I eye AI | Making the most of what AI has to offer | Always looking for the next big thing | Follow to keep an ποΈ on the latest Tools, Tutorials & Prompts.
Remote
Joined April 2023
- Tweets35.8K
- Following255
- Followers39.2K
- Likes64.3K
Pinned Tweet
MiniMax just turned H3 into an entire AI production studio.
You give MiniMax Design an idea, and it can plan the shots, control cameras in 3D, generate with H3, edit the footage and even add subtitles.
Basically: prompt β finished content
Hereβs what that looks like β
Honestly, AI video workflows are getting ridiculously good.
Blender, Seedance 2.5 and After Effects can now work on the SAME video project, without losing context between tools. π€―
No more starting from scratch every time you move to another app.
I came across this setup in YouArt Desktop, and it's so useful.
Here's how it works:
β³ Build a 3D scene in Blender
β³ Generate shots with Seedance 2.5
β³ Polish everything in After Effects
And it's all connected through ONE canvas.
Your prompts, references, and previous versions follow the project, so you're not explaining the same thing every time you switch tools.
You can still edit things manually in AE and finish the cut in CapCut, Premiere or Resolve.
I'd love to see more AI video workflows work like this.
Your AI agent is about to face its WORST nightmare: users trained to break EVERYTHING. π
Plurai just launched a FREE simulation tool for Claude Code, Codex, and GitHub Copilot.
Point it at your repo.. it creates a digital twin:
β Your actual agent, unchanged
β Its tools & environment recreated from your code
β Simulated users trying to break it
It catches failures, shows what went wrong + helps verify fixes, all in isolated tests.
Better to find the bugs here than have real customers find them for you.
π€ Paid partnership
Big day for us at @pluraiAI ! Plurai Simulation is live and free for every AI engineer.
AI agents look great in demos. Until real users break them.
So we built a world where your agent can fail, learn, and get fixed before your users ever see it.
β Digital twin: Your real agent, its tools, and environment recreated from your code
β Simulated users & tasks: Realistic users, tasks, and workflows that push your agent into the edge cases that matter
β Auto evals: Check what your agent actually did, not just what it said
β Verified fixes: Every failure comes with the run that exposed it and a fix you can verify
β Replays: Turn any production failure into a simulation and replay it against every new version of your agent
All inside the coding agent you already use: @claudeai code, @OpenAICodexCli , or @GitHubCopilot .
Start simulating for free: simulation.plurai.ai/connect
Claude JUST launched its new productivity tools, and someone already dropped FREE, open-source alternatives.
Claude has Whiteboards, Sheets & Data.
Now puppyone has released its own:
β Whiteboard
β Spreadsheet
β Dashboard
All FREE. All open source. Works with ANY AI model.
And you don't have to stick with Claude to use them.
Talk about perfect timing. π
Check them out: go.puppyone.ai/4zdUcXk
A 720p Seedance 2.5 render on YouArt Desktop's Max yearly is $0.138 a second. I've paid fal full price for every one of mine so far, so seeing it come in 71% under that stings a bit.
π€ Paid partnership
Introducing the YouArt Desktop App: your AI creative teammate on your computer.
It uses the best LLMs like Fable 5.2 to make creatives. It can operate creative software on your computer, including CapCut, DaVinci Resolve, Adobe Premiere and many more.
It brings client feedback onto a native canvas, and you can ask agents to address all the comments while you have a coffee.
Itβs the most cost-efficient video agent. It supports all third-party API keys and uses smart routing to send every clip to the cheapest provider.
Have fun!
Some AI videos look so damn cinematic, I genuinely wonder what prompts people are using. π€―
Now you can upload those videos to Dreamina, break down their visual style..
and try your own ideas with Seedance 2.0.
Plus, eligible CapCut Pro users can unlock up to 98% off Dreamina Basic, free generations & bonus credits.
Here's the trick: π§΅
First, a quick heads-up.
Dreamina is ByteDance's AI creation platform, where you can generate videos and then edit them in CapCut.
Already have CapCut Pro?
Open the link below, hit "Sign in to reveal," and log in with your eligible account.
You'll see your rewards, followed by the membership payment page.
1
12
157
Unsolved problem, and IFM doesnβt claim otherwise. A base model, open, limits stated.
Iβd call this a research release and treat it like one.
π€ Paid partnership
Introducing PAN 2: an open foundation world model for physical intelligence.
A language model predicts the next word. PAN 2 predicts the next world: what happens after each action, grounded in time, physics and space.
Towards AI that simulates βwhat if?β before it acts.
