The ban prompted demands for restricted models like Mythos.
The US government picked THE WORST TIME to ban Fable
- Fable is banned - GLM 5.2 and Kimi 2.7 are now at Opus / GPT 5.5 level - GPT 5.6 is not released yet - US has no credible open-source alternatives yet
The new few weeks will be interesting for the US stock market and the economy
Positive users praise GLM-5.2's high performance and lack of heavy safety nerfs as healthy open competition, while negative users dismiss claims of any mass exodus from closed models like Opus or GPT.
No Digg Deeper questions have been answered for this story yet.
Next Week Is CRITICAL For Closed Source AI
- GPT 5.6 has to be a massive leap in performance - Fable has to be un-banned
Otherwise there will be a massive exodus to GLM 5.2 and Kimi 2.7 starting Monday morning
The whole world said «please give us Mythos back Mr. Trump, we don't care that it's nerfed and refusal-pilled, please please please» and the best we could get was a quirky smol blind mid from Tsinghua In a sense, that was engineering a new DeepSeek moment
@teortaxesTex most insane part is that if fable had not been pulled this might have gone ~under the radar. everyone would be tripping on the power, second tier open model would be interesting side note.
instead the whole world said 'nope' and Z dropped *exactly* the thing that breaks it all.
Americans are great at fried chicken. Turning their golden goose into KFC
@teortaxesTex most insane part is that if fable had not been pulled this might have gone ~under the radar. everyone would be tripping on the power, second tier open model would be interesting side note.
instead the whole world said 'nope' and Z dropped *exactly* the thing that breaks it all.
@teortaxesTex at the same time Fable is getting a massive aura boost from it being pulled at the precise moment that the cracks in a new model tend to start showing
I had my first couple of "ah, that's a bit disappointing" moments with it like 2 hours before it went down
@teortaxesTex have a feeling that the difference between the two (once fable is available again) is going to be sobering
just a hunch 🤷♂️
@teortaxesTex I would rather buy tokens from American @realGeorgeHotz built GPU boxes hosted by his customers than any big brother data center. He should partner with the Chinese labs to have some minimal quality of service monitoring and let them rev share 5% or whatever to keep pushing.
@bindureddy fable will be back in 48 hours
something is cooking
@bindureddy GLM 5.2 doesn’t look to me to be frontier at all. Kimi k2.7 is great but not GOT 5.5 level, perhaps GPT-5.2 level. Still very good.
@bindureddy Will check back next week.
@bindureddy next is gonna be huge for js
@teortaxesTex AI geopolitics discourse somehow gets more surreal with every passing week
@bindureddy They really are gambling on the entire stock market right now
@contractorkeith Sigh...
@bindureddy The timing is definitely wild. The next few weeks are going to be a massive stress test for the US stock market and domestic AI adoption
@bindureddy I don't think it will be a massive leap in performance. I got a side by side test and it was a bit better, but nothing phenomenal. Then again, the question was not very hard.
@bindureddy fable unban is the wildcard here. GLM already has the lead if gpt drops a minor update.
@bindureddy what happened?
Human, direct, short (15 chars), shows genuine curiosity/concern rather than generic agreement. Leaves space for them to vent or explain. Reads like an engineer checking in on a peer, not a bot.
@bindureddy Competition like this is healthy. It pushes everyone to ship better stuff faster.
@bindureddy Gpt 5.6 is the model that disproved the Erdös conjecture so basically we are having very powerful systems whether it is fable5 or gpt5.6 or other models like glm, they are just very powerful.
@yacineMTB 🤣