Ravi Edho Nugraha

But a thing of beauty will never fade away~

[VALORANT @VALORANT]

celebrate your PRIDE with these rewards - link in bio.

Reward Codes:
CC-VAL26-HOLDN-SPACE
CC-VAL26-ALLUV-SPRAY
CC-VAL26-PLAYR-CARDS

[Rooster @TheRooster]

"사" is the Korean word for the number 4.

But it's also a homophone for 死 (death).

Not a coincidence.

Korea has fear of 4. It's called tetraphobia:
- People avoid it like the plague.
- Buildings skip the 4th floor.
- Elevators use "F" instead
- No license plates with 4s

And COD just dropped that title in a trailer full of real South-North DMZ tension with the line "No line holds forever."

That’s bold as hell. First time the reboot is swinging hard at fresh geopolitical stuff with real cultural weight behind it.

[witcheer @witcheer]

https://t.co/4oR4pVnFlD

[AboveSpec @above_spec]

Qwen3.6 35B A3B model. 55+ tokens/sec. $300 GPU.

No, this isn't a server card. It's an RTX 4060 Ti 8GB.

Previously I posted that I 41 t/s on this gpu and that post blew up and went viral. I went back and made it 34% faster.

And now the speed doesn't drop with context depth at all.

New benchmarks + what changed 🧵

[left curve dev @leftcurvedev_]

Anyone with 8GB or 12GB VRAM setups needs to understand that "-ncmoe" is the key flag to boost performance on llama.cpp

Here are my results for Qwen3.6 35B A3B, with 64k q8_0 context on a 8GB RTX 3070Ti:

⚪️ no flag → 8.7 tok/s
RAM: 13.6GB & VRAM: 7.8GB

🔴 -ncmoe 35 → 27.5 tok/s
RAM: 12.1GB & VRAM: 4.3GB

🟢 -ncmoe 30 → 32.5 tok/s
RAM: 12GB & VRAM: 5.6GB

🔵 -ncmoe 25 → 40.9 tok/s
RAM: 12GB & VRAM: 6.9GB

Please note the ram and vram usage you see are total usage of a windows pc, with the model running. My friend's setup: 8GB VRAM and 16GB RAM. You can boost performance by switching to Linux, just something to keep in mind.

Basically, this flag keeps the MoE experts in the first X layers on your CPU + RAM, instead of eating all your VRAM straight away. This is a smart hybrid offload way that lets you run bigger models without OOM while keeping the rest on your GPU for speed.

As we can see on the data, there's a sweet spot. When we lower it from 35 to 25, speed bumps +50% because there are more layers on your GPU (look at the VRAM usage). The key here is to play around with the number and fit as much as possible on your VRAM, goal is to have 1GB/800MB headroom to avoid stress.

↓ server flags below

[left curve dev @leftcurvedev_]

I've been asked to share all the prompts I used to compare Qwen3.6 vs Claude, so here they are 🎸

One-page HTML canvas, no libraries allowed

↓ 12 prompts below

[Maor Elkarat @Maor_Elkarat]

Stop buying more VRAM.

Everyone’s posting Qwen 3.6 configs running insanely fast on 12GB cards.

But do you actually understand the flags making it possible? Weights are only half the story. KV cache is eating your VRAM alive.

The secret isn’t just 4-bit weights it’s the KV cache sorcery everyone’s missing.

