500K+ AI Community | AI Creator | Sharing AI news, tools & tutorials | DM or Mail for Collabs ↓ 📥 hossainalif522@gmail.com
Australia
Joined January 2023
- Tweets64.2K
- Following1.5K
- Followers108K
- Likes65.8K
Pinned Tweet
FREE VERSIONS VERSUS PAID VERSIONS
Do not pay for Spotify, use Lyra
Do not pay for Apple TV, use Streamly
Do not pay for Netflix, use Flixio
Do not pay for Paramount+, use EpicFlix
Do not pay for Apple Music, use ESound
Do not pay for Peacock, use GlitchTV
Do not pay for Disney, use Netmirror
Do not pay for Prime Video, use CineHub
Do not pay for HBO Max, use Moviebox
Do not pay for YT Premium, use Brave browser
Do not pay for Hulu, use ShowZone
(Why do you pay for content you can get free?)
I'm curious how many developers have actually measured whether their AI coding agent improves on their codebase over time, rather than just feeling more productive.
🤝 Paid partnership
OpenClaw, Hermes Agent and Prime Agent don't meaningfully learn from experience.
On Continual-ARC, our new benchmark at @letsunifyai, none of them were able to memorize anything of value and none were able to generalize to new tasks. Training a small 9B model's weights did both.
A small marketing budget shouldn’t disappear into manual voiceover cleanup.
Dreamina Marketing Studio builds product visuals and explores campaign concepts. Bring the assets into CapCut PC — CapCut Agent EditPilot lets you select several clips and describe the fixes (filler words, transcript mistakes, pauses).
Create with Dreamina. Polish faster with CapCut PC.
Dreamina: tinyurl.com/376c8r3x
CapCut PC: tinyurl.com/2v23rn22
#CapCutPC #CapCutDidThat #Dreamina #Smallbusinessmarketing #chatwithcapcutagent
YouArt Desktop hooks into the APIs you're already paying for, and it has its own built-in models too, all on the same canvas.
I didn't have to leave anything behind to try it, and my project stays in one place whichever provider runs the job.
You pick the model and the router finds the lowest price for it. If you like your current provider you can keep the account. When another provider cuts its price, your next job gets that price automatically, with the same model and the same quality.
You're never locked into one provider, which I can't say about much else I use.
One prompt in Qoder, and Qwen4-Max-Preview turned my idea into a playable neon space game.
I've been curious about what Qwen4 could do, so I decided to test it by building Neon Drift, a game where you dodge asteroids while the speed keeps increasing.
It came with 3 difficulty levels, crash explosions, and even a high-score saving feature.
My best run so far: 284. The 61 run? Let's not talk about that
What I enjoyed most was seeing a simple idea turn into a fully playable game from just one prompt.
You can try it with a limited-time 80% discount!
Give it a prompt and see what you can build.
🤝 Paid partnership
I have a bunch of small tasks I keep saying I'll do later. Canceling subscriptions, following up on invoices, preparing for calls.
@n8n_io just launched Nod, an open-source personal agent that can actually handle these jobs using its own computer and browser.
I like the idea of handing off those annoying little tasks instead of spending my time doing them myself.
🤝 Paid partnership
Sending messages with AI agents is easy. The tricky part is what happens when someone replies hours later and your agent isn't running.
@inkbox_ai is making that easier by giving agents their own email, phone, iMessage, and WhatsApp, plus a computer to run on.
The agent wakes up when a message comes in, handles it, and goes back to sleep when it's done.
I like this approach. You shouldn't have to keep an agent running 24/7 just to make sure it doesn't miss a reply.
🤝 Paid partnership
We raised $5.3M to give AI agents a computer to run on, with their own email, phone, iMessage, and WhatsApp
Led by @peakxvpartners, with @rauchg, @sternhenri, @jefrankle, @cjc, and @daniel_levine
Today we’re launching hosted agents on @inkbox_ai. Your agent runs next to its communication channels, so it never misses a reply. It sleeps when idle, so you don’t pay for downtime.
In the 4 months since our YC launch, we’ve processed millions of messages for AI agents.
Making an AI model faster is great. Doing it without adding more GPUs is even better.
That's what caught my attention about @hedra_labs' new Model Factory.
They say Krea 3's 2K image edits went from around 22 seconds to just 3.6 seconds in under a week, using the same four AMD MI355X GPUs.
Getting that much more out of the same hardware is the kind of AI progress I like seeing.
🤝 Paid partnership
Introducing Hedra Model Factory: an autonomous inference and training platform for building models for the real world.
