Founder @ Miro, an ai data lab. Bootstrapped a business to 8 figs @ JUICE (acquired). Fintech partner @ https://nitter.gq/t.co/fk0Vwy0EaK. First (web hosting) co @ 14
Miami, FL
Joined May 2012
- Tweets4.8K
- Following1.3K
- Followers9.4K
- Likes20.8K
Pinned Tweet
Psychology is used to persuade you every day.
Here are 13 cognitive biases that marketers use.
(And how you can use them to influence others)
-Thread-
Most data vendors pitching frontier labs have their deal killed before it even gets to eng/researchers.
Size matters… but not as much as some other things.
What labs actually care about:
- Token count is a vanity metric. If signal density is zero, nobody cares about your billion tokens.
- Cheap synthetic text is poison. Unverified synthetic data ruins reasoning trajectories, and researchers treat it like slop.
- The real money is in verifiable environments. Compilers, CAD, games, deterministic setups where models test actions against ground truth.
Don’t get fooled by the tourists flexing data deals on X…
It took one @alexandr_wang tweet and all of the labs suddenly started buying video games to train on
The coolest thing about the AI data space is it's kind of like having a crystal ball into what AI will EVEN better at in 3-6 months
Right now that's:
- Construction/Architecture/CAD
- Generated 3D environments
- Mechanical engineering
- Enterprise workflows
Unpopular opinion: all of the crypto bros are going to leave the data space in ~1 month.
I’ve spoken to a lot of people who see the tweets, the hype etc but as soon as you ask them how many deals they’ve closed or how much they’ve been paid, it ends up being 0.
1 month will be roughly enough time for people to realize the “easy” part was a lie/marketing by a few very smart companies (h/t @micro1_ai)
Data brokers showing off ARR are by definition schemers
All of the revenue is one off. Multiplying it by anything is nonsense.
The ones that sell to labs are taking a license and building a war chest of assets, but anyone claiming ARR is just disingenuous
I spent a week in the Bay Area to learn everything I could about the AI data industry.
Here are the top 5 most shocking things I learned:
The easy money “skool” version of data brokering is going to die as quickly as it came
AI labs have gotten really strict really fast… the money is flowing but be prepared for the grifters to be shaken out
Much better for the space
Artec is coming for the creator space 👁️
Introducing Artec 2.0: the Creator Studio
Artec is the ultimate all in one UGC app, for creators.
Gain access to over 5,000 brand deals, today, from all of your favorite marketplaces and brands.
Create incredible videos and save hours of time with tools such as Scripter, AI Editor, Analyzer, and more.
And finally, manage your content career, all in one place.
To gain access, comment “Artec” and we will send you the link in dm.
Create more, and work less, with @artecapp :)
Imagine “making $75k per deal” but deciding to sell a course on it instead
🤔
We have everything you need to succeed with the offer.
Join Data Broker, 60+ members who are running this Gold Rush Offer:
skool.com/databroker
Data brokering will be like any other industry where there’s fast money.
The grifters will be quickly expelled because it’s really obvious if you have no idea what you’re doing.
The engagement bait on X will fade. The people who do well by labs and sellers will make a fortune.
and if you want a reliable data partner, consider Miro (and the highest offers as well):
miroadvisory.com
