Founder of Junwatu. Sharing AI engineering and creative experiments.
Remote
Joined July 2011
- Tweets21.7K
- Following2.7K
- Followers5.1K
- Likes37K
Pinned Tweet
Little Dragon.
You never know what you'll find in the woods.
Made with MiniMax H3 to save credits, while Seedance 2.5 handled the more complex scenes.
Created mainly with PixVerse.
#PixVerseCPP @PixVerseCreator @PixVerse #PixVerse #PixMoreChallenge
Carmen (Hearts2Hearts) ngaku di sela fan sign Jakarta kalau dia nggak peduli sama celah keamanan PHP, soalnya semua sistemnya udah dipindahin ke Rust sejak 2024.
"Fansku nunggu aku ke Jakarta 2 bulan. Tim dev nunggu rewrite Rust udah 30 bulan, masih santai aja. Ternyata aku lebih cepet," katanya.
DΞV retweeted
No tricks. Just tracks. 🎃
Sonilo’s Halloween giveaway is here:
- Free users: 1 month of Pro, free
- Active paid subscribers: 100,000 bonus credits that never expire
Follow @Sonilo_music, repost, & leave a comment to join.
Ends Oct 31. Let’s make something spooky.
AI creativity, it turns out you can do literally anything you want!
For this video, Seedance 2.5 still struggles a bit with the Javanese language, but visually, the result is already pretty close to what I imagined.
Now, as long as you have enough tokens, the only limit is your wallet! 😂
Created with PixVerse's Agent.
@PixVerse @PixVerseCreator
#PixVerseCPP #PixVerse #PixMoreChallenge
🤖 Made with AI
Eksplorasi Rumah Joglo 3D
Di Indonesia itu sumber inspirasi luar biasa banyak, mulai dari rumah tradisional, pakaian adat, artefak, dan masih banyak lagi.
Potensi edukasi, pelestarian budaya, sampai pariwisata itu seriusan gede banget!
Dan untuk eksperimen ini, saya pakai Gemini 4 Argon (Pro subscription) lewat Antigravity untuk bikin web app eksplorasi 3D dari rumah joglo.
Saya gak pake referensi apapun jadi di-handle sama Gemini. Akan lebih baik kalau misalnya ada real aset buat referensi.
(Thanks to @thebuggeddev for initial the source code)
🤖 Made with AI
Used my last credits on @Hailuo_AI to create this short animated movie.
The entire video was generated with MiniMax H3.
🤖 Made with AI
DΞV retweeted
grok bot 0.71.2 is out
- smarter notifications
- files up to 250 mb (1 gb for video), 20 per message
- forms arrive pre-filled
and a lot more!
x.ai/changelog/bot
ALT Grok Bot 0.71.2: smarter notifications and larger file attachments
ALT Smarter notifications: only on the device you're using
ALT Files up to 250 MB, videos up to 1 GB, 20 per message
ALT Also in Grok Bot 0.71.2: a list of smaller features plus lots of fixes
DΞV retweeted
Your iPhone has had a video studio button this whole time
Put Grok @bot on the action button, connect the HeyGen MCP, and one press turns a sentence into a finished avatar video
DΞV retweeted
Introducing FLUX 3 Max, the latest post-trained video model from fal Research in partnership with @bfl_ai.
FLUX 3 Max is tied for #1 in image-to-video quality on independent third-party benchmarks, generating a 10-second, 720/768p video in ~12 seconds.
For the next three weeks, FLUX 3 Max is 75% off, making it the most affordable frontier video model at its quality level.
DΞV retweeted
MiniMax M3.1 Flash, aka Space Bunny, is now available on GMI
Built for coding, agentic workflows, and computer use
70% lower agentic task costs, 73.7% on DeepSWE v1.1, 82% on CyberGym
Run it now 👇
MiniMax M3.1 Flash: Space Performance + Bunny Speed
🚀🐰 Yes, Space Bunny - which has topped global leaderboards with 200T+ tokens - was an earlier version of M3.1 Flash.
💰Redefining the performance economics of agentic AI. M3.1 Flash achieves frontier-level intelligence across key tasks such as Cowork (Automation-Bench: 62%; ALE: 30%), Coding (DeepSWE: 74%) and Visual Reasoning (BabyVision: 91%), at 1/3 the cost of M3 and 3x the speed.
API pricing (flat rate): $0.15 / $0.60 per 1M input/output tokens, with cache reads at $0.015 / 1M tokens.
Model weights will arrive on Oct 23.
API: platform.minimax.io
M Plan: minimax.io/m-plan
MiniMax Code for desktop: agent.minimax.io/download
DΞV retweeted
Step 5 Preview from @StepFun_ai is now available on GMI Cloud
A 600B-parameter MoE model activating just 27B per token, with a 1M-token context window and native text and image understanding.
Built for complex agentic tasks across coding, finance, and professional workflows.
Try it now 👇
Step 5 Preview is now open weights.
BF16 and FP8 weights are live on Hugging Face and ModelScope, together with our technical blog.
• Sparse MoE: 600B total, 27B active per token
• 1M-token context window
• Native text and vision input
• Built for long-horizon agentic work: coding, software engineering, professional knowledge work and finance
Frontier intelligence, open weights. Build what's next.
Hugging Face: huggingface.co/stepfun-ai/St…
ModelScope: modelscope.ai/models/stepfun…
GitHub: github.com/stepfun-ai/Step-5…
Sneak peek of Dunia Sekawan, a catalogue of Indonesian ghosts.
There are so many of them that I couldn't fit them all into one video, so I just picked a few that came to mind and included some basic information about each one.
The original ghost references are actually much more terrifying than what you see here, but I think the results turned out pretty great!
Video shots generated with HeyGen Video + Seedance 2.5. Reference images created with GPT Image 2.5.
Made with HyperFrames Studio.
@HyperFrames_
🤖 Made with AI
How to create it?
It's pretty easy, actually. Just tell the studio what you want to create. Here's the prompt I used:
Create a 16:9 horror video featuring Indonesian horror characters, 60 seconds long, 720p, using Seedance 2.5.
You can customize the prompt, add your own reference images or generate them directly in the studio. From there, you can add the title, description, and other details.
Also, I didn't use any special motion skills here, just the default settings.
One thing worth mentioning is that I've connected my studio to the HeyGen API, but you can also use Grok, Claude, or Codex with their respective plugins or skills.
