Ryan Carey
983 posts

Ryan Carey
@oceanb3st
5 years in DTC - $40m spent on Meta for clients Directing a movie: https://t.co/d77EOLNrgx
Global Katılım Nisan 2020
954 Takip Edilen331 Takipçiler

New project delivered for @TheDRAfiliatee 💪
Turns one winning static ad into 100+ Meta-ready variants — new hooks, claims, and angles, rendered in every aspect ratio in under a minute.
Here’s how it works 🧵

English

@andrewxroas ABO's are better for creative testing. More control and you can scale/make decisions faster. Every account I work with spending >$1m/mo uses them
English

@metamantra Brother, a flop? They'll break even on sales in the pre release. This will be the most profitable game of all time.
GTA V generated $1 billion in just three days lmao
English

@SuppsCopyChief How much UGC are you testing per week? We saw a pullback post MD on most brands but also ecom is generally slower June. What's YoY telling you?
English

Currently going through the absolute, without a doubt, worst extended beat down on meta I've experienced in 14 years of doing this.
Anyone who is doing well on meta right now, would love a DM, happy to pay for your time.
Less than 60 days ago we were spending well over $30k a day profitably on meta.
We've run founder ads, creator ads, AI ads, static ads, tested pages, offers, FB pages, ad accounts, pixels, all on new angles, new concepts and old angles we know scaled for us.
Either something is very very wrong, or after 14 years of doing this I just don't know how to make ads anymore.
Feel like I'm taking crazy pills - anyone else going through it this bad?
Truly just a skill issue?
English

@scalesurfer_ Tested on bunch of accounts not seeing any impact still using standard
English

@ZssBecker Lol, no. Obsession wasn't made by AI or five guys sitting in a room pressing buttons on a laptop. It was shot with real physical cameras, real physical actors, real locations, and hard, grueling, traditional filmmaking labor on a micro-budget. GG
English

I shat you not. This will be the biggest money making opportunity in the next 5 years.
In 1-2 years when AI is good enough.
We will see small teams making billion dollar Animes and live action movies at sub million dollar budgets.
The moat in entertainment will be gone and so will the cost/bs that stops creators from creating.
Imagine Game of Thrones. But made by 5 guys and the author, with them all splitting the billions it makes.
The HR-lady-ran cess that are Hollywood and Publishing are breaking. The creators are going to take power back and it's going to make them billionaires.
Good.

English

I've been sitting on a skill that fixes Claude's biggest weakness for 3 months.
Claude is the best model in the world at thinking. It's one of the worst at writing like a human. Every LLM is !
Even Anthropic's own leadership has been open about this : writing with genuine taste and voice is one of the harder things for these models to crack, long after they've conquered code and math.
Default Claude output reads agent written
Em dashes everywhere.
"It's not just X, it's Y." Soft hedges.
Every sentence shaped like every other sentence.
Prompting alone doesn't fix it. You can write a 2,000 word system prompt begging for casual tone and Claude will still betray itself and hallucinate
Prompting matters more than anything that goes without saying but prompting hits a block there
So I built a Claude skill that breaks through it.
We use it internally for everything @Threadslab007 ships.
Every Reddit post. Every comment. Every X post. Every LinkedIn ghostwrite. Every client deliverable. Every cold DM.
This is the single most refined skill in our stack. Nothing else comes close to the leverage it gives us on output quality.
Single SKILL.md file.
Settings → Customize → Skills → Upload. Drop it in
Comment "HUMANIZE" + follow and I'll DM you the file

English

@codyschneider I know some Info offer owners who might shit their pants for this
English

three weeks age I deployed a LinkedIn outbound agent for a startup, and since then it's produced 55 positive responses
this is real GTM engineering
agents autonomously driving pipeline
this is how the agent works
1. The Agent checks 20 linkedin influencers profiles daily for new posts (people engaging with this content their target customers)
2. It then extracts all net new engagers from the posts
3. For each engager it researches them by searching the web and does lead scoring, and if they pass adds it to Heyreach via API
4. Cold DMs the person with the offer
5. Agent manages the inbox via API
6. Meetings get booked
if you want this more below

English

@PhantomStays Even better, staying at the W in Shanghai they have HEATED seats meaning you are also ridiculously cozy. W
English
Ryan Carey retweetledi













