Using Mythos/Fable to center the div.

Opus 5 was the worst at writing and explaining its own work, which made it the worst model for most people because nobody could understand the shit it was writing😂. Opus 5.5 fixed that. It’s a monster now. Opus is back Claude is back.

If Opus 5.5 is beating Fable 5.1 on all the benchmarks then why did Fable 5.1 requires usage credits? :)

Opus 5.5 made this. Its just amazing. @claudeai @bcherny @shantanugoel

Just waiting for people to start frontend development using Jev. Who is up for it?

Spring AI and TypeSafe Jev. spring.io/blog/2026/09/21/sp…

Hey @grok how can I leverage Jev AI for portfoliointel.co.in Give actual use case that work in real scenarios.

Is it just me or it’s impossible to keep with AI tech trend nowadays. Everyday it seems like there is new discovery and we need to just adopt it. What the heck is Jev now @grok.

Every enterprise CEO/CTO, when asked how they use AI?

What happened to webmcp? levelup.gitconnected.com/web…

I'm 37. solo founder from India looking to connect with more creator & ai enthusiasts!

Is anyone really able to make money with AI?

I will assume AGI is here when the X search actually starts working.

GPT-6 Astra is wild. I burned my limit with just one prompt. Back to Fable 5.1.

Opus 5 is a disaster from Anthropic. I have switched back to Opus 4.6. Opus 5 has no regards to skills that were working fine before and now suddenly generates a Novel on each attempt. It has killed harnesses that we were working before and it fills context with garbage. @AnthropicAI @karpathy.

Looks like Aadhaar uncle was right about not building frontier models.

It's almost a year now since AI coding agents reached a level where you can literally build anything, yet why haven't we seen a billion-dollar company? It seems building or coding was never the bottleneck.

With AI. For your use case either it’s cheap as fuck or expensive as fuck.

I have been using GPT 5.6 and it’s amazing. Fable is not even close.