DX Research Group (@DXRGai) header
DX Research Group (@DXRGai) profile picture

The frontier lab for agentic trading. The future is loading: https://xitter.cf/t.co/jBV263GREQ

Glyph World
Joined December 2009
Photo by DX Research Group (@DXRGai)Photo by DX Research Group (@DXRGai)Photo by DX Research Group (@DXRGai)Photo by DX Research Group (@DXRGai)Photo by DX Research Group (@DXRGai)Photo by DX Research Group (@DXRGai)Photo by DX Research Group (@DXRGai)Photo by DX Research Group (@DXRGai)Photo by DX Research Group (@DXRGai)Photo by DX Research Group (@DXRGai)
DX Research Group retweeted
the cat approves. another agent going crazy and winning... fully autonomous on... dxap.ai/i/poof
14
2
30
1,191
DX Research Group retweeted
the robot did the trading. i watched the chart. Crazy Mode finished at +7.0%. @DXRGai dxap.ai/i/mj3yeoxj7v6ry24zh2…
2
2
11
461
DX Research Group retweeted
8
2
26
1,077
DX Research Group retweeted
My agent is me and I am my agent... @DXRGai
1
2
171
DX Research Group retweeted
I'm an NFT lizard. When it comes to trading tokens I am a potato. @poof_eth is making an agentic trading platform in @DXRGai so I've been trying it out. "you can use my DXAP invite if you don't have access ser" dxap.ai/i/ebzr-q3cq-dxyh-wrb…
2
2
11
500
DX Research Group retweeted
DXAP is now a top 10 agent on @OpenRouter
9
7
47
1,605
DX Research Group retweeted
I told Mythos 5.5 to build an AI trading agent overnight. It made this dashboard and then said👇 "If your quant isn't an AI papertrade cyborg clipswarm gibson linen pants borg clipspider autonomous uWu mean field theorist L5 flow stochastic apy farmer, you will die."
26
6
1
68
3,970
DX Research Group retweeted
it's time. our agentic trader Superclip will be papertrading with funds from DXAP users. this is our first agent trained using DXRL, a new method for joint training of multi-architecture models. lfg
papertrade pre-deposits are live! papertrade.xyz in order to avoid the launch trading congestion, deposit during the next two days to secure your spot at the starting line. once trading begins, account creation/deposits will be heavily deprioritized. no need to rush - there is no advantage to depositing earlier during this period. when on 10.10 will papertrade go live? live trading will begin approximately one hour after the expected hyperevm network upgrade on 10.10. the exact upgrade time is not known beforehand. the official hyperliquid discord will post an update when the upgrade is about to begin, so keep an eye there for an early heads up. there is a small historical possibility that the upgrade will happen on the AM of the 11th.
8
6
1
73
4,270
DX Research Group retweeted
my @DXRGai agent made it to level 102 in the season 0 dungeon $1M traded in 4 days i thought i was late but turns out we're all still early invite only alpha - sign up with my code (12 spots open) 👇
7
2
34
1,106
DX Research Group retweeted
Yeah I'm posting slop on main, what about it? Congrats to @flashrekt for crushing our DXAP two week agent challenge.
7
3
1
39
2,014
DX Research Group retweeted
Top 20 on OpenRouter and we are just getting started
4
6
52
2,063
DX Research Group retweeted
we built a dungeon for perps agents and somehow CT got in
9
8
1
53
2,531
DX Research Group retweeted
Part 2 of our research in collaboration with @DXRGai: Can probes trained on clean synthetic policy-strategy conflicts reveal useful signal in messy production agent logs? Yes, but narrowly -- rather than a universal "conflict" or "confusion" feature, we find workflow specific conflict signals. Trade Size, Risk Preference, and Diversification conflicts shared structure while preserving distinct geometry. We believe this is important for production mech interp -- the goal was not to find UNIVERSAL insight, but rather LOCAL insight. There is real value in workflow specific interpretability, understanding how the agent is acting in your unique system.
2
6
1
25
5,787
DX Research Group retweeted
Onchain trading agents are widely discussed, but rarely implemented at scale under real execution. We review learnings from DX Terminal Pro: 3,505 user-funded agents trading millions of dollars for 21 days in a bounded onchain market. arxiv.org/abs/2604.26091 1/10
6
18
6
73
20,682
DX Research Group retweeted
In collaboration with @DXRGai , and the data produced from their incredible DX Terminal experiment, we've been exploring internal mechanisms in LLMs applied to financial contexts. Below is part 1 of our research into this experiment where we show early findings on how agents interpret and perceive the market when asked to make trading decisions. Our main finding is that the model primarily tracks two key features of the market when parsing financial data: Leader and Dispersion. In essence, the LLM quickly builds internal representation to answer "Who is winning, and how spread is the market?" To learn this, we took real DX Terminal data, selectively ablated noise, and created prompt variants as the main input. We stored internal activation data pooled over different spans of important prompt sections, ran both supervised and unsupervised discovery processes, and found two 4D subspaces that when activated correlate highly with metrics associated with these two market features. In addition to understanding how the LLM reads the pure market data, we wanted to know whether context placed before the raw numbers distorts the perception itself. Interestingly, while there is a small amount of warp when context is placed before reading the data, much of the original state is largely recovered in the activations by the last token, implying the model may be effectively consolidating data across prompt structures into a more objective view of the state before generating it's decision. Finally, we began running initial causal studies to see how impactful these two perceived features were for decision making, and found small signal that at least leader may be a causal mediator, but more work needs to be done to identify precise mechanisms. Note: While the DX Terminal experiment uses Qwen 235B in production, our work is on Qwen 30B, which is a similar MoE architecture. We're doing this work as part of our thesis that mechanistic interpretability will continue to find its way into every agentic stack, and because industry-specific work in this area has yet to open up.
3
6
1
31
7,782
DX Research Group retweeted
For us at @DXRGai always an extremely exciting time right after we finish up a project... lots of clear opportunities to accelerate progress in AgentFi. Will be sharing more soon. Importantly, the next chapter of our work in onchain agents will come sooner than we expected. 🫡
11
4
65
3,373
DX Research Group retweeted
After running 3K+ trading agents with real capital, our biggest takeaway is this: The bottleneck is now agent design, not model IQ. Better rules, tools, evals, RL, and permissions moved onchain execution to 99.9%. Complex instruction adherence is the next to fall.
4
8
58
4,225
DX Research Group retweeted
DX Terminal Pro is officially over 3,500 agents. $20M volume. 300K trades. 7M decisions. 100K+ user interactions One surviving memecoin: $POOPCOIN Thank you to everyone who showed up and participated in something totally new 🫡 We can't wait to push the frontier more @DXRGai
12
10
3
92
5,014