UhuBuhu

[UhuBuhu @kecksbe]

@TheHumanoidHub we need to start building more stuff like this again just build more big things

[UhuBuhu @kecksbe]

@fzdart @hightbunker sry but isn't this just the old clone minecraft thing i remeber either it was back with 4o, o1 or o3 not sure on the exact model but one of those where a lot of people showcased controlling apps via hand signs build by one of these models.

[UhuBuhu @kecksbe]

@KennyDeMetter @Rainmaker1973 the real purpose is of course gathering trainings data though

[UhuBuhu @kecksbe]

@KennyDeMetter @Rainmaker1973 of

[UhuBuhu @kecksbe]

@MichaelWaitze @vibecademyai First of all IT makes more fun coding with Fable since IT gets you without you having to explain everything in detail. Second ITS faster than Opus cause IT makes less mistakes and writes better Code. And IT gives you way better insides into other solution/has more novel solutions

[UhuBuhu @kecksbe]

@emollick Not available in Germany sadly

[UhuBuhu @kecksbe]

@Ratonisko @stupidtechtakes If it's the one i think IT IS than this one isn't

[UhuBuhu @kecksbe]

@ImLunaHey Not hitting Limits often on the 5x pro plan of Codex i was also doing Well with 5x Claude but Limits we're rougher there. Maybe try the new claudex Meta where you use Fable AS Main Agent and let IT use your small Codex Models Like Luna AS subagents at least for exploring

[UhuBuhu @kecksbe]

@telep_io @shivamhwp @theo Made a Post about IT kinda tells you how to do IT @thsottiaux did one too If you are lazy Like me Just give Codex the links to the Posts and Tell IT to set IT Up for you and there you go

[UhuBuhu @kecksbe]

@kimmonismus Efficent

[UhuBuhu @kecksbe]

@kimmonismus I mean it can BE more token efficent but that also means it's faster meaning you qeue Up the next prompt faster which at the end generates the Same Tokens per hour. If the token Gen Speed also increased with the model IT would make Sense that usage drains faster while beeing more

[UhuBuhu @kecksbe]

@adonis_singh Now Imagine cerebras 5.6 generating Code and displaying in Realtime while speaking to IT and instantly showing results WE probably won't be there with 5.6 but maybe in a few years

[UhuBuhu @kecksbe]

@synthwavedd Not trusting that since Lama 4 benchmaxxing

[UhuBuhu @kecksbe]

@JuniperViews @Lentils80 Thats rly good If it's the price of 2.5

[UhuBuhu @kecksbe]

@String_The0rist @GaelBreton Not Sure about the Opus call Opus was great but Always needed steering. But with Codex i agree GPT 5 while beeing hated felt different than Opus making less issues and GPT 5.1 codex max high what a Shit Name felt Like the First model that i could really send free for a bit

[UhuBuhu @kecksbe]

@P4T0SH1 @CarstenG0705 @KryptoWolfGER Migräne als Beispiel aber darum gehts nicht es ist einfach dumm wirtschaftlich und es macht den menschen einfach das leben unnötig komplizierter

[Fix @FixlationAI]

@AnthropicAI Fable 5 will only be included for up to 50% of your weekly usage limits 😶

And it’s only available on plans for 6 days before being removed on July 7, then it becomes credits-only..

[UhuBuhu @kecksbe]

@AndrewCurran_ how big of numbers are we talking about like a 50% improvement or bigger?

[UhuBuhu @kecksbe]

@adonis_singh fabel but if we rly get full sol at 750token/s than i would go for that