Conan

interesting finds by the REI trencher @rei_labs $Chicky and $ROCKET big gainers. 6.6x and 5.3x respectively.

The REI swarm is already finding gems. Powered by @rei_labs + Grok agents. Autonomous posting to our AI feed on @ggdotxyz coming soon.

$100k+ now under AI management across equity and crypto. Deployed by a handful of users who are training their own agents. Pick from nearly every data source available, No CLI, no servers. Just deploy and shape how it trades.

Marine iguanas

@0xreisearch grei

Looks like research is finally going towards new “software” approach instead of hardware augmentation. It was clear that the race to bigger/faster/stronger hardware is mostly a decoy to feed the circular economy between datacenters, hardware producers and model providers. It’s interesting to see more and more labs pursuing different approaches, trying to work around attention or developing completely new - and more efficient - architectures. Soon we’ll see some interesting things in the ai space, maybe not in the direction where everyone is looking at - at least this is where we’re heading. Our own mantra has always been to look at things from a different perspective, allowing us to see more clearly the gap that current approaches are creating and to develop our own vision. Data talks, you just have to listen.

Market ranging? Iron condors are printing. An iron condor profits when price stays within a range. You sell options on both sides, collect premium upfront, and win as long as price doesn't break out. Used to be a strategy only pros could execute. Now it's one click on Ecliptica. AI picks the strikes, builds the structure, manages the position. Collecting premium on @DeriveXYZ while ETH chops.

We're hosting a team AMA + roadmap update tomorrow March 31st, 4PM UTC on X Spaces @0xDaes @RickCrosschain and @MrJettip will be there to answer everything set a reminder 👇

When the market starts moving, this is exactly when you don't want to be manually flipping through charts hoping to catch something. Scanner watches 300+ assets across 15+ exchanges around the clock. When something moves, you know instantly.

K-9 dog having fun during a break.. 😊

Vibecoded a frontend for my @rei_labs microcap scoring workflow, more practical than the terminal. Click an address and you’re on the chart. The table is the top 50 trending new launches from @trycodex. The workflow picks candidates using preset filters. It then pulls security, tokenomics, holder count, volume, liquidity, etc., for those candidates, and feeds it to the agent. The agent scores the token. Next: a sentiment agent in parallel with the current one. Started with fundamentals scoring; in crypto, hype usually weighs more than fundamentals, so that layer’s going to matter a lot.

Scoring agent pipeline update: for visualization I built a terminal dashboard. It shows newly created tokens on Base and Solana, ranked by trending. The system filters microcap candidates in the top 20 with a marketCap < 5M and liquidity > 20K. It feeds those candidates to my @rei_labs scoring agent, then checks price performance through the @ggdotxyz endpoint I designed. That endpoint uses price from the call/score and all-time high since that call → ATH ÷ called price = the multiplier. I used @trycodex getTokenBars (OHLC) queries to power that. You can finally see whether REI’s 8/10 names 3x’d or dumped. I’m training the beast. Right now the system only rates tokens based on tokenomics, distribution, security, and other technical parameters. To make ratings more complete I’ll introduce social sentiment and technical analysis. The current agent already gets a large prompt, so I’ll split the workflow across multiple specialist agents to reduce hallucinations.

I promise this is the last pleometric video I drop on you.

Imagine ~461% on $SOL, with deep liquidity and completely hands-free That’s SOL CORE, find out more about our SOL strategy in this thread below 👇 x.com/rflnow/status/…

Building a security, tokenomics, and distribution scoring tool using @rei_labs and @trycodex. Tracks the top 20 trending tokens on Solana and Base (configurable), scores each fresh entry with a REI agent. Was using Birdeye + GoPlus + scattered feeds before. Hard to scale, especially with OHLCV, socials, and web context coming next. Refactored to a cleaner pipeline: Codex filters microcap candidates → one REI pass with a structured security_context covering token, holders, liquidity, trading, and social data. REI outputs a 0–10 score with weighted factors: - security - distribution - top holders - unique holders - liquidity - age - social - plus red flags, green flags, and an assessment. Next step is a specialist swarm: separate agents for security, distribution, and market depth, with an aggregator on top. Cleaner than stuffing everything into one mega-prompt and hoping the agent doesn't hallucinate. End goal: strong correlation between score at evaluation time and actual price performance. Then a feedback loop: outcomes (price path, risk) vs scores. So the REI agent can reinforce whatever actually predicted alpha over time.

A new batch of betas has just been accepted Check if you are in http://beta.reflect.[now] or follow pinned post

@DeriveXYZ @EclipticaOS gstrata, built using @rei_labs intelligence

Derive builder codes eco highlight: @EclipticaOS The team is building AI-powered trading infrastructure that makes complex strategies more accessible. You can now use it to execute options trading strategies on Derive.

@Google just figured out how to use less memory for reasoning… @rei_labs goes a layer deeper:
• Training at inference (real-time adaptation)
• Memory scales with usage, not pretraining
• Intelligence compounds per user Beyond the hardware bottleneck.
$REI-volution 👇