Thumbnail artists ARE NOT cooked.
GPT-4o and Gemini image gen is really cool, but humans who are good at thumbnail design will wield these tools the best.
Left: made by a the talented 4rizon
Right: made by gpt-4o
most ads suck! 😵
But, it doesn’t have to be this way. I can help you and your company make short video ads that actually convert.
I’ve made thousands of ads and have learned through real experience what gets better CTR and lower CAC.
➡️DM to work with my new Short Ads Agency.
If you need help creating and running ads on Facebook, Instagram, TikTok, YouTube--DM me. Or maybe you know someone that owns a company that needs that.
Check out my site to learn more shortads.io
If you're not using video to run paid ads for your business, you're leaving customers and sales on the table.
Video has the highest engagement and best CTR on social media today.
I've created thousands of videos and ads for my own brand and clients.
@mreflow You uncovered something Very profound here—AI can’t generate new, novel ideas, nothing it spits out sparks that creative imagination. It’s an amazing tool at so many parts of the creative process, but it can’t originate, only regurgitate.
Going to ramble for a minute here. Feel free to ignore.
I've been having a real creative block lately when it comes to YouTube. I love making my Friday News breakdown videos where I share everything that happened in the week in a single video. No plans to stop making those.
Outside of those videos, I want to produce more but just haven't had a lot of great ideas. I could do what most other YouTubers (focused on AI) do and rush out a new video every time there's a new announcement. But there's just too many announcements lately and when people rush, they often share inaccurate info. I won't name names but I've seen a lot of videos on YouTube that are rushed out as soon as an announcement happens that share inaccurate information. I want to prioritize accuracy over speed. Besides, I did a poll on YouTube and people seem to prefer all the news bundled into a single weekly video anyway.
I want to differentiate from all the other channels that share all the news. News videos have become a commodity. When a new LLM is released, you'll quickly find 15 videos within the first hour of the announcement. I'm burnt out by that game and trying to race people to be one of the first. Friday news roundups allow me to really process all the news and share accurate info about what's going on.
My channel was originally built on tutorials and news roundups and that's what I want to get back to. Unfortunately, I just have this mental block where I'm so close to everything that's happening that I struggle to determine what tutorials or guides people will actually find valuable. So, instead of making videos during the week, I procrastinate by "vibe coding."
I ask AI for suggestions but it hasn't been great at that still. It's great for helping with title and thumbnail ideas and giving me outlines for videos once I have a concept but it's still not great at giving me concepts that I actually want to make or that I feel would be valuable.
Anyway... Just a little ramble about why I haven't been posting as much lately. Feel free to share feedback or insights. I wanted to share this publicly because I know a lot of creators go through this struggle and it's not unique to you or me. We all have these peaks and troughs. I've been a full-time content creator for 16-years now. I quit my day job back in 2009 and haven't looked back since... And to this day, even with AI, I still haven't figured out how to overcome the creative blocks that hit me from time to time.
I'm always open to advice and ideas though. Sorry for the public rant. I find that putting these types of things into the world helps others see that these problems are pretty universal... And it usually leads to some amazing ideas from people that might read it.
@mrjonfinger THIS. People’s view of what AI video can do right now is greatly screwed by the quick paced 3 second clips back to back with no continuity or relevance to each other. Just trailers. AI will truly shine when you can make a slow paced basic multicam dialogue scene with it.
I think this year Ai trailers need to start showing multiple clips of the same subjects in the same environment from different angles.
The value of isolated clip montages is really hitting saturation.
I wanted to try something a bit longer with Pika additions and swaps. It is still a serious headache, but it was a good challenge just to test the moment.
What on earth is the point of the new Veo 2 model in YouTube Shorts if you can’t generate realistic people? That’s lame…
I get trying to stop deepfakes, but just block names of public figures so we can generate a real looking subject. Spamming didn’t work😔
#veo2@TheoMediaAI
Been debating making a subreddit exclusively for serious YouTubers only…
No hobbyist…
No easy mode
No anti capitalist
No “do it for passion not money”
No whining/complaining
Exclusively people who want to be full time creators or already are…
It was nearly impossible to get such epic full VFX shots and huge wide shots of fantasy worlds years ago—now you can do it in a moment.
#Ray2#DreamMachine#LumaAI
Recreating 15M Scene in 15 mins with AI.
As a filmmaker, I’m always testing the latest tech like cameras, but now it extends much further into the world of AI because how much it’s changing this industry.
I just got my hands on the new Luma AI Ray2 model and it’s impressive!