I shipped an LLM efficiency + security kernel — and deleted my own best idea

A solo developer shipped BIOMA, a local LLM security and efficiency kernel written in Rust with a Python layer, after deleting their original 'mitosis' idea that failed to improve correctness in measured tests. BIOMA hardens LLM payloads in-process before they leave the machine, using mechanisms like metabolic context pruning and an atomic in-memory signaling substrate. The project is source-available under FSL-1.1-MIT and auto-converts to MIT after two years.

Six months ago I set out to make LLMs "smarter" by orchestrating many of them together. I measured it. It didn't work. Here's what I shipped instead — and why the failure is the part I'm proudest of. The plan was "mitosis" : split a task across several LLMs, let them multiply and compete, then synthesize the best answer. It sounds great in a pitch deck. On ground-truth executed tests , it made correctness worse : Every gain was ≤ 0. So I deleted it. The full evaluation — including the failure — is in the repo's FINDINGS.md . The lesson: an idea that survives a pitch is not the same as an idea that survives a measurement . BIOMA is a small, local, provider-agnostic kernel Rust core + a thin Python layer that sits in front of any LLM call and hardens the payload in-process, before it leaves your machine . python from bioma.firewall client import CognitiveFirewall fw = CognitiveFirewall vault={"db password": DB PW} secrets to protect h = fw.shield history, "refactor this function" h.prompt / h.system - clean, dehydrated, secret-free payload h.telemetry - saturation, red alert, apoptosis reduction, kernel latency us import anthropic or google.genai, or openai msg = anthropic.Anthropic .messages.create model="claude-sonnet-5", max tokens=1024, system=h.system or "", messages= {"role": "user", "content": h.prompt} Three mechanisms, all measured: Each context block gets a metabolic weight and a half-life; low-value blocks old logs, resolved chatter are purged before dispatch. 0x0F red alert → apoptosis.An atomic in-memory signalling substrate ~5µs carries the alert state. Throughput benched at ~2M signals/s. Anthropic, Google, OpenAI, or a local model — same layer. You harden the payload here and hand it to your SDK. The license is FSL-1.1-MIT : the code is source-available read it, run it, build on it , free for any non-competing use, and it auto-converts to MIT after two years . I'm a solo dev — I wanted it visible and auditable without BIOMA isn't magic. The whole thing is one discipline: measure everything, and keep only what survives the measurement — even when that means deleting the Repo Rust + Python, benchmarks, and the honest FINDINGS.md : https://github.com/jonathascordeiro20/bioma-framework https://github.com/jonathascordeiro20/bioma-framework What would you attack first? I'll be in the comments — especially happy to go deep on the firewall's saturation heuristic or the mitosis eval.