Filip K

The #1 reason why advertisers struggle with their Meta Ads strategy? They don't test enough creatives. 5 reasons why creative volume is king:

@superderevo i wonder if in the future brand will have as many meta pages as target segments

Steal this from top Meta advertisers Do not run every ad from the same Facebook page Coursiv is a good example They use multiple pages for different use cases: • AI for career • Simple AI explanations • AI for busy moms • AI certificates 1️⃣ This creates cleaner signals for Meta to learn from A person seeing an ad from “AI for busy moms” enters with a different expectation than someone seeing the same offer from a generic page The comments are different The reactions are different The first clicks are different The page also feels more targeted to the audience 2️⃣ It makes the creative system harder to copy Most competitor research starts page by page If all your winners sit under one page, the pattern is easy to map If the system is split across audience-specific pages, competitors see fragments before they see the full strategy The lesson: The Facebook page is not just a wrapper for the ad For top Meta spenders, it’s part of the creative strategy

@superderevo Good news for meta ads mcp wrappers!

Meta just lost months in the AI ads race Yesterday China blocked Meta's $2B acquisition of Manus One sentence from Beijing that rewrites Meta Ads roadmap overnight For most people this is a geopolitics headline. For marketers, the only official MCP connector to Meta Ads just got pulled Manus was already live inside Ads Manager since February It was Meta's fast track to agentic ads. An external AI agent that could autonomously run your Meta campaigns from the outside Now that foundation is gone. Meta has to rebuild from scratch If you're thinking about building agentic workflows for Meta Ads, don't wait for a platform-native solution. Build on top of the Marketing API or use external products Spiral is releasing agents in May. The agentic Meta ads future is still coming. The path just changed

@superderevo easy peasy

reverse-engineer any app's paid growth in 5 steps: 1. open Spiral 2. creatives launched last 30 days: video or image, ios, android or web? 3. which ads are trending this week? 4. what their users say in app store reviews? 5. remix what's already scaling that's 80% of success

@levelsio the problem is trying to sell the AI capabilities as a product, rather than a feature. instead of trying to solve a narrow specific problem, companies are better off using AI to automate the entire workflow in a verticle domain. but seems like everyone just wants to sell shovels

AI UGC is a great example of an idea that was groundbreaking in 2023 when HeyGen pioneered it And then got commodified so fast with open AI models being able to do it That just 3 years later anyone can make an AI UGC generator startup Which means the life cycle of medium AI startups if they don't constantly pivot is shorter than it was with regular startups I think Regular pre-AI startups weren't really based on open models, they were based on open source tech for sure, but there was more custom code tying it together Interesting times

Imagine spending 12/h a day orchestrating agents but still not having clipboard history

@zuckpayer @Meta @finkd hey man, sorry to hear this. something that has worked for me in the past: create an instagram account, pay for premium and that will get you customer support access. explain to them that you got unfairly banned and what you were doing i managed to get unbanned twice doing this

🚨 @Meta @finkd JUST PERMANENTLY BANNED MY ACCOUNT. 16+ years on Facebook. Over $1,500,000 spent in ads across my account. And what do I get back? An automated email. "Your review was unsuccessful." Why? Because I connected Claude + MCP to my system to get more out of Meta's platform. TO SPEND MORE TO THEM! That's it. That's the whole reason. No explanation. No human on the line. Just some underpaid kid behind a laptop making copy-paste decisions about accounts spending thousands of dollars per day. Meta has completely lost sight of who their customers are. And they simply don't care. It's time for a new platform to rise. The Meta monopoly needs to end. We deserve better. #CancelMeta #FacebookBan #MetaMonopoly #DigitalRights #AntiMeta #BigTechAbuse #MetaFail #ClaudeAI #AIRights #BreakUpMeta

A single view of your competitor's creative tests makes it stupidly easy to understand which ads are working for them. Recent green are in testing. Older green are winners. Done.

@jackfriks this is cool

i have a folder on my computer named "keep going" here are some of the images inside of it: (part 7)

@codyschneider and apparently the docs are not enough to not go against meta tos

let me get this straight so what you're telling me is an MCP is an API wrapper with the API docs saved locally so the AI model knows how to use the API without searching on google via exa AI for the API docs got it

