Hacker News Reader

Comment by segmondy | Hacker News Reader

In the last few days, Chinese labs have given us MiniMaxM3, KimiK2.7 and now GLM5.2. Meanwhile US is censoring models. Reads like fiction.

View on HN

I don’t understand how I grew up thinking USA is the gold standard is good and China just make cheap copies and is bad.

But these news really changes my view on China and USA. I can’t believe it almost.

> I don’t understand how I grew up thinking USA is the gold standard is good and China just make cheap copies and is bad.

Hollywood.

> I don’t understand how I grew up thinking USA is the gold standard is good and China just make cheap copies and is bad

You did not grow up in the 80s ... Where it was the same about US vs Japan. Look how it turned out for several of the US industries. The US tends to sleep, look down on other countries, and then it loses key industries because of that attitude.

It’s not just the US. Honda recently announced they’re not able to make internationally competitive (=== high quality and affordable) cars anymore, and abandoned their entire EV line.

I guess they’ll just milk the ICE assembly lines until they are bailed our or go under, Detroit-style.

Well china still making cheap copies (distills)

Distillation helps, but is only a minor part of overall training effectiveness. It's not like everything suddenly collapses if distillation is made impossible.

propaganda

Any idea how kimi2.7 compares with GLM5.2?

It seems to be on par with Sonnet from my (short) experience. If GLM 5.2 matches Opus that is quite a step up for more complex and longer tasks.

The Chinese models are censored (too?).

> US is censoring models

For the current Anthropic issue, I’d say that’s more likely to just be generic corruption, revenge, shakdeown, and/or incompetence from the Trump admin. ‘Censoring’ might be technically correct, but I think one of the aforementioned verbs is a better fit.

> The Chinese models are censored (too?).

This is MUCH less of an issue if they're providing the weights though.

They can still be fine-tuned & ablated.

> corruption, revenge, shakdeown, and/or incompetence

Sadly, I think it's all four at once.

Also, IPO prep work. The whole industry is speed running enshitification.

It’s not just the models. Try copy pasting stuff out of the claude app, or sharing a conversation. It’s completely broken now.

It feels like the difference is really just the competence level of the corrupt government.

It’s not like the American regime is anti-censorship but pro-shakedown.

china is doing political censorship on specific well known topics - their current government, corruption and workers rights, taiwan, tiananmen square, human rights in xinjiang, old communist leaders and a couple others. not defending them just to be clear. the rules are the same for everyone, if your model cant give you anti ccp views with normal prompting you are safe. they never tried to restrict access to a model thats already public for vague security reasons.

Tbh if we had a Harris admin I expect we'd have some sort of locking down by now.

Probably. But it would be at least somewhat thought-out and apply to all the AI providers. Not just the one currently disfavored by Captain Dipshit and the Sycophants.

I really don't know why business cozies up to Trump so much, given how unbelievably unreliable and mercurial he is about...everything.

Because he’s running a racket. If you don’t cozy up to Trump you will suffer and it’s as simple as that. This administration is a mafia.

Nobody thinks the leopards will eat THEIR faces, just their neighbors'.

I didn’t follow the news continuously enough to know what 5:21 or your comment meant.

Background reading:

https://www.anthropic.com/news/fable-mythos-access

tl;dr: Anthropic supports government centralized government control over models, Amazon produced a probably bogus request to pull down Mythos and Fable, so Trump pulled it down.

It’s probably bogus because no evidence of effective jailbreaks were provided, and also Fable/Mythos isn’t any more capable than OpenAI’s pre-jailbroken 5.5 offering, making it a moot point.

Anthropic can put it back up once they institute citizenship checks for their customers and ban any foreign nationals they employ from using it.

(All of the above according to Anthropic)

I’ll editorialize and say that this is blatant illegal retaliation on the part of the admin, and also that anthropic brought it on themselves with their “this model will kill us all” Mythos marketing stunt.

I guess in this story, Amazon is the useful pawn/idiot. Maybe it’ll go Shakespearean, and we’ll get some lowbrow comic relief from Bezos.

Reply on news.ycombinator.com