Doesn't help how the us gov is saying Kimi 3 stole IP of Fable even though it's not possible to distill Fable in 1 month.

▲

Someone1234 13 hours ago | parent | next [-]

And they've provided no technical explanation of how...

This may be an odd reference, but I remember this song and dance with Roomba Vs. Roborock. "Roborock stole our IP!!!" but the problem was that song was slightly believable in the very early days, then Roborock started releasing products far better than anything Roomba had, but the tune in the US never changed. Literally for years after Roborock's products were two generations better than Roomba, but somehow they "stole" that?

China has money, talent, and cheap electricity/superior infrastructure. The US's big AI built their products on stolen copyright material and social media scraping. China likely got a catchup boost via distilling, but recently a lot of their AI ventures have come fully online, and they've just throwing tons of compute/power at the problem, and we're starting to see legitimately competitive products without shortcuts/cheating.

But the US won't change its tune; and that's convenient because they don't really want to compete, but rather roll out protectionism and fear per normal.

▲

chlee 13 hours ago | parent | prev | next [-]

the rumor in chinese web is the following.

Z.ai found a way to crack/overcome the cryptography that safeguarded Claude/Fable's thinking tokens in Claude Code, so Z.ai had the ability to peak into decrypted thinking tokens that Fable produced in Claude Code. Z.ai also shared their findings with other big LLM companies like Moonshot AI, which had 20,000+ accounts speed running through Fable to get their hands on as many high quality thinking tokens as possible for distillation.

https://github.com/rgreenblatt/which_claude_is_k3/blob/main/...

https://x.com/alex_prompter/status/2079625670918046048

Like all rumors, who knows what's true or what's not.

	▲	maxloh 13 hours ago \| parent [-]
		[dead]

▲

echelon 13 hours ago | parent | prev [-]

That's wild. It was only 15 days. It would have been impossible to gather enough data AND clean it up AND train AND do post training in just 15 days.

edit: the Anthropic employee downvotes are coming in now. This was +8, now it's +3.

	▲	benjiro29 13 hours ago \| parent \| next [-]
		> It was only 15 days. That, and usage was also half. Given the fact that we seen not any evidence beyond people "see, it one shot X game looks similar to Fable" (a lot of one shots look similar to other models). You expect to see something from Fable its restrictions/fallback behavior popping up in K3. You do not filter out all that behavior. Especially on such a short notice. For me the milk example, where somebody asked if some EU cheese's had type A/B milk type (they are allergic), triggering bio restriction prompts. That just shows you how hard it is to find all those restrictions because its not in obvious spots where Fable goes off. Yet, not a single person has shown this behavior in K3... The funny thing is, other Chinese companies can train on K3 when the open weight are released next week. So even if K3 was or was not trained on Fable, they already have a alternative to get their models to the same level.
	▲	luckystarr 13 hours ago \| parent \| prev [-]
		Then it's a case of trying to uphold the narrative?