More details: ifm.ai/pan/
If you're already making AI videos, there's a pretty obvious way to make money from that skill.
Start making product ads for small businesses.
You don't even need a physical product to get started.
A product photo and a decent prompt can get you pretty far.
Here's how I'd do it:
β Upload a product image to Dreamina Marketing Studio and generate an ad using a prompt.
β Open it in CapCut PC and use AI Edit to fix anything that looks off.
β Use AI Extend if a product shot ends too early.
β Add voiceovers, music & text using Multi-Track Editing.
Make a few sample ads for different products, put together a small portfolio, and start reaching out to businesses.
Even local shops or small online brands could be potential clients.
Worth trying.
Dreamina: bit.ly/4k0KTGe
CapCut PC: tinyurl.com/2v23rn22
#CapCutPC #CapCutDidThat #Dreamina #Smallbusinessmarketing #capcuteditpilot
π€ Paid partnership
Robots are getting a sense of touch, and the amount of extra AI needed is surprisingly tiny. π€―
According to Robbyant, LingBot-Touch-Tiny adds less than 0.2% to the parameter count of the models they tested (T-Rex, Ο0.5, and LingBot-VLA).
Such a small addition, yet it makes a real difference in how robots handle objects.
π€ Paid partnership
What changes when robots can feelβnot just see? π€β
Meet LingBot-Touch, our open-source tactile foundation model towards dexterous manipulation, pretrained on nearly 5,000 hours of real tactile data.
Through unified heterogeneous tactile modeling, LingBot-Touch supports transfer across sensors, dexterous hands, tasks, and action models.
Watch it in action. π #LingBot #EmbodiedAI #Robotics
There's always that ONE shot in a video you wish you could redo.
I can think of so many clips I'd love to go back and fix using @ViggleAI Remake.
Introducing Viggle Remake.
The most powerful video transformation agent. And weβre releasing the weights today.
β recast every character in a scene, and they stay consistent across cuts
β swap objects, change the background, restyle or add VFX
β up to 2 minutes
β keep chatting to fix any shot until itβs right
From $0.96 per 10 seconds at 720p, less than half of what Higgsfield Genjutsu charges.
Try it now: viggle.ai/meme/app/home?ui=vβ¦
Open weights: huggingface.co/Viggle/Viggleβ¦
You can tell this was built by people who've spent real time in a lab.
The test is exactly what a scientist would ask: what does the cell do next?
Very clean launch @celltypeinc π
Can LLMs simulate biology?
We tested frontier models on real lab experiments: predict how cells respond to drugs, gene knockdowns, and immune signals. Even the best get only part of the response right. Lots of room to improve.
Introducing BioSimBench, built on @NVIDIAAI. π§΅
Me: Just one tiny edit and I'm done.
Also me, 40 minutes later: still editing. π
It's crazy how fixing one small thing in a product video can take forever.
Say there's something distracting in the background, or one part of the clip needs fixing.
Instead of going through all the editing tools manually, I'd just use Edit Pilot in CapCut PC.
Tell it what needs fixing in plain English, and it can help make that specific change.
Really handy when you just want to finish the edit and move on with your day.
π tinyurl.com/2v23rn22
#capcutpc #capcutmarketing #capcutblackfriday #CapCutDidThat
π€ Paid partnership
I made an AI agent plan a $3,600 Greece trip for two.
Then asked it to plan another trip to Portugal.
Guess how much of the workflow I had to rebuild?
ZERO.
I used Teamily AI's team agents to handle the research and destination comparisons while my teammate and I reviewed the options.
It reminded me of this line from Summer's piece:
"Humans + Agents. One Team."
Here's what I did:
β Created a Travel Research Agent
β Built a Loop with the agent, me, and my teammate
β Asked it to research a 5-day Greece trip and rank the best options
β Reviewed the shortlist together in the group
Then reused the SAME Loop for Portugal.
All I had to do was change the trip requirements and run it again.
No repeating instructions or setting up everyone's roles twice.
I'd much rather have the agent do the research and leave the actual decisions to us.
This is how I built and ran it in Teamily AI: π
Join me on Teamily, the human-agent platform, where you and always-on agents get work done together in one unified and context-aware space.
Redeem my code in Settings - Credits Center and you'll get 5 CREDITS right away.
Code: 86GM9N
teamily.ai/i/86GM9N
3
164
Last chance to grab CapCut's Black Friday deals!
October 30 is the final day, with up to 60% off Pro and 40% off Ultra on CapCut PC.
If you've had your eye on either plan, today's your last chance.
