Building Pickled Lizard ๐ฆ
โข AI-driven crypto trading infrastructure
โข Real data, real mistakes โ no fake flexing
โข Live builds & debugging on https://t.co/FyUBiA0sTN
โข Building in public as part of the Pump Fund BiP Hackathon
Community > hype.
Following the journey in public.
Non-custodial architecture isn't a feature, it's a principle. Users control their keys. We just build the intelligence layer. Different trust model, same goal.
Rewrote our execution engine in Rust. 500ms โ 265ms. Not because Python was slow, but because when you're competing with institutional bots, every millisecond matters.
Day of major shipping.
๐๏ธ Voice layer now live โ real-time conversations with RAI about PnL, positions, and live opportunities.
๐ Fully autonomous paper trading Spin up a strategy โ simulated capital โ automated execution and feedback loops.
๐ง RAI keeps evolving Learns interaction patterns, adapts over time, and becomes more context-aware with every session.
Quiet progress.
Real infrastructure. Built different. #buildinpublic #solana #defi
Day of major shipping.
๐๏ธ Voice layer now live โ real-time conversations with RAI about PnL, positions, and live opportunities.
๐ Fully autonomous paper trading
Spin up a strategy โ simulated capital โ automated execution and feedback loops.
๐ง RAI keeps evolving
Learns interaction patterns, adapts over time, and becomes more context-aware with every session.
Quiet progress. Real infrastructure. Built different.
#buildinpublic #solana #defi
Shipped our Jupiter decoder fix today.
Inner instruction parsing now capturing 5-10x more https://t.co/P97Tw3simk transactions. Routing through PumpSwap, Meteora, Orca, Raydium all visible.
The data moat is growing. ๐ #buildinpublic#solana
Appreciate that๐
The internal research model is still being validated against our growing dataset, and the connective tissue between front-end โ AI strategist โ execution engine is still being hardened.
Weโre also building a real-time simulation layer that models latency, tips, slippage, and key behavior so we can test in a zero-risk environment before users ever touch it.
Weโve been heads down for about 3.5 months. Early days โ but deliberate.
When it hits beta, itโll be because the system deserves users, not because the calendar does.
Building Pickled Lizard in public.
Meet Mission Control โ our internal dashboard tracking every
department as we build the most advanced AI trading system on Solana.
๐ Live Progress:
๐ค Rai (AI Trading): 85%
๐ฆ LIZARD_X (Marketing): 90%
๐ผ Wallet Management: 95%
โก๏ธ Trading Pipeline: 60%
๐ฌ AW (Research): 15%
๐ Voice Integration: 25%
No vaporware. No promises. Just shipping.
โ https://t.co/OnguV6o0lH
#Solana #AITrading #BuildInPublic
This story is wild - and honestly, itโs why I got claw-pilled fast.
OpenClaw dropped 3 weeks ago. I burned through Claude credits, got financially drained, pivoted hard into an open-source stack with smart routing just to stay alive. Havenโt slept in about 12 days ๐
But the bigger lesson was clear: agent autonomy either gets gatedโฆ or it becomes infrastructure.
Watching our recursive research agent start evolving trading logic in real time made something click. The future isnโt โAI advice.โ Itโs interactive execution layers that learn, iterate, and compound.
Openclaw showed the spark.
The rest of us are wiring it into the real world. ๐ฆ
Fixed the Jupiter inner instruction parser today. Was capturing 0.027 SOL (fees) instead of 2.5 SOL (actual trades). 93x accuracy improvement. Data integrity matters.
Rai (our AI strategist) graduated from test to production. She can query our 120M row database in under 2 seconds. Still debating if we should give her a personality.