Social media is changing… FAST. There is ONE thing to focus on above all that will help you stand out in 2026.
…the same thing it’s always been:
THE STORY.
But perhaps, it’s more important now that it’s ever been.
AI can generate infinite content, but it can’t generate why it matters.
And that part is always determined by the human element.
You want to stand out in the era of the infinite content machine?
Learn how to create meaning, not just things.
STORY STORY STORY.
Create to serve. I think this message is more relevant now than ever.
Which is why I’m re-uploading one of my most popular videos (with a slightly different intro & sound edit)
Do you want to build an audience? Ask yourself — what’s in it for them? How can you best serve them?
Is it education? Inspiration? Entertainment?
When you think about it, success is always rooted in service.
When you’re building a startup - any product or service - you’re first identifying what problem you are solving for others.
It’s the same with The Great Online Game.
Creating to serve helps combat burnout too:
When you’re in it for yourself - a certain number or metric - you’re more susceptible to burnout.
When you’re in service to others, there’s a certain strength that comes to you.
&
Fall in love with the journey. You can’t lose.
This tool feels like Claude Cowork for creative teams. It’s called Skills, by @air.hq
Air is your headquarters for all creative operations.
Think of it like Dropbox + Frame io + Claude all-in-one.
Here’s what they figured out with Skills.
Right now, every time you make an AI edit, it’s basically a one-off.
There’s no real way to save it, or share it to re-use later.
It’s just a single prompt at a time.
Skills changes that.’
Once you crack a winning workflow, you can save it as a skill for anyone on your team to be able to reuse automatically and endlessly across all of your different assets at once.
And you have every cutting-edge model available to use, to build skills with.
Which is why Air is so popular — it’s used by Cursor, Warby Parker, Ramp and 3,000 other brands.
If you want to try Air, comment “SKILLS” and I’ll send you a link.
I got a chance to visit the NEO factory, just unveiled by @1x.technologies — the first true vertically integrated robot factory in the United States.
The full factory tour is coming soon — the first episode from an original series by @metav3rse 🦾
Freepik just became Magnific.
• More than 1 million paid subscribers.
• More than 100 million monthly visits.
• Over 175 million images and videos generated monthly.
This is the story of how a small stock photo company founded in Málaga, Spain in 2010, became a powerhouse with $230 million in annual recurring revenue.
Note: I have a long-standing partnership with the team at Freepik… err Magnific, but this video was made out of love for my friends over in Málaga
LTX-2.3 is without a doubt the best open-source audio-video model available.
It has been downloaded 1.65 million times –– in the last month alone.
It’s the first, highest benchmarking, and most downloaded truly open-source audio-video generation model, and is designed to run locally on consumer GPUs.
If you’re a creator who no longer wants to pay AI video subscription, a developer building real tools, or a tinkerer who wants to experiment with running AI models locally on your machine, this is obvious and best choice.
Comment “LTX” for a link to the model, and resources on how to install it.
@ltx_model
You can now use Claude Code (with the new computer use feature) or Claude Cowork to automate entire ad creation end-to-end.
Meaning, Claude actually takes over your computer and browser to automate what you normally would have to do manually.
So your imagination is the limit. You can get your work done while you sleep.
The awesome part about this workflow is that you can customize it to produce any type of creative you want –– not just UGC.
I used @higgsfield.ai for it because they currently have the cheapest access to unlimited Seedance 2.0 (up to 70% off)
Comment “AI” and I’ll send you the guide I put together on how to create and run this workflow.
#HiggsfieldCreator
Seedance 2.0 is the video model that shook the film industry.
It is the world’s highest benchmarking video model by a wide margin.
It is finally available globally except for – unfortunately – the US (which is coming very soon)
You can try it now on @higgsfield.ai
Here’s what the model does very well:
Director-level control: precision over performance, lighting, shadow and camera movement.
Up-to 15 second videos: Currently only available at 720p
Multi-camera storytelling: Generate a storyline with natural shot connections
Temporal consistency: My European team experimented with this model for an entire day and this was what was most impressive to them.
In a way, Seedance 2.0 is a cinematography simulator.
Even for anti-AI filmmakers, I think it can be helpful for immersive storyboarding.
Comment “AI” and I’ll send you the link to try it.
#HiggsfieldCreator
Anthropic just killed it – again. Claude can now control your entire computer autonomously.
It can open and use any app, browser, file, folder, spreadsheet – anything you’d do sitting at your desk.
But the real magic, is when this is paired with another recently released feature by Anthropic called Claude Dispatch.
A single conversation thread on your phone orchestrates multiple AI agents completing tasks, using your computer, in parallel, while you’re away from your machine.
In this video I show you why this is so important, and how to set it up.
Now a few notes from me:
➝ This is currently only available on Mac and to Pro and Max subscribers.
➝ Don’t expect a miraculous experience that will execute everything flawlessly. We’re early – this is being positioned as a research preview for a reason.
➝ You’re giving access to your computer and files to Anthropic. That capability comes with risks worth understanding.
If you like this video, follow @rpn or comment “signal” for my newsletter to stay two steps ahead on the incredibly disruption happening in tech and how to leverage it in content and business.
Perplexity just reinvented the computer, with Computer.
Think of it like OpenClaw without the security vulnerabilities. To me this is the closest thing we have to Jarvis for the average consumer.
The magic of Computer is the multi-model orchestration. Meaning, when you give it a high-level task, this beast breaks it down into multiple sub-tasks to run in parallel, tapping into 19+ frontier models to automatically assign the right one to each sub-task.
Combined with thousands of app connections, @perplexity Computer is capable of end-to-end execution across the entire context of your work.
Comment “Computer” and I’ll send you a link to try it.
Everybody’s talking about Claude Code but I’m surprised at how little people are using Claude Skills – which is honestly the feature that makes the whole thing worth using.
Instead of re-explaining how you work every time you open a new session, you write it down once in a simple markdown file and Claude automatically understands your preferred structure, format, standard etc for the output.
This video breaks down what Skills are, how to create your first one in a few minutes, and how to start building a system that actually works the way your brain works.
Save this for later.
Quick note: Skills also exist on regular Claude, but those are pre-built — things like creating docs and spreadsheets. Claude Code is where you build your own custom Skills tailored to how you actually work.
So regular Claude gives you Skills that are essentially like system prompts at the project level. Claude Code is where you can really get their full value.
Second note: I recommend having between five and 10 clearly defined skills that don’t overlap in their purpose. The more skills you have (especially overlapping ones) — the less reliable they become.