@gregisenberg would love a pod with someone who is using this consistently in business

i found a github repo that lets you spin up an ai agency with ai employees engineers, designers, growth marketers, product managers each role runs as its own agent and they coordinate to ship ideas 10k+ stars in under 7 days 1. engineering (7 agents) frontend, backend, mobile, ai, devops, prototyping, senior development 2. design (7) ui/ux, research, architecture, branding, visual storytelling, image generation 3. marketing (8) growth hacking, content, twitter, tiktok, instagram, reddit, app store 4. product (3) sprint prioritization, trend research, feedback synthesis 5. project management (5) production, coordination, operations, experimentation 6. testing (7) qa, performance analysis, api testing, quality verification 7. support (6) customer service, analytics, finance, legal, executive reporting 8. spatial computing (6) xr, visionos, webxr, metal, vision pro 9. specialized (6) multi agent orchestration, data analytics, sales, distribution what i like about this approach is the framing instead of one big ai agent trying to do everything, you structure it more like a company. specialized agents, clear responsibilities, workflows between them im curious to see what this actually feels like in practice and if its any good (do your own research) github.com/msitarzewski/a… but as always will share what i learn in public and on @startupideaspod one thing is for certain and it reminds me the future belongs to those who tinker with software like this

Claude code taking over parallel industries is a good insight into what will happen if AI-native devs continue getting laid off

It would be interesting to hear you talk about freelancing in Europe. I know of devs that were able to quite easily get to 250k€/y freelancing in Europe, especially since AI started increasing productivity in development. It just seems quite obvious to me that getting 5 clients paying 5k/month is much easier than finding a job that pays 25k/month. Would love to hear your thoughts

I’ve met so many software engineers in Europe who assumed the same (“no way anyone in EU can make more than €100K/yr) Then read my article on the Trimodal nature of dev salaries, realized which companies to apply to, and broke in there. This article: blog.pragmaticengineer.com/software-engin…

@williamkast_ you can get away with AI for statics because at this point they are indistinguishable from real human edits. but for videos this 100% true

Never try to imitate real ads with AI. People can still often tell, and you make them feel as if you're lying to them. AI ads perform best when you don't try to hide that they're AI. If you go for AI, make it unrealistic on purpose. Now you turned "lying to your customers" into "hooking them with something new and entertaining.

@andrewxroas yup, it's the ozempic crowd

The popularity of peptides is caused by the fact that most people are lazy mfers always looking for shortcuts in life It’s the same exact type of people who tried getting rich off crypto and trading instead of learning skills and monetizing them

the issue with having this fully automated is that it's hard for the AI to figure out how to present the data without full context about what you are looking for. and if it has full context then why bother presenting the data and not just give you the insight? but for sure dynamic UIs will be a thing, its just that they we will still probably have to guide the AI on how to present the data to get meaningful insights

im a firm believer that on demand UIs are the future of data analysis I just had claude code make this paid ads dashboard using live data from my data warehouse in about 30 seconds first I built a data pipeline and a data warehouse to unify the data then I connected to this via the Graphed MCP which wrote the SQL for every chart and returns a live URL so this data is constantly updating in the dashboard then I wanted to add date filtering and to make it so I could bucket by daily, weekly, monthly so I prompted it to add that to the dashboard then I wanted it to be mobile friendly so I could take the screenshot that you're seeing in post and it did it on demand UIs custom to your exact needs at that exact moment in time are how we're going to visualize data and turn it into decisions in 2026

If you think big companies have their shit together try integrating with meta's api

@shl and with AI you can speed up the process by building many bad things at the same time

You must build bad things before you can have good ideas

@adamdotdev @ThePrimeagen there's huge upside to not going as fast as possible for the sake of being able to understand your codebase. maximizing speed while minimizing the devs distance to the code is the key

I find it very funny when anyone feels confident that they've figured out agentic programming, even funnier when they're trying to teach others how to do it. I've been working on OpenCode since May of last year and I still have days (like yesterday) where I'm not even sure any of this is a good idea lol I end up landing on "yes, these models are an incredible tool" but it's still all very confusing, lots of tangled thoughts and emotions and realities. I badly miss the mundane coding tasks that broke up my days/weeks, the ones where you put on the headphones and just bang out 600 lines of code. But, no question, replacing those hours of my time with a few minutes of waiting on an agent is a boost and worth being excited about, despite the mixed emotions. Then there's the distance that can creep in between you and the codebase if you start getting apathetic. I think it's pretty common at this point to make even small changes by prompting the models. It's less friction than finding the relevant code and making the change yourself. And less friction seems to win, must be some law of the universe or some shit. When most or all of your interactions with a codebase start flowing through the models, you start to lose track of where things live, which abstractions/components are carrying the weight, etc. It's a scary feeling to wake up and realizing you can't even reliably @ a precise file for a change you want to make, and you have to get more vague, leaning harder on the model. It all creeps up on you, there's an undeniable dopamine hit from using these things, and the resulting come down is predictable, like coming off a sugar high. On the positive side, it's really nice seeing other devs go through the same cycles, knowing we're all in this together and we'll ultimately figure it out.

Let your agents build the software while you build the office