who will shut all this down.
China will be dropping the open models blather any day now. The autonomous agential coding Ai just turned out to be way too easy to perfect. They will give Kimi K3 to the Sinaloa cartel on Monday, but that's going to be the end of it.
Gatekeeping in the name of security is coming to the Chinese models too. Kimi k3 will be the last major step to go wild. It is clear that the state will quietly drop the soft huggable open models blather as the models get stronger. This is why Trump admin whining about oMg cHiNeSe OpEn ModEls etc is irrelevant; he should just leave it to China to draw the conclusion it in fact is drawing. Even six months ago things looked completely different.
I hadn't thought of it as named that way, but super long dash is a reality of ye olde typesetting, e.g. bibliographies with one author appearing many times, but named only with the first work, or in indexes where there are subordinate concepts etc.
Here's a googelensed bit of the index of a philosophical work from the thirties. The entries after 'Aristotle' are all megadashes, but google lens reads them as hyphens. The point is you are to retain 'Aristotle' in your head as you read this part of the index — the same words often have their own entry. Many of these subheadings appear in the same index under the 'Kant' entry — in his case, of course, the subheadings have subheadings so the typesetter doubles the triple-dashes up without however employing a six-em-dash. Current taste I think would just use indentation for all the purposes mentioned they are all line-beginnings.
You are the one hypocritically crying foul about Anthropic digesting the internet, while pretending Moonshot didn't do the same --- when of course it did, it's an LLM with inter alia perfect English.
Anthropic has made no statement on K3 I think. But if it does it will be to claim to be the real source of K3's merits.
Anthropic was claiming they were being drained for distillation specifically by Moonshot in February. I don't credit this guys charts but the supposed paradox is a mere repeated word-meme pushed across social media by the Chinese state.
The actual ruling had pretty much the opposite significance of what you suppose.
Its result is that now the LLMs chew through used, often rare, books pulping them immediately - which of course people also complain about in the same shrill tones.
Moonshot has 'pirated' just as much. Its what an LLM is: a training on all human composition. How can Deepseek or Claude give me very accurate answers about e.g. 19th c german idiom? But they do.
The point is quite different: that everyone is harmed by advanced open weight agentic models even as we are benefited by them. This holds of African states as of any other. Many of them, by the way, are under attack from well-funded moderately insane groups like JNIM. Also by the way, groups like this are famously really good at recruiting the specific psychical type of the engineer.
It is not possible that releasing open weights will continue forever, or probably even to the end of the year.