I met with @VadimNotJustDev about the playbook behind Postiz, it's a long hour full of growth hacks. Here is one I reverse engineered by watching Higgsfield. With this hack I am guaranteed to have at least 100k views per post.

This is what I tried to understand, too. Because you can get a JSON response from all the big classification models, it's an easy task. Jev is 100x faster / cheaper. But nobody is talking about the quality. I don't write a single line of AI in my tweets, yet all those Jev classifiers tagged them as AI. We need someone to compare smart models' (Fable / Astra) classification with Jev. I am sure Jev is great for dumb classification, but maybe even a simple machine learning model can be good.

Two days ago, I talked with @willbakerships about B2B SaaS, and we both agreed we won't see many companies "making it" past $10k MRR. But the ones that do will become giants. SaaS added a big filter, as it did before COVID. Be innovative, spot trends, do hard marketing.

Here is the thing about Muse that everybody should understand. So far, we had @bot from X (tech Twitter); that's a big, sophisticated market. Now, the boomer market (Facebook, Instagram) @Muse will start to get it. You have to understand that the boomer market is much bigger than X, as it's no longer "early adopters" They are also making buying decisions faster. Play it right, be early, and you can make loads of money.

I grew 3k followers since yesterday. But that doesn't mean it's good. When you are making money and post about it, you get a lot of engagement, even more if you have a big milestone like "200k" MRR. I plan for it to go viral. By the end of the post, I add a CTA; in that one, I added >>> follow me, with a good explanation of why. CTA is important; it's how people take action, so if you ever have content that can go viral, put a CTA at the end. But what did I see from the CTA a few posts after? My reach didn't grow, and a lot of unverified members followed me. So, is getting a lot of followers always good in 2026? Not necessarily. I think you should have people follow you because they are genuinely interested and were never prompted to do so; those people are much more likely to engage with you. Compared to "dead followers" who see your posts and don't engage, which lowers your reach. BTW, welcome all the new followers ❤️

↳ SaaS is Dead. ↳ You are one step away from being replaced. ↳ I can vibe-code Postiz in one day ↳ If I were you, I would sell Thank you for your attention to this matter.

You should use the new MCP UI option in ChatGPT and Claude not because it looks good, but because it can provide missing logic. For example: - You can implement an upload component, which is super useful because MCP can't upload files. - You can implement a poller - usually the client side of ChatGPT and claude, won't pull for a video status if you generate something; a UI widget can solve it. Thank me later :)

What does it mean to become rich? And not this 50% Europe tax, 20% profit from revenue. No. Offshore 0% corporate tax in Hong Kong (PTR), and taxed on the money you remit to Thailand. And a company that has an infrastructure of $3,000 with a dynamic marketing spend (up to me) Pros (feels good): - Never care about it - Your wife doesn't ask you by the end of every month how much money we made, and you say, same as last month because we invested it back in the business or had some extra trips. - I can pay for mounjerou (it's pretty expensive every month); in general, I want to do a hair transplant and other body modifications. - I know I will never have money fights with my wife (unless we divorce) - I can pay for my family to visit, and even fly my friends! - I can rent a better apartment - I can't fly business; I've never done it - My daughter can live her best life (we will try to make her less spoiled, though) Cons: - Can't go back to my country, to my friends and family because of the "honeypot" situation. - Operating on a larger scale is much more stressful - Afraid to lose it all because of AGI - High churn - stressed that if customers stop coming, I will die fast. - Hard to find time for more than business; it's sucking you in! All in all, very happy!

Postiz reached $200k MRR!!!!!!!! Actually, it was crazy because on the day I was supposed to hit 200k, I grew $3k (so it's $202k now) 😅 I am planning to do a big video about everything I have done to get there, plus my plans to reach $3m ARR. Follow me an put the bell icon so you won't miss it 😉 I promise you some crazy-ass marketing tactics.

There are a few arbitrage opportunities today in the market that I think people are sleeping on for marketing. 1. B2B games - not sure if you know @threejs or not, but since their graphics engine is code-only, a few years ago nobody would even think of dealing with it. Today, together with Meshyai, you can create awesome marketing assets for your SaaS. The last bidding trend was just an example. 2. Seedance 2.5 - Everybody is still doing screen recording/video motion graphics. You can create amazing stories with AI, with truly cinematic results. Almost nobody is doing it, and the people who do usually grab a lot of attention. 3. UGCs with OMNI - there is a big trend specifically on @sideshift_app to create videos with many, many content creators, but you can easily grab more attention by modifying your first shot; for example, add crooked teeth or a huge pimple. It always feels like everybody is doing the same thing without any innovation. And the ones that do grab all the attention. Soon, you'll see amazing new marketing stuff from Postiz.

So a little after I published it, somebody from @OpenAIDevs checked my plugin and declined it :( I don't know why; there was some weird bug between organizations in OpenAI. I resubmitted it, if that nice annonymous person wants to help me again, that would be awesome ❤️

I think we are living in a crazy time, especially for established companies. When I started Postiz almost 2 years ago, I didn't want to handle video and photo processing because working with videos is always a headache. So I used Transloadit; it's an FFMPEG service for handling files. The service is great, but it isn't scaling well; it's getting very expensive, and for me, since I have many people scheduling posts from MCP and CLI, you can imagine the insane number of calls I get with videos. So I didn't offer Transloadit for the API/CLI/MCP users because I definitely couldn't absorb the cost. Today I said enough. I used Claude Fable 5.1 to build a system, then deployed it on RunPod because I need a GPU. Everything is in Python; I don't even know Python 😂 It took me the whole morning, literally. We are living in an insane time. Here is the code: