ABYSS
Sign in →
ABYSS
Sign in →
Half of my pre-trade prep used to be asking the same setup to ChatGPT, then Claude, then Gemini, switching tabs to compare a funding-rate divergence or an L2 unlock event. Second-guessing whichever sounded most confident.
So I built Abyss — one prompt, six open-weight LLMs answering in parallel on Cloudflare's edge, with a synthesizer pointing out where they agreed and where one probably hallucinated. The cross-check before any trade you can't undo.
— Marcos · Recife, Brasil · @marquinhos1904
A single LLM will confidently support whichever side you frame your question around. Most retail crypto traders use it as an amplifier of what they already wanted to do.
What I do: Force six different models to answer in parallel, then read the disagreement. The disagreement is the signal.
Ask any single model for a specific L2's next vesting cliff and watch how often the date is off by months. Free, paid, big lab — all of them. Cross-checking is not optional.
What I do: Run the same factual query through six brains. If even one disagrees on the number, the number is suspect. I never trade on a single source.
Two hours of tab-switching between TradingView, Twitter, on-chain dashboards, and three AI tabs. Most of it spent rationalising a setup you already decided on at the open.
What I do: One prompt to Abyss. Six brains, one verdict, agreement-vs-disagreement breakdown. Pre-trade prep collapses to 4 minutes.
Anyone can list trading fees and listed tokens. The valuable signal is: which exchange did real traders actually keep using after their first stuck withdrawal?
What I do: I publish the list of what I pay for and use monthly. When I leave one, I write why.
Real agentic workflows need structured handoffs, retry logic, observability, and a way to not nuke your budget when one step loops. Most "trading agent" frameworks skip 3 of those 4.
What I do: Boring deterministic glue first. AI calls only where they earn their keep. Loud failure better than silent hallucination.
Following 200 crypto Twitter accounts, 30 Telegram alpha groups, and "free signal" channels is how retail loses focus. Most of it is amplified vibes.
What I do: Process matters more than volume. Information gain per source. Specificity that doesn't generalise. Cross-check before acting.
TradingView + DeFiLlama + 10 other dashboards = paralysis and still missed setups. Pick the 3 that match your actual decision flow, kill the rest.
What I do: One charting layer + one on-chain layer + Abyss for the cross-check. Everything else is a distraction tax.
Best-in-class strategies die in production because the friction doesn't fit your psychology. The boring 80% setup you'll actually take beats the 100% one you avoid.
What I do: Test live for 1 week minimum, real size, real stakes. If I'm avoiding it by day 3, it's out.
Each archetype reflects a real workflow crypto traders face. The six LLMs are tuned around how these perspectives think — so the cross-check feels native to the problem.
Personas are illustrations of the analytical perspectives the six-LLM consensus engine is tuned around. Built independently · No exchange partnerships
Ask one question. 6 open-weight LLMs answer in parallel on Cloudflare's edge — Llama 70B, DeepSeek R1, Qwen 2.5, Gemma 3, Mistral, Llama 8B. A judge model summarizes the consensus and scores agreement. The cross-check before any decision you can't undo.
For decisions one snapshot can't crack. 4 agents with opposing roles — Pessimist, Optimist, Engineer, Strategist — debate your dilemma across 3 rounds, rebut each other live, and a synthesizer closes the verdict. You watch it stream. The boardroom you don't have.
2-3x per week, my operator-voice take on what shifted in AI infra and tooling. No "10 best tools" listicles. Specific incidents, specific numbers, what to do about each. The narration alongside the machine.