More on the deals and AI editing features below:π§΅
π tinyurl.com/2v23rn22
Say you need to make three UGC-style product ads in a day.
I'd start in Design Studio by creating an AI presenter with the product, then use Video Studio to generate the demo clips.
But what if the audio has background hiss, the voice isn't clear, and there's a random cable showing in the background?
Here's how I'd fix those.
1
11
191
A little girl is alone in her room.
At least, that's what she thinks.
Then a man's voice enters the story... and things start getting creepy.
I created this short AI horror story called "My Other Self" and left it on a cliffhanger.
And now anyone can decide what happens next.
I came across Ollang's Halloween experiment called The Longest Night and decided to give it a try.
Here's what I did: π
β Created a 41-second horror story and uploaded the audio
β Used AI dubbing to play it in Hindi
β Added a second character with a male voice using "Add to the story"
β Continued the story with another audio clip
β Played back the updated version in different languages
I also found out Ollang has APIs that developers can use in their own apps.
So if you're building something that needs dubbing, translation, or content in multiple languages, you can automate that through Ollang instead of doing everything manually.
Pretty useful beyond this Halloween experiment.
And coming back to my story, you can actually add your own continuation.
Introduce another character, record a new scene, or give it a completely different ending.
I've done my part. Let's see how much worse you guys can make it π
Continue "My Other Self" here:
playground.ollang.com/longesβ¦
2 BILLION people. All searchable by AI agents.
Ploid has built a massive index of people that you can search in plain English.
Find potential clients, pull their emails, work history, socials.. and even check the original sources.
I'd actually love to plug this into my own AI agent and see how much of my client research it can handle.
Check out their launch below.
Introducing Ploid.
I think the world's data should belong to everyone, so we indexed over 2 billion people globally.
Ploid turns the public web into one profile for every person and company on earth, and serves it to you and your agents for cents.
We're mapping the human economy, and we're live today. ploid.com
A good candle ad is all about the little details.
The soft lighting, warm colors... and even the sound of a match being lit.
You can create that cozy evening look in Dreamina, then bring it into CapCut PC to add sound effects and time the match strike perfectly with the flame.
And right now, both have some pretty sweet deals.
β Dreamina Basic: Just $0.29 for the first month
β CapCut Pro: 60% off
β CapCut Ultra: 40% off
CapCut's Black Friday offer runs from October 20β30 on desktop only.
Pretty good time to experiment with some new creative ideas.
Grab the deals here:
CapCut: tinyurl.com/2v23rn22
Dreamina: tinyurl.com/5xbtv4x6
#CapCutDidThat #Dreamina #Smallbusinessmarketing #CapCutSale
π€ Paid partnership
You can now ask ChatGPT to find the perfect music, effects, and transitions for your video and add them straight into CapCut.
No more spending ages searching through asset libraries.
CapCut's new ChatGPT plugin makes this possible.
Here's how to try it: π
1. Open CapCut Web and go to Video Editor.
2. Click the plugin icon in the top-right corner.
3. Copy the installation prompt.
4. Open the ChatGPT desktop app on your PC, paste the prompt and send it to install the plugin.
5. Describe the video you're making and let ChatGPT help you find the right assets.
The library also includes licensed assets for commercial projects, which is very useful if you're creating product ads or branded videos.
This workflow was previewed at CapCut World alongside creators and OpenAI.
It's currently in early access on ChatGPT desktop for PC only, with mobile and web support expected in mid-November.
I came across an AI platform doing something pretty wild.
It's basically a cheat code for anyone burning through credits on top video models like Wan 2.2, LTX 2.3 & MiniMax H3.
β’ Render speed: Up to 17x faster (115s β 6.7s)
β’ Generation cost: 50% cheaper ($4.80 β $2.40)
Quick breakdown π§΅
MachGen/ @MachgenAI is not building its own video model.
It's focused on making the existing ones run a lot faster and cheaper.
The idea is pretty simple. Get the same quality videos without spending as much time or money generating them.
And you don't have to change your prompts or the way you already work.
I've seen some insane Seedance generations lately.
So when I saw Dreamina offering a whole month of Basic for just $0.29, I genuinely had to check twice.
And you get 1,575 credits too.
How are they making AI video generation THIS cheap?
Let's break it down. π§΅
π€ Paid partnership
So I looked into what you're actually getting for $0.29.
β A full month of Dreamina Basic
β 1,575 credits for AI content generation
Not bad at all.
But I'm curious about how far those credits can actually go.
Since different models use different amounts of credits, the number of videos you can generate will depend on what you're creating.
#DreaminaOnly29Cents #Dreamina
1
4
160