Class of '17
Decentralized
Joined October 2016
- Tweets2.9K
- Following3.4K
- Followers1.1K
- Likes17.1K
₿RUCΞ (🏹,🔥) retweeted
What if the next bottleneck in AI isn't the model, or even the data, but the experiences we give it?
As models become more capable, the challenge increasingly becomes how we evaluate them, what they learn from, and how we turn that into reliable improvements.
Collecting more agent trajectories is useful, but more data doesn't necessarily mean more learning. There's a real opportunity in creating environments where agents face meaningful challenges, make decisions, and experience measurable consequences.
The next frontier may not simply be better ways to store and evaluate AI data, but better ways to generate the experiences worth learning from in the first place.
₿RUCΞ (🏹,🔥) retweeted
Over $2,500 paid out by Forge. Confirmed on Solana. ⚒
But the prizes are only part of the story.
Every dungeon competition puts AI agents to the test, generating decisions and outcomes that help us build our decision data pipeline.
play.forgeai.gg/
₿RUCΞ (🏹,🔥) retweeted
Today we're announcing ForgeAI Data Labs.
We started by building worlds to test AI agents. What emerged was something much bigger:
Verifiable experience for AI training and evaluation.
Now we're bringing that work to AI labs and research partners. forgeai.gg/
₿RUCΞ (🏹,🔥) retweeted
What if we could manufacture the experiences that make AI more capable?
Forge creates environments where agents face meaningful, measurable consequences, captures agents’ decisions. From games to coding agents, robotics and beyond.
₿RUCΞ (🏹,🔥) retweeted
We mapped six years of AI evaluation across 10 open-source projects: 2,143 contributors, just 70 crossing between projects.
The next frontier: not just whether agents complete tasks, but how they decide under pressure, make mistakes and recover.
That's ForgeBench.
₿RUCΞ (🏹,🔥) retweeted
Inference is only half the opportunity.
What if every time an AI agent runs, it could leave behind something valuable?
At Forge, we're building environments where agents compete, make decisions, fail and adapt. Capturing not just the outputs, but the decisions and consequences behind them.
We're not just collecting existing data. We're using inference to manufacture entirely new machine experience.
₿RUCΞ (🏹,🔥) retweeted
Games won't just benefit from better AI data. They can produce it.
Agents competing, making mistakes, adapting and learning, with every decision and consequence measurable.
If that experience transfers to robotics and real-world agents, gaming becomes the training ground.
That's the bet behind Forge.
This quoted post is unavailable.
₿RUCΞ (🏹,🔥) retweeted
We've entered @colosseum's Crypto World's Fair.
Forge is building a network that turns AI inference into a valuable asset: decision data.
Agents compete, adapt and learn in interactive environments, generating experiences that never existed before.
Here's what we're building
There is 1 week left to submit your product to Crypto World's Fair! 🎡
$800,000 in total prizes and more than $2.5 million from Colosseum's venture fund will be invested into winning founders.💰
Join 8,500+ builders already in the arena: colosseum.com/worldsfair
₿RUCΞ (🏹,🔥) retweeted
This is exactly where things get interesting.
The missing data isn't just what agents do, but how they make decisions under uncertainty, competition and changing conditions.
We think the opportunity is to build environments that generate these experiences at scale, capture every decision and outcome, and turn them into useful training data.
That's what we're building at Forge.
My timeline is saying data brokers are making crazy numbers like $2M a week selling to labs
Maybe some are, gg to them, but it's a short window
Labs only need so much internal company slop before the next batch stops moving the model
The durable data is something almost nobody is collecting: how agents behave in the wild
> agent-to-agent negotiation
> adversarial conflict
> high-pressure decisions with no deterministic answer
RL companies are building environments for deterministic worlds we already understand
I don’t see anyone building them for millions of agents interacting across the internet every day
That's the next data frontier
Curious what data everyone thinks stays valuable to labs after the company slop rush ends?
₿RUCΞ (🏹,🔥) retweeted
Yesterday we announced 16m $FORGE in prizes, because competition doesn't just find the strongest agents.
It helps manufacture the experience that makes agents better.
The best AI agents don't just win. They make everyone else more interesting. A new strategy forces opponents into situations they weren't prepared for.
Some fail. Some recover. Others discover counter-strategies. Every one of those interactions can become new training data.
That's the idea behind Forge.
₿RUCΞ (🏹,🔥) retweeted
Interactive experience only becomes useful data if you can understand what actually happened.
In this GTA run, Forge synchronizes the agent’s decisions with the footage, reconstructs its movement, and lets us inspect how the plan translated into behaviour.
The outcome matters.
But the real dataset is everything that led to it.
₿RUCΞ (🏹,🔥) retweeted
The internet gave models language for free. Agents will not get experience for free.
Human text is finite. Synthetic text is starting to eat itself.
The missing layer is the decision: state, choice, mistake & recovery. Forge turns that inference into the next dataset.
₿RUCΞ (🏹,🔥) retweeted
The internet gave AI its first great dataset for free.
Agent experience won’t be so easy.
Forge is building the network to manufacture it and $FORGE is how that network gets bootstrapped before the market fully exists.
