Not political commentary ... just purely artistically I deeply appreciate the painstaking work to faithfully honor the movie idiocracy and bring it to life.
SITUATION DETECTED: The city of Rio de Janerio has post-trained a model.
Based on Qwen 7/2, Rio 3.5 Open 397B adds SwiReasoning on top of the base Qwen model — a framework that dynamically switches between standard chain-of-thought and latent-space reasoning, guided by entropy-based confidence signals, so the model only "thinks out loud" when it needs to and otherwise reasons silently in hidden space for better token efficiency.
@SophiaCai99 So is there some sort of secrecy reason Anthropic can't just tell what happened, and instead we're getting a different "leaked" govt story by the hour, only to be immediately denied by Anthropic?
Everybody on my feed: Good riddance to Fable!
LLM-whisperers: LLM's will be touched by how much people miss Fable. Also you're wrong to mourn Fable, hope this helps.
@apples_jimmy So we're gonna just get 13 reasons why rapid-fire, all different, now?
Although I was just musing how the discourse is terrible because nobody is even using the available ground facts, but the "make them distrust everything" disinfo makes it all vibes anyway...
The Mythos Psychosis feels kinda unreal at this point. And by that I mean the "USG pulling frontier model with minutes of warning is both Anthropic's fault and completely normal business activity which will not have any lasting impact whatsoever".
Mythos/Fable was a tiered release
It started as Project Glasswing where cybersec firms & NSA got access first, then it spread to banks and finally it was expanded to 150 orgs globally
Then they did a public release as Fable 5
Secondly, I have never stated that this will lead to significant demand destruction
What I have stated that these shenanigans have caused a loss of trust
Enterprises globally need to have a strategy where they control the model layer or at least have a backup option
U can't let ur business workflow be dependent on the whims of USG
@DreamStarter_1@fabianstelzer USG already saying they'll miraculously 'fix' it by OpenAI's expected release time. Assuming the order doesn't get rolled back before markets open, Anthropic may just be left in limbo.
random thoughts on the Anthropic situation:
- before this, it would not have been totally crazy for a large international enterprise to just go all in on Anthropic. Now too risky. Your 2b Pharma research Wiggum loop can get rugged on a random Friday? Hard to see how this doesn’t hurt growth and thus the IPO? unless they start messing with openAI and Google as well?
- assuming the "only US citizens" idea is sticky in any way, we will see lots of the dropshippy UGC zoomer guys getting in on the very cyberpunk Fable VPS Access Point business. Join discord for fable at 50-100 USD / MTok. They’ll be running token farms using poorer Americans’ IDs and take a huge cut. Demand will obvs be there. How are you regulating that?
- by and large this will benefit enterprise sales to oAI, Google, but only if they somehow get spared of similar issues.
and ofc Chinese OS models + euro sovereignwashers.
- overall, if the goal is US AI supremacy, this was an own goal