Erick F. (@OneShot_Erick) header
Erick F. (@OneShot_Erick) profile picture

I built the software that helps instagram creators to increase their reach 馃殌

Oporto, Portugal
Joined May 2011
Photo by Erick F. (@OneShot_Erick)Photo by Erick F. (@OneShot_Erick)Photo by Erick F. (@OneShot_Erick)Photo by Erick F. (@OneShot_Erick)Photo by Erick F. (@OneShot_Erick)Photo by Erick F. (@OneShot_Erick)Photo by Erick F. (@OneShot_Erick)Photo by Erick F. (@OneShot_Erick)Photo by Erick F. (@OneShot_Erick)Photo by Erick F. (@OneShot_Erick)
vibecoders, fill this in for me: the thing that finally made me ship was ___ 馃憞
33
Still the best advice I'd give someone starting out alone: build the thing you'd use even if nobody else ever did. That's how this started for me. I needed it at the agency. Everyone who showed up after that was a bonus I never planned for.
40
Why do so many founders ask AI what to build next when their users already told them?
45
I used to treat every post on a new account like it had to be perfect. Now I think the first few weeks are for experimenting. The polished grid can come later. A new account has no audience to lose. Nobody is studying the grid yet. That makes it the best time to be a little messy on purpose. So take one idea you actually believe in and make it a few different ways. A different first second. A different caption style. A different sound. Spread them across the week and give each one a fair shot. Then read what the numbers are telling you. Views are only part of it. Look at where people drop off, which version got shared, and which one reached people who had never seen you before. Reaching non followers is the part that grows an account. When something works, don't upload the exact same file again. Make new versions of it and keep testing. That's how one winner slowly turns into a direction. The method is simple. Sticking with it is the hard bit, because the first weeks usually look like nothing is happening, and that's exactly when a lot of people stop. If you're growing something from zero right now, what are you testing this week?
1
70
be honest: when a post flops, do you make a new version of it, or move straight on to the next idea? 馃憖
42
One edit gets you a video. A batch gets you that same video in a whole wardrobe of outfits 馃帹 Bold white captions, small lowercase ones, a warmer grade, a different track. On Trial Reels each one gets shown to a different crowd of people who don't follow you yet.
1
35
That's what VideoBatcher is for. Pools of hooks, captions and music mixed across your videos automatically, with color, zoom and framing changes on top, rendered on your own computer. videobatcher.com
7
Some nights at the agency were nothing but exporting. Version after version of the same video, late into the night. I didn't hate the work. I hated that none of it could be organized or automated.
1
47
That feeling is still why I build. Every time I add something, I ask the same quiet question: would this have saved that version of me a night?
7
Creators who grew an account from zero: what did the first month actually look like for you? 馃憖
62
doing support, product and marketing alone for something people pay for every month is a strange mix. one hour you're answering a question about a render, the next you're deciding what to build next. wouldn't trade it. ask me again at 2am though 馃槄
44
Your best hook is probably one you haven't tested yet.
47
A lot of people vibecoding right now are building for problems they only read about. I built for one I had every single week at an agency: hundreds of versions of the same video, all made by hand.
1
35
AI made building cheap for everyone. It didn't give anyone else your problem. That part is still yours. If you're stuck on what to build, start with whatever annoyed you most this month.
3
Builders: ship the rough version this week, or the clean version next month? 馃憞 No wrong answer. Curious which one you actually do.
41
The part of VideoBatcher I like most has nothing to do with editing. It's that you can ask Claude or GPT for 30 hooks, and they go straight into a batch.
1
27
Through the MCP, the AI sends those hooks over with caption styles and music, and your own computer renders every version. No pasting 30 lines into 30 exports by hand.
1
4
Honestly, most of those hooks won't win. You don't need them to. You need the few that do, and now testing them costs you one sentence instead of an afternoon. videobatcher.com
3
Next time a video flops, try this before giving up on the idea 馃挕 Keep the footage. Change only the first second and the caption, and post it as a new version a few days later. Sometimes the idea was fine and the opening wasn't.
40
Founders building something right now, what is it? Drop the link and one line on who it's for 馃憞
5
3
160