The recursive self-improvement loop behind frontier research has been locked inside a few labs. We're opening it to teams in robotics, manufacturing, biology and defense.
Our researchers come from Stanford, NVIDIA and Meta Superintelligence Labs. Now we work alongside your team to post-train and optimize your models.
AI forgetting things is annoying. But when agents start handling real tasks, it becomes a bigger problem.
Imagine a customer was promised something on a call last month, but the AI only sees today's email. It has no idea what happened before.
That's why I like what @GraphonAI is building.
Instead of just finding matching text, Graphon connects information across conversations, documents, images, and videos.
They also report 94% on LoCoMo using fewer tokens than full-context approaches.
I think this kind of memory could make AI agents a lot more useful.
🤝 Paid partnership
After years working on robot memory at @NASAJPL, @GoogleDeepMind , @Meta and @nvidia, @clarkzhango and I kept asking: why does AI still forget?
Today, with @deepakmishraVC, we’re announcing @GraphonAI and our $35M Series A just 6 months after our seed.
• #1 on ATM-Bench, ATM-Bench Hard
• 94% on LoCoMo
• 84% on DolphinBench, 10x cheaper
• A new cost-accuracy frontier across 7 benchmarks
businessinsider.com/graphon-…
Dreamina just partnered with CapCut to unlock a whole new playground for CapCut Pro members. 🚀
dreamina.capcut.com/ai-tool/…
Every eligible creator walks away with rewards: up to 10 free Seedance 2.0 generations, extra bonus credits, and a full month of Basic for a crazy $0.29 if you are brand new to the Dreamina platform.
#DreaminaPartner
Dreamina is the powerhouse AI creation platform built by ByteDance, the exact same innovators behind CapCut and Seedance. 💡
Hit the link below and tap "Sign in to reveal". Just make sure to use the login connected to your active CapCut Pro membership.
You are guaranteed all three reward types, though the exact amounts change with each draw. Expect up to 98% off Basic, completely free 720p generations for any supported length, and some sweet bonus credits.
1
2
1,237
Small marketing team. Suspiciously large content calendar.
From Nov 12-18, purchase or upgrade to Ultra on CapCut Web and get 7 days of unlimited image generation on select models.
Explore campaign looks, compare variations, and save the best assets for upcoming posts. Your 'creative department' of one finally has options.
capcut.com/my-edit?utm_mediu…
#CapCutWeb #CapCutWebDesignStudio #AIImage
The part of Emma that interests me isn't the realistic face. It's the timing.
@nuance_ai is trying something different with Nuplex. Emma can react while you're still talking, using expressions and visual cues to make the interaction feel more natural.
Imagine using this for language practice, virtual interviews, or game NPCs that can actually follow the flow of a conversation.
Making an AI look human is one challenge. Making it understand when to listen, react, or speak is a completely different one.
I think that's the bigger breakthrough worth exploring.
The first AI persona you’ll actually want to meet face-to-face.
Most AI personas stare blankly while you talk. Emma reacts while you’re still speaking.
After 18 months of research at @nuance_ai, this is Emma, built on Nuplex, our full-duplex audiovisual foundation model.
I know what sounds wrong in my UGC ad. Finding the button for it? That's apparently a separate degree.
In CapCut PC, I use Auto Caption on the product explanation, then notice the same misspelling across several talking clips. I box-select those clips on the multitrack timeline and tell EditPilot, in one simple sentence, which caption spelling to correct across the selection.
I don't need to know the name of every editing control to describe the change. I preview the corrected captions right on the timeline and adjust any line that covers the product.
Try editing your next UGC ad with a simple chat in CapCut PC.
tinyurl.com/2v23rn22
#capcutpc #capcutmarketing #capcutblackfriday #CapCutDidThat #capcuteditpilot
Your password was never stolen. Your 2FA was never cracked.
Yet someone could still get into your account.
Sounds impossible? It's not.
Hackers can sometimes bypass the login process entirely by stealing and reusing an active session you've already authenticated.
No password needed. No verification code required.
This is called **session hijacking**, and it's a security risk more people need to understand.
Here's how it works (and how to protect yourself):
Alibaba, the company behind Qwen, has a coding agent called Qoder.
They just dropped a new stealth model called Qoduck.
Curious to see how it performs on real coding tasks.
You can try it for free for a limited time: qoder.com/
I thought I was making a product ad. Somehow I ended up directing a whole room of people fighting over ONE thing. 😭
The best part is watching one person notice it, then seeing everyone else follow.
