Spacesheepboy

[Spacesheepboy @Spacesheep20441]

@laaaarissaaaa Daily dose of life reality

[Spacesheepboy @Spacesheep20441]

That’s the first app I made. It’s called “Dying soon”. It sends you daily notifications to remind you of the fact that nothing is permanent, and each of us may die one day or another. Not to scare people but to encourage us to be present and cherish each moment that we have .

[Spacesheepboy @Spacesheep20441]

Hi everyone, this is my first post. I’ve decided this year to step out of my comfort zone. Posting publicly on social media wasn’t an easy thing for someone introverted like me, but you gotta to try new things out :)

I love vibe coding and building small (and silly apps).

[Max 🇮🇪🇱🇻🇨🇾 @maks6361]

After automating my TikTok posting pipeline, I got engaged and went for Instagram posts automation 😀

That was actually easy as well but the best thing is that it's full 100% automated!

I have my mini pc that runs 24/7 firing cron jobs for 2 pipelines:
1. Slideshows into the 12-15 secs video.
2. Video content : 4 secs of the hook (AI generated with Veo 3.1) and then 6 secs of the app CTA glued together.

Started this pipeline only 3 days ago so a bit early to see the outcome.

I get around 50 views per both slideshow play-through and the video content but hopefully it'll grow later!

[Andrej Karpathy @karpathy]

Don't think of LLMs as entities but as simulators. For example, when exploring a topic, don't ask:

"What do you think about xyz"?

There is no "you". Next time try:

"What would be a good group of people to explore xyz? What would they say?"

The LLM can channel/simulate many perspectives but it hasn't "thought about" xyz for a while and over time and formed its own opinions in the way we're used to. If you force it via the use of "you", it will give you something by adopting a personality embedding vector implied by the statistics of its finetuning data and then simulate that. It's fine to do, but there is a lot less mystique to it than I find people naively attribute to "asking an AI".

[Cody Plofker @codyplof]

I've been thinking a lot about how ad creative is evolving rapidly. If you're active here, you've probably seen 100 posts about Andromeda and creative similarity. And for good reason.

Things are changing and changing quick. Here are a bunch of things I currently believe about ad creative right now and going into 2026:

1 . Variation and iteration testing is a big cost center and the majority of it is a waste of time. Meta is being very clear now about what it considers a unique ad. I was personally a bit shocked to learn that ads that I thought looked very different and communicated different things were considered the same ad/ entity ID in Meta. So if even seemingly larger changes are considered the same ad; all of those small variations like hooks, copy, etc seem mostly pointless.

We've gone from 4-5 variations of each ad to 2-3, and often we make no variations. And when we do iterate now, we only find success on a pretty significant swing. Usually it entails taking the same messaging and finding new ways to say and show it.

2. The editor era is over. The creator era is here. It used to be that the most valuable person on a performance creative team as a great video editor. A great editor could go into a library; find a bunch of footage, and make a bunch of banger ads from it. There is still a bit of value in this approach,but I think much less so. Today the value is in diversified content, and you can only get so far with the same footage.

You need volume, diversity, and creativty and that is what creators get you. People are engaging with creators way more than brands on social, and you can only get so much out of editing. We need net new, social first content across a wide variety of personas, demographics, and messaging and that is a near impossible thing to do with just brand led footage or production. Creators are the solution here.

3. Speaking of creators, you know I am a huge fan of partnership ads. As much as people are talking about them and testing them, its still underrated. But they are only one type of creator asset; I do think you can work with creators for non-partnership ads and should have a creator strategy from your brand page. But I think almost everybody should be investing in partnership ads.

4. DR UGC is dying. I think the super scripted, cheesy direct response ads that a lot of the big name agencies produce is past it's heyday and that is a good thing. Consumers are numb to it, it is doing nothing for brand consideration, and it's no longer as effective. UGC started because it was super effective; but somehow it's become super produced, with strict scripts, still edits, round of edits by comittie, and endless iterations. We've lost the plot.

