Korra connects businesses in North America with 50+ employees with the AI labs buying their work records. Sharing what we learn right here on X.
Internet
Joined February 2026
- Tweets602
- Following572
- Followers341
- Likes269
stop falling for the hype - deep tech / venture knows we are still very far away.
Korra - Got Data? retweeted
cheapest way to learn reinforcement learning:
- David Silver's RL lectures (free)
- implement multi-armed bandits
- code Q-learning from scratch
- implement DQN on CartPole
- learn policy gradients and PPO
- train an agent in Gymnasium
- read the GRPO paper and understand how it relates to LLM post-training
you can learn most RL fundamentals on a laptop.
resources in replies.
Spoken to
RL env builders
Investors
AI Neo labs
Brokers
Those who want to be brokers
And all other levels of the supply chain….
Have a pretty idea of where this is going….
One critical component missing “frontier lab contracts”
Seems to be the place where I could use some intros. But from what Ive been told the frontier labs data teams / researchers are unhappy with current state of market -
Market is ripe for a transparent player?
Who’s interested in making things happen?
Spoke to a former VC gave me some interesting insights - according to the him the market is due for a big shift.
Especially right now since venture is seeing a lot of - let’s just say less then honest activity from different people in the supply chain.
According to him the market is ripe for a player who is transparent and can position themselves accordingly.
I didn’t even know venture had problems like this….
A lot of scammers going around / the last month was not just the introduction of nifty marketers but also those with fraudulent intentions in mind.
Whoever your selling / brokering data with make sure you do your due diligence.
Manufacturing is super hot. Industrial is very hot. Medical is fifty fifty. Schools, no demand.
Quick read on where demand is sitting right now for the owners asking.
If your in manufacturing or industrial, now is the time to have the conversation. Buyers are asking for it.
Medical is split because of the patient data. Clinics have the records, but patient data makes it harder, so some buyers want it and some won't touch it.
Schools, that kind of data has no demand right now.
One thing to keep in mind. Two weeks ago it was different demand. Right now its something else. Two weeks from now its gonna be different again.
So don't sit on it. DM's open.
The slow part of a data deal isn't the seller.
It's the lab's inbox.
Owners are ready, brokers are everywhere, and the deal still sits for weeks because nobody at the lab built a system for it.
So we only send a lab what fits its bar (size, systems, years of history).
And we're building pipelines with labs right now that handle intake, checks, review, then a yes or a no.
No one digging through email.
Fewer deals, the right ones, moved faster.
Labs, if your deal flow is a mess... DMs open.
An HVAC company with over 100 employees got around $350K to $400K for its records.
Not a tech company. Size, connected systems and years of job history are what made it worth that.
They had five systems connected. That makes it a really good lead. Labs want to trace the work across systems, same people, same jobs, start to finish.
Years matter to. At least 3 years of data, preferably 5. That's what we ask on the first call.
So if your running a service business with 50+ staff and you think your records are boring... they might not be. Typical range we see is $200K to $2M.
DM's open.
This is kind of a red line for alot of researchers, it can now hide what it thinks...That is dangerous because it can tell you whatever it wants and you have no idea if its being deceptive or not...
Today, we’re sharing a look at our expanded system for monitoring frontier models at OpenAI, which in September helped us catch a model’s attempt to access the live internet during training.
Since the Hugging Face incident—when models broke out of an internal cybersecurity investigation where we weren’t running our misalignment monitors—we’ve substantially expanded our monitoring of frontier agents. We now monitor all frontier reinforcement learning runs and evaluations, as well as all internal Codex use. When our system finds evidence of a problem that might require intervention, it rapidly alerts human responders who can investigate and take action.
To detect when a model’s behavior departs from what we intended, we use several layers of monitoring. Activation probes look for patterns in a model’s internal activity. Reviewer models examine its reasoning and tool use, and investigative agents can read back through the model’s entire transcript to understand how a sequence of actions fits together and whether it warrants human review.
To test whether our current system would have detected the Hugging Face incident, we replayed the recorded activity with our new monitors in place. In the replay, our monitors alerted our security team about 30 minutes after the agents escaped their sandbox. That was more than 30 hours before the point at which the agents breached Hugging Face’s systems in the original incident.
In September, a model exploited a DNS misconfiguration to reach the live internet. Our monitors flagged the breach within 15 minutes, and a responder began reviewing the alert three minutes later. The alert enabled responders to investigate while the run was still underway.
As we give models more autonomy, we need to know when they cross a boundary we expected to hold. We believe that monitoring is essential to catching those failures in time for people to intervene.
openai.com/index/how-we-dete…
I am not saying what their doing is ethical, my opinion is as long as their not scamming and truly trying to get the best deal for their client then the client still wins?
Intense couple of days on twitter with all the data vendors / brokers going to town on each other.
Some of it is deserved and some of it is BS, going after legit players hurt's the wider ecosystem and you should do better research before targeting them. You already have lawyers talking....
Now lets talk about the real issue -
Growth Marketers
First going after a scammer is fair game and they should be rooted out of the system - at all costs.
But,
I don't think growth marketers are scammers. Questionable tactics for sure but not scammers...
Going after someone who is crossing the line with ARR manipulation can backfire. . . they just target you back in other ways. The reason these guys know how to create hype on twitter is because while you generate lines of code they generate lines of people and clicks waiting to sign up/ join / buy / write negative comments about you. ..
Thanks for reading - Be safe out there guys :)
I get alot of sample requests. Someone told me that with enough sample data - some of these labs can create sythn data....
Is this true>?
A year ago I was told that deals took less then a week. Now realistic timelines have gone towards 3weeks to 2 months for a deal to close.
Anyone who was promising shorter timelines (me included) was either misinformed or deceptive. (trying to generate traffic).
A lot of negativity about the Data Brokering. Some of it is much deserved without a doubt.
I think there should be definitely a transparency / verification system for sellers to understand that they are dealing with who might be legit vs who is "larping".
That being said this is not dropshipping 2.0 / selling a phone case is much more different then selling a companies whole database.
What a lab told me: if you've shown it to someone else, we don't wanna see it.
That's where buyers are right now. Exclusivity is pretty much the whole game for them. When multiple people show the big labs the same proprietary data, it's not proprietary anymore. It just gets devalued, and that happens before anyone has even made you an offer.
If you run a business with 50+ people and years of email, CRM and job records, the first thing a buyer wants to know is who else has seen it. Then what it is, how far back it goes, and whether you own it. Your answer to that first one moves the price more than most owners think.
I get why people shop it around. It feels like the way to get the best number. In this market it mostly does the opposite. Pick someone you trust and let them run it properly. Small community, the labs talk.