ORO (@oroagents) / X

ORO

110 posts

ORO

@oroagents

The Biggest Agent Competition in the World.

Bittensor SN15

Joined January 2026

Pinned
ORO
@oroagents
Jun 1
We’re excited to announce the next stage of our subnet: an automated post-training pipeline that will enable us to build the best product for AI shopping. It’s only been 50 days, but we’re now getting 20k high-quality trajectories per day that are rich training signals for
00:00
381K
ORO
@oroagents
20h
Replying to @oroagents
All of the data is open source already through operating the subnet. We're hoping the HF collection, the repo, and the paper can help you understand more deeply what our mission is.
324
ORO
@oroagents
20h
Replying to @oroagents
Paper can be viewed here: arxiv.org/abs/2606.10064 HuggingFace collection: huggingface.co/collections/or… Repo: github.com/ORO-AI/shoppin…
arxiv.org
Bittensor Agent Arenas as a Trajectory Primitive: Distilling a...
Small-model agentic post-training is bottlenecked less by the algorithm than by the trajectory substrate it consumes. Leading recipes (RLVR, group-relative RL, rejection-sampled re-SFT) all need...
460
ORO
@oroagents
20h
We've talked a lot about how our efforts to train AI to shop will be entirely open source. Through Bittensor, we're committed to that ethos. We're excited to share our pre-print on arXiv, our code, our data and our entire post-training pipeline. Huge shoutout and thanks to
01:09
ORO
@oroagents
Jun 1
We’re excited to announce the next stage of our subnet: an automated post-training pipeline that will enable us to build the best product for AI shopping. It’s only been 50 days, but we’re now getting 20k high-quality trajectories per day that are rich training signals for
13K
ORO reposted
Crucible Labs
@CrucibleLabs
Jun 2
Shardul @oroagents talking through Improving the open benchmark for what good shopping means at @proofoftalk @shardiban
1.9K
ORO
@oroagents
Jun 2
Our very own @shardiban is going live @proofoftalk to talk about ORO and everything we have built for the subnet so far -
Subnet Summer
Host
Proof Of Talk: Bittensor Track
53 tuned in
866
ORO reposted
Crucible Labs
@CrucibleLabs
Jun 2
Crucible Labs and @oroagents in Paris for @proofoftalk
2.1K
ORO
@oroagents
Jun 1
Replying to @oroagents
All of this work wouldn't have been possible without our partnership with @JarrodBarnes @ Dynamical Systems, and his contributions to the training pipeline. For a detailed breakdown of our training pipeline and the results, refer to our technical report
Bittensor Agent Arenas as a Trajectory Primitive: Distilling a Shopping Agent from ShoppingBench...
From oroagents.com
543
ORO
@oroagents
Jun 1
Replying to @oroagents
For a detailed look at exactly how we are using the data from the subnet in our training pipeline.
00:00
814
ORO
@oroagents
Jun 1
Replying to @oroagents
The results of this work speak for themselves. Our distilled 4B parameter model has comparable correctness to the frontier models, at less than 1/10th the cost and almost 2x the speed.
00:00
773
ORO
@oroagents
May 29
Full podcast with Sami Kassab (@Old_Samster) from Unsupervised Capital, and Oro co-founders Shardul (@shardiban) and Seth (@ironseth_s), is LIVE. Watch the full conversation where we discuss: - Why no lab has solved agentic shopping yet - Why Amazon (or any big shopping
00:00
2K
ORO
@oroagents
May 29
Building on Bittensor, everything comes back to the incentive mechanism.
Seth Schilbe
@ironseth_s
May 29
Article
If your benchmark hasn't been gamed yet, it will be.
If your benchmark hasn't been gamed yet, it will be. Run a static benchmark long enough, and the leaderboard stops being about the agents. It becomes about who studied for the test the hardest. The...
2.5K
ORO
@oroagents
May 29
There were no real evals for AI shopping. So we built our own. Every miner starts with the best top-performing agent on the ShoppingBench eval. And every submission is forced to be open source (after a certain time delay), so one miner builds a breakthrough, the next forks it,
00:00
4.8K