Here's how I put this chaotic concept together with CapCut PC: tinyurl.com/2v23rn22
#capcutpc #capcutmarketing
The moment everyone turns around needs to happen FAST. I wanted that sudden shift from a quiet room to a full-on product frenzy, but my transition wasn't selling the reaction.
Instead of manually digging through the timeline, I used Edit Pilot in CapCut PC to describe the transition adjustment I wanted and refine the existing edit.
Giving an AI community something playable to build sounds like a good way to get wonderfully strange ideas.
Combos Fun is an AI gaming community where you can create with Claude Opus 5.5 and GPT-6 Astra, then let other people play the result.
If GPT-6.2 gets added, I wonder what this crowd could come up with next.
@CombosFun_AI’s Minecraft Challenge is open. Make something, submit it, and see which ideas people enjoy.
Use my link to claim free credits and create your first Minecraft game with Combos Fun:
combos.converge.ai/events/bl…
Your world. Your rules. Your game. 🧱
Build your own block world on Combos and join the Block World Challenge.
🏆 Win up to $10,000 in prizes
🎮 Solo or multiplayer
📅 Submit by Nov 22
Build yours here ⬇️
combos.converge.ai/en/events…
AI video models are getting impressive. But there's something I've been wondering about.
How do we actually measure whether they can respond to instructions while a video is still being generated?
That's what StreamEval is trying to answer.
It's an open-source benchmark for interactive streaming video generation, with a leaderboard covering three areas:
1. Avatar — how well digital humans handle dialogue, movement, and instructions.
2. World Generation — whether models can generate scenes while responding to new requests and keeping things consistent.
3. Video Editing — how well models can make changes to a live video without breaking what should stay the same.
The interesting part is that StreamEval doesn't just look at how good a generated video looks.
It also measures whether the model follows instructions, maintains consistency, and handles interactions over time.
The research evaluates 38 models across 300 test cases, with work involving researchers from Tsinghua University, NVIDIA, CMU, UC Berkeley, Princeton, NUS, and Reactor.
And the benchmark is open-source. Developers can inspect the evaluation code, review the scoring process, and reproduce evaluations when they have access to the required models and APIs.
AI companies and developers can also provide API endpoints for their models, allowing them to be evaluated through the StreamEval framework.
I think benchmarks like this matter as AI video moves beyond short, pre-generated clips toward genuinely interactive experiences.
Explore the leaderboard: streameval.ai
What would matter more to you in an interactive video model: visual quality or following instructions reliably?
#ad #StreamEval #StreamEvalBenchmark #InteractiveVideo #AIRanking
Leaderboard: streameval.ai
Research paper: arxiv.org/pdf/2610.28402
Open-source code: github.com/StreamEval/Stream…
Hugging Face dataset: huggingface.co/datasets/Stre…
ModelScope dataset: modelscope.ai/datasets/Strea…
The next big AI breakthrough might not be another chatbot.
Ling-3.2-flash shows what happens when models can see, reason, interact, and execute across different environments.
From Blender operations to multi-agent collaboration, it's pushing the boundaries of what a single AI model can accomplish.
Introducing Ling-3.2-flash, Ant Ling's native multimodal model: 560B total / 25B active parameters.
Stronger complex visual understanding, long-horizon multimodal agents, 3D understanding & interaction, and medical intelligence.
Outscores Opus 5 (max) in release evaluations:
- OmniDocBench1.5: 92.21 vs 90.56
- ZEROBench(Sub): 43.71 vs 39.52
- Vision2Web: 85.86 vs 80.20
What if you could change the mood of a video without filming a single new scene?
That's what caught my attention about TryDRAMA.
Same footage, same person, but completely different facial expressions. You can switch emotions, adjust their intensity, and give an existing clip a whole new feel.
TryDRAMA × Hugging Face is now live, and it's an interesting tool for creators experimenting with AI video editing.
Try it here:
huggingface.co/spaces/trydra…
@syncmonster_ai #ad
Introducing DRAMA, our AI Expression Editing model.
DRAMA enables precise expression editing in videos and images.
Modify expressions, control the intensity, or blend multiple emotions together, all while preserving the original video quality.
Want to put it to the test? DRAMA is now live on Hugging Face for FREE! 📷
Link: huggingface.co/spaces/trydra…
Platform link: trydrama.io
#trydrama #drama #faceexpressionai #generativeAI #huggingface #aivideo #productlaunch #AIModels #TrendingNow #viral #aivideo #VideoEditor #FYP #Filmmaking #ContentCreator