Blake s

[Blake s @S86067S]

@matthewcp There are many people that are unaware agentic coding harnesses exist who still do this

[Blake s @S86067S]

@lectroidjohn @alextoussss Bats aren't flying around my yard during the daytime when mosquitos are about to bite me.

[Blake s @S86067S]

@_mateothegreat @kimmonismus Why wouldn't it be 5? When the first Opus version came out it was 3.

[Blake s @S86067S]

@pvncher @ignacioaal What would also be useful is a subagent auto-selection mode where you can define in advance what types of tasks should map to specific subagent models and reasoning levels, so that it doesn't need to guess. Since everybody has different workflows and preferences.

[Blake s @S86067S]

@TesseractUnfold @AWSNewsroom @OpenAI Try reading the first sentence, it's from openAI not Amazon.

[Blake s @S86067S]

@BooBakerBiz @alex_verem Get back to work wageslave

[Blake s @S86067S]

@66cjg @theinformation The capabilities are jagged, so in some areas it's already far past human level, in other areas it's lower than human level, so it's not an apples-apples comparison in the first place

[Blake s @S86067S]

@adamawadd @kimmonismus More than enough for you doesn't mean more than enough for me

[Blake s @S86067S]

@TinoWening That strategy won't work for open-ended tasks like network/kernel troubleshooting, where the issue could end up being very simple or very complex and usually you won't know until afterwards.

[Blake s @S86067S]

@stark4833 @gdb Did they promise that by this year?

[Blake s @S86067S]

@DanielLockyer You forgot Pro, with standard/extended

[Blake s @S86067S]

@thebuggeddev The benchmarks measure objective metrics, and this task is subjective, so I don't see the connection. And this model isn't known for its UI design abilities in the first place.

[Blake s @S86067S]

@johncrickett @svpino The reality is some people are effective at using it and some people aren't

[Blake s @S86067S]

@plotarmordev @kimmonismus Because as the models get more capable they'll be used researchers for things like drug discovery and curing diseases, which would still benefit anyone who isn't using the models themselves

[Blake s @S86067S]

@andrewlee07 @kimmonismus Are you saying 5.5 is good enough to perform every possible task?

[Blake s @S86067S]

@AcePapers96261 @ChrissGPT https://t.co/fRr891qAbs

[Elon Musk @elonmusk]

Grok 5 will be out before the end of this year and it will be crushingly good

[Blake s @S86067S]

@SaronAdir @twlvone @thsottiaux You can use it for anything that can be done on a computer

[Blake s @S86067S]

@kimmonismus No it doesn't, Fable 5 doesn't have the same abilities as Mythos

[Blake s @S86067S]

@tszzl Well we probably need to keep our brain in order to retain our consciousness

[Blake s @S86067S]

@chetaslua @Zai_org Are the gears just a visual effect or is it actually using the physics of the gears to control the time?