5. I think a bunch of teams across organic and paid social are behind. They're stil trying to curate pretty IG feeds for followers. That era is over. TikTok is everywhere, and everywhere is TikTok. What I mean by that is every platform has a short form, algorithmically generated feed where followers don't matter. Engagement is all that matters, and if people engage with content it will be shown to more people like them.

This is true across organic and paid. I think we will continue to see less difference between good organic and paid content. Both need to be top of funnel, hyper personalized, and give users a reason to share it with others.

6. More teams should have creators on staff. I think it would be a significantly better investment to have 2 full time creators on staff pumping out TikToks/Reels/Ads for 40 hours a week than most UGC agencies. I am testing this right now in some smaller capacities, but I'm likely going to do this.

7. More brands should hire creators as creative strategists. Ridge does this and it's really smart. I think it's easier to teach someone selling and data than it is so teach an ad person good content. I also want to do this. I think it would be a good idea to hire a creator in that audience cohort as the strategist for it.

8. I think we'll see more success with more produced content. I think the typical DR UGC agency is in the middle; too produced/edited to be authentic but too janky to invoke trust. You either want to go high and produce your stuff; or go low and make it super raw and unedited.

BTW, I'm on the hunt for a good video agency who can shoot on DSLRs or fancier but also gets performance. If you know of any, hit me up.

[Charly Wargnier ♨️ @DataChaz]

This guy literally built a system prompt that makes Sora 2 create mind-blowing videos.

(prompt in 🧵 ↓)

[Jordan Menard @jordanmenard]

"Natural CAC" = the cost Meta wants to charge you to acquire a customer.

Hit this number, and Meta will give you nearly infinite scale.

Raise budget 100% → CAC only rises ~10%. For subscription brands, the game is finding this number; then scaling into it.

Trying to push it lower is a fool’s errand. Meta’s algo knows the price.

[Sam Mendelsohn @SamMendelsohnW6]

I have a theory about Facebook ads that sounds like a conspiracy but might actually be true.

One of my buddies built a tool that gave advertisers perfect event matching. Like 100% match rates. Every conversion tracked flawlessly.

Every brand that turned it on saw an immediate lift.

But 2-3 months later, that lift disappeared. Every single time. Performance came back down to exactly where it was before.

So they experimented. Turn it on for a month, off for two months, back on again.

Same pattern every time. Temporary bump, then reversion to baseline.

His theory: Meta knows your margin profile and will squeeze you right up to the edge of what you can afford to keep spending.

Think about it. Meta has 10+ years of data on who actually buys from ads and who never will. I'm on Facebook every day and I've probably never bought anything from a Facebook ad. They know that about me.

But they still need to monetize those users. So they show them ads anyway, diluting your high-intent audience with enough low-intent users to keep you hooked, but not quite profitable enough to scale infinitely.

Any efficiency gain you get just gives Meta room to push you into slightly worse audiences until you're back at your margin ceiling.

It doesn't matter what you do. Meta will always bring you back to whatever performance you started with.

I don't know if this is actually true. Maybe it's seasonality or survivor bias or a hundred other explanations. But everytime I hear a pitch for the next great data enrichment tool this is the first thing I think about.

Part I'm not sure of is whether brands actually have any control here, or if we're all just playing a game where the house always adjusts the odds.

If Meta's always going to revert performance to equilibrium, the only real leverage we have is outside the ad platform.

Better product-market fit.

Healthier margins.

Higher retention.

Operational excellence.

The stuff that actually lets you afford to spend more than the next guy.

The bad news is those things are really hard. The good news is they're a much more defensible moat than how you structure your ad account.

[Ernesto Lopez @ErnestoSOFTWARE]

This is actually crazy🤯

This Ig account is generating over
100M views a month using Sora

In the gym niche.

Step 1: make a gym app using cursor
Step 2: make 100 viral Sora videos
Step 3: post them with a CTA to ur app
Step 4: print money.

[Lotanna Ezeike 💳 @lottsnomad]

i think i’ve just seen a masterclass in ugc content

- the shock
- the subtle placement
- the ragebait (to the exact audience)

1.5m likes
4.6k comments
119k bookmarks
146k shares

[Spacesheepboy @Spacesheep20441]

@past_funny2046 125买的 要割吗🥲