Ahmed Ibrahim

[Ahmed Ibrahim @le0ndaz]

@harry_ai_nerd @cursor_ai @Kimi_Moonshot Very cheap move

[Ahmed Ibrahim @le0ndaz]

@tak_dcxi https://t.co/tbVME4Wiq2

[jhey ʕ•ᴥ•ʔ @jh3yy]

a little css trick i like ↓

adapt the radius gradually based on available space

shrink from --r to 0 within --g 🧑‍🍳

.card__surface {
--r: 42px;
--g: 24px;
border-radius: clamp(
0px,
(100vi - 100cqi) / 2 *
var(--r) / var(--g),
var(--r)
);
}

[Ahmed Ibrahim @le0ndaz]

@thegenioo @blueemi99 It was available on most providers, not only opencode, Cline and many others

[Ahmed Ibrahim @le0ndaz]

@kiri49x86 @Alibaba_Qwen @arena wait for unsloth I guess

[Ahmed Ibrahim @le0ndaz]

@Da7_Tech how good is devin's quota compared to codex?

[Ahmed Ibrahim @le0ndaz]

@CDPROJEKTRED_IR Love you

[Ahmed Ibrahim @le0ndaz]

@NVIDIAGeForce I mean, why not #RTXPowersPlay

But, this is cheap for a multi-trillion company though, it should have been something more valuable

[Ahmed Ibrahim @le0ndaz]

@designwkarthick https://t.co/Csj2CM6xzR

[Ahmed Ibrahim @le0ndaz]

@designwkarthick Use it from: https://synthetic
new

This will get you to test it, but comtext is 500K here

[Ahmed Ibrahim @le0ndaz]

@Cyrusontherun @togethercompute DS is definitely faster, GLM is 2.5x the size and 3x active params, not to mention DS has vision

But, GLM is way smarter, premium vibe, pretty much knows everything from the first glance, this was my experience with GLM5.2 working on a math-rendering heavy app (vulkan, metal)

[Ahmed Ibrahim @le0ndaz]

@hqmank If you mean the $20 then I will assume you're not a heavy user, I consumed 3% of the weekly in ~10 minutes so I refunded😭

[Ahmed Ibrahim @le0ndaz]

@hqmank Which plan?

[Ahmed Ibrahim @le0ndaz]

@EmperoAI This is around 1 token per business day, I think people are hammering it?

[Ahmed Ibrahim @le0ndaz]

@CreeCoder Good luck with it, I will support for sure

[Ahmed Ibrahim @le0ndaz]

@rezoundous Use https://t.co/z2pBdVgll9 which is exactly what you describe

[Ahmed Ibrahim @le0ndaz]

@togethercompute Can you show the time taken as well?

[Ahmed Ibrahim @le0ndaz]

@MarcosHernanz It sucks and I get angry at it almost hourly

[Ahmed Ibrahim @le0ndaz]

@ahmd3ssam GLM, Kimi are my goto models rn but Codex is the superior one in this chart imo, their limits suck though