That makes it sound like there’s something actually new here, but it seems like it’s just a Chinese company getting closer to what "Open"AI / Anthropic already have.
It’s nice that these things are open source, but from a technological view this isn’t a breakthrough.
It feels like nitpicking, but there’s a relevant difference here. If you can only access the weights (and not the training data and the learning algorithms used), it’s hard to get (say) a Chinese model to freely talk about Tiananmen Square Massacre. (Just tried it locally on Deepseek R1; it’s doable, but takes some prompt hacking.) The models may be freely available, but there’s bias and censorship baked into them.
That makes it sound like there’s something actually new here, but it seems like it’s just a Chinese company getting closer to what "Open"AI / Anthropic already have.
It’s nice that these things are open source, but from a technological view this isn’t a breakthrough.
Dunno about getting ‘closer’ but glm-5.2 is indistiguishable from what I got before from claude and openai while being cheaper.
Open weights != open source
Thanks, didn’t know that!
It feels like nitpicking, but there’s a relevant difference here. If you can only access the weights (and not the training data and the learning algorithms used), it’s hard to get (say) a Chinese model to freely talk about Tiananmen Square Massacre. (Just tried it locally on Deepseek R1; it’s doable, but takes some prompt hacking.) The models may be freely available, but there’s bias and censorship baked into them.