Erwan Novianto

[Erwan Novianto @akuerwan]

@nopisuhadianto @hanifproduktif pertanyaan yang sama

[Erwan Novianto @akuerwan]

ini gila sih qwen3.8 27b bisa 9x lebih kecil dengan kemampuan hampir mendekati normalnya

[PrismML @PrismML]

Today, we’re announcing Ternary Bonsai 2 27B.

Based on Qwen3.8 27B, Bonsai 2 27B is 9x smaller than its full-precision counterpart while retaining 98.2% of its aggregate benchmark performance.

Two months after the first Bonsai 27B release, the biggest change is quality. The footprint remains 5.9 GB, but the gap to full precision has narrowed materially, with particularly strong gains in agentic coding, multimodal reasoning, and long-horizon tool use.

Ternary Bonsai 2 27B is available today under Apache 2.0.

[Erwan Novianto @akuerwan]

https://t.co/h7DXubEaBa sangat2 murah hati, coding plan member dapat free glm 5.3 flash malam hari

[ZCode @zcode_ai]

GLOBAL BUILD is live 🌍🔥
GLM-5.3-Flash, FREE in ZCode. 10 hours a day. Sep 3 – 18.

Who gets it
👑 Coding Plan members — free every day, all 15 days. Our VIPs go first.
🥚 New users — 100M free tokens on sign-up. One-time, valid until the window closes

When it opens
🇺🇸 8 AM PT · 11 AM ET
🇬🇧 4 PM London
🇨🇳 11 PM Beijing

How to claim
1️⃣ Open ZCode → tap the card in the bottom-left corner 👀
2️⃣ Not there? Restart the app 😏

https://t.co/9y0HPmcBxz

[Erwan Novianto @akuerwan]

@airplanestar_ sampe macet begini wkwk, sampe sy chat baru gak bisa dicompaction klo dah gini

[Erwan Novianto @akuerwan]

@airplanestar_ betuul sama gitu

[Erwan Novianto @akuerwan]

@airplanestar_ iya udah 1M sih biasanya jarang nyampe 500k juga pake glm 5.3 flash, ini baru bbrp iterasi udah nembus 700k-800k aneh bin ajaib

[SSSS.CRYPTOMAN⚡️AI @SSSS_CRYPTOMAN]

なんとMiniMax H3 Maxの無料枠がかなり増えました！

✅無料のサインインで1日15回作れる
✅10秒���動画が生成可能
✅T2V,I2Vに対応
✅日本語の文字表現も完璧！

生成も爆速で楽しいのでこの機会に是非やりましょう✨
個人的にSora2の再来を感じています

👇ここから使えます
https://t.co/f4g2vumgYO

[Erwan Novianto @akuerwan]

@levifikri yang pertama kali terdampak AI ya programmer, kalau programmer gak embrace AI ya dipastikan perlahan akan sangat sangat tertinggal. Programmer skr bukan tukang coding tapi org yg paham bisnis proses, arsitektur tepat, tech stack tepat, and how to scale, welcome vibe coder

[Erwan Novianto @akuerwan]

@kangcahyakeren Install dua2nya dan pakai dua2nya sampai suatu hari agent gw macet2 klo pake omniroute padahal gak pake compression apapun, akhirnya lanjut pake 9router aja

[Erwan Novianto @akuerwan]

Nice qwen3.8 27b rilis, dunia ai open weight menggelora hebat

[Qwen @Alibaba_Qwen]

We promised open weights for Qwen3.8. Now, time to meet them! 🎉

⚡ Qwen3.8-27B:
- A native multimodal dense model. With just 27B parameters, it outperforms Qwen3.7-Plus overall and shines in real-world coding & office workflows.
- 262K native context, easily extendable to 1M tokens via YaRN.
- Built for builders. Highly efficient, high-quality, and licensed under Apache 2.0.

🚀 The open weights for Qwen3.8-2.4T-A95B (Max-level) have also been released recently.

Whether you're shipping lightweight applications with Qwen3.8-27B locally or building agents with Qwen3.8-2.4T-A95B, they're yours now!

Download, deploy, and build something we haven't imagined yet. 👀👇
- Hugging Face：
https://t.co/4kaAcqYEVj
- ModelScope：
https://t.co/eRIMZCGkhC

[Erwan Novianto @akuerwan]

@Cupang1997 Ekspresinya sih terlihat realistis tapi

[Erwan Novianto @akuerwan]

@Cupang1997 @direkturcrypto Haha keren bang, ditunggu tutor lipsync nya pake H3

[Erwan Novianto @akuerwan]

@direkturcrypto aeowjdiiwo2eown ai asowiehdkwi ai

[Erwan Novianto @akuerwan]

wah beberapa hari lalu baru rilis Minimax H3 sekarang LTX-2.5 rilis, dunia video ai makin seru nih

[LTX.io @ltx_io]

Introducing LTX-2.5, the world model the world builds on.

One of the biggest upgrades yet to the model already powering film, robotics, and real-time workflows. Higher pixel fidelity, multishot scenes that hold together across cuts, and a pretrained foundation built to be finetuned across domains.

Plus: Diffusion Fidelity Rendering, a new rendering approach built for pixel quality that holds up frame by frame, even on a cinema screen.

This isn't a tool you rent. It's a foundation you build on, open and yours.

[Erwan Novianto @akuerwan]

dan bikin sub agent ini juga gampang banget, waktu kita chat dgn agent utama dan dia discover pengetahuan baru, kita bisa langsung suruh dia bikin sub agent baru atau bikin skills baru biar bisa dipake di project lain, dan itu otomatis langsung jadi ga pake ribet

[Erwan Novianto @akuerwan]

ini fitur zcode yang paling berasa bedanya dengan code agent lain, kita bisa bikin sub agent sendiri dengan kemampuan beda-beda, nanti agent utama otomatis manggil sub-agent2 itu untuk kerja sesuai job desc masing-masing.

[Erwan Novianto @akuerwan]

Use case nya misal pas lagi nyuruh deepseek v4 flash tiba2 dia mentok pas liat gambar, nah itu dia otomatis manggil sub agent vision untuk minta bantu liatin gambar itu

[Erwan Novianto @akuerwan]

@airplanestar_ enak parah sih zcode betah asli, cuma kmrn buntu waktu disuruh install app desktop dia kayak ngerjain di sandbox intinya ada yg gagal build, akhirnya pindah sebentar ke opencode, selebihnya balik zcode

[Erwan Novianto @akuerwan]

@airplanestar_ Iya bener yg bagus memang begitu backup beda pulau, biasanya batam. Setauku data negara gak boleh host di luar negeri, jadi harusnya punya data center sendiri, instansi punya masing2 biasanya kerjasama dgn Komdigi juga ya