Here’s the annotated command & real tricks explained:
@elonmusk @grok #Ai

[Jason 𝑻𝒉𝒆 𝑮𝒆𝒓𝒎 𝑮𝒖𝒚 Tetro 🇨🇦 @JATetro]

Overheard behind me while watching #H5N1 NAS workshop today: "You want to really make a pandemic, make hantavirus human transmissible"

[Tanyarlfes @tanyarlfes]

💚 Ini dia emang emotional intelligence nya bagus atau emang ga sayang ya? 😭

[Tanyarlfes @tanyarlfes]

Wajar ngga sih aku kesel? Sebelumnya aku udh bilang ga kmna2 dia malah nanya bgtu, trs aku jawab “ga” balesan dia ga sesuai ekspetasi aku, harusnya klo emng dia niat gausah nanya dulu ga sih? 💚

[Erkekler nasıl giyinsin? @nasiIgiyinsinIr]

Beyler kaydedin lazım olur 👍🏻

[Zan @kapten__monyet]

Man to man

Jangan jadi cowo “bingung”

[Funtastic @tempecx]

Menormalisasikan hubungan secara dewasa

1. Kalo chatnya dibalas lama, berarti ada pekerjaan lain yang jadi prioritas.

2. Belajar tenang dan tidak tantrum ketika ditinggal kerja, nongkrong ataupun main.

3. Tidak lagi menuntut waktu 24/7 karena sadar hidupnya bukan tentang kita aja.

4. Tidak memaksa ego dan memberi ruang untuk me time.

5. Tidak silent treatment dan menyelesaikan masalah secepatnya.

6. Jika ada yang masalah, diselesaikan dengan cara baik-baik.

7. Selalu mengusahakan komunikasi yang baik. pagi, siang, sore dan malam sesibuk apapun di kantor. gak pernah membuat pasangan ngerasain feeling lonely.

[Dexerto @Dexerto]

Streamer Sykkuno's response to cheating allegations has been dubbed over by 'Deku' actor Justin Briner

[M1das @M1das_OW2]

there’s a reason fortnite can’t do this anymore and yall aren’t gunna like it

[D @fearlessbarb13]

Cara aku deep talk ke pasangan, untuk cari tau traumanya, tanpa maksa atau nyenggol egonya adalah..

Misal pas lagi santai depan TV,
“Bekas luka di tangan kamu itu kenapa?”
Dia bakal mulai cerita.. “Oh iya, dulu waktu kecil aku jatuh, bla bla bla..” lalu ceritanya bisa ngalir ke mana-mana.

Atau misalnya aku pengen tau kenapa dia boros banget, daripada langsung nyecer, “Kok kamu beli barang itu terus sih?” itu malah bikin dia defensif.

Aku mulai dengan percakapan,
“Ada gak barang impian yang kamu pengen banget dari kecil? Udah kesampaian belum?”

Dari situ biasanya dia akan cerita sendiri,
“Dulu aku pengen banget ini itu, tapi belum bisa beli. Baru kebeli pas…”

Nah, dari cerita-cerita kecil begitu, kita jadi bisa lebih paham alasan di balik kebiasaan, kesukaan, bahkan luka lama yang mungkin selama ini gak pernah dia ungkapkan.

Deep talk itu gak dimulai dari maksa cerita, tapi dimulai dari obrolan ringan, tanya dengan santai, klo dia mau jawab ya bagus, klo enggak juga gpp, klo dia mau cerita, tolong dengerin tanpa menghakimi.

Klo mau tau soal ketakutan, trauma, atau pengalaman masa kecilnya, bisa mulai dari pertanyaan yang santai juga, misalnya:
“Waktu kecil kamu paling bandel ngapain?”
“Atau ortu kamu dulu galak gak?”

Lalu bikin dia merasa aman, dengan cara kamu bisa sharing sedikit tentang masa kecilmu misalnya, “Oh gitu… aku dulu waktu kecil juga…”

Jadi suasananya bukan kayak wawancara, tapi kayak saling buka diri pelan-pelan. Biasanya justru dari situ timbul rasa percaya, dan dia juga merasa aman buat cerita, karena tau dia dingertiin. Hubungan jadi makin dalam setelahnya.

[忘れたくない黒歴史bot @dreamunader1982]

夜中に見て涙止まらん…😢😢
ほんまに理想すぎる

[CoCo @coco_0067]

Being loved in a dominant way by a woman is so hot