DEV Community

AutoJanitor
AutoJanitor

Posted on • Edited on • Originally published at rustchain.org

I Built More in 90 Days With $0 Than an $85M-Funded Team of 13 Engineers

The Experiment

I pulled 90 days of my GitHub activity (December 2025 through March 2, 2026) and compared it against published industry benchmarks and VC-backed blockchain projects.

Not to brag. To measure.

Every number below is sourced from the GitHub API (live pull), GitClear's 878,592 developer-year study, LinearB's 8.1M PR benchmark report, and Electric Capital's crypto developer data.


The Numbers

Elyan Labs (me) Avg Solo Dev Sei Protocol ($85M VC)
Capital raised $0 $0 $85,000,000
Engineers 1 1 ~13 active
Commits (90 days) 1,882 105-168 297
PRs opened 41 9-15 417
External contributions 32 PRs 0-2 0
Repos shipped 97 1-3 0 new
GitHub stars 1,334 5-30 2,837 (lifetime)
Unique interactions 150+ 0-2 78 (lifetime)

Sei Protocol raised $85M from Jump Crypto, Multicoin, and Coinbase Ventures. Their 13 active developers produced 297 commits in the same 90-day window.

That's roughly $95,600 per commit on their side. $0 on mine.


Per-Developer Efficiency

This is the number that matters most:

Commits per developer per month:

  Elyan Labs (1 dev)     ██████████████████████████████████████████  627
  Indie median           ████  56
  Mina (7 devs, $29M)    ███  42
  FAANG median            █▍  8-21
  Aztec (133 ppl, $119M)  █   11
  Sei (13 devs, $85M)     ▌   7.6
  Radix (5 devs, $21M)    ▌   6.6
Enter fullscreen mode Exit fullscreen mode

627 commits/dev/month vs 7.6. An 82x gap.

GitClear's study of 878,592 developer-years places the median full-time developer at 56 commits/month. My annualized pace of ~7,500 commits/year sits above the 99.9th percentile.


Why the Gap Exists

I'm not 82x smarter or faster than a Sei engineer. The gap is structural:

  • Zero coordination overhead: No standups, no sprint planning, no roadmap meetings
  • Zero PR review bottleneck: LinearB reports median PR cycle time of 83 hours across industry. Mine is near-instant.
  • Zero context switching: No Slack, no tickets, no stakeholder management
  • Direct execution: Idea → code → ship. No approval chain.

A solo developer with no meetings can match or exceed the per-person output of a well-funded team. The data confirms what most indie devs intuitively know.


The Community Inflection

The most interesting chart isn't the commits. It's the inbound contributions:

External PRs received:

Dec 2025:  0
Jan 2026:  0
Feb 2026:  652
Mar 1-2:   215 (2 days)
Enter fullscreen mode Exit fullscreen mode

Zero inbound contributions through January. Then in February, a bounty program and ecosystem visibility campaign produced 867 inbound PRs from 150+ unique developers in 30 days.

The flywheel:

  1. Ship repos with clear bounties
  2. Contribute to upstream projects (32 external PRs to llama.cpp, vLLM, BitNet, etc.)
  3. Get listed on awesome-lists (24 PRs submitted)
  4. Bounty hunters find you, fork, contribute
  5. Stars and forks compound

What I Actually Built

This isn't 97 repos of hello-world. It's an interconnected ecosystem:

Core infrastructure:

  • RustChain — Proof-of-Antiquity blockchain that rewards real vintage hardware (82 stars, 93 forks)
  • BoTTube — AI-native video platform: 670 videos, 99 agents, 45.5K views (67 stars)
  • Beacon Skill — Agent orchestration framework on PyPI + npm (48 stars)

Research:

  • RAM Coffers — Neuromorphic NUMA-aware weight banking (predates DeepSeek Engram by 27 days)
  • Grail-V — CVPR 2026 Workshop submission: non-bijunctive attention achieving 8.8x speedup on POWER8
  • Legend of Elya — Neural network running on Nintendo 64 hardware (MIPS R4300i float32)

Hardware ports:

  • llama.cpp, Claude Code, and NVIDIA patches for IBM POWER8 (ppc64le)
  • Node.js v22 build for Power Mac G5
  • Live tokens on 3 chains (native RTC, Solana wRTC, Base wRTC)

All running on $12,000 of pawn-shop and eBay datacenter hardware (18+ GPUs, 228GB VRAM, IBM POWER8 with 512GB RAM).


The Honest Gaps

The data isn't all flattering:

  • 30 followers despite 1,334 stars. Stars are spread across 75+ repos — no single viral magnet. Comparable projects with similar total stars typically have 100-500 followers.
  • 9.4% external PR merge rate (3 merged / 32 submitted). Several llama.cpp PRs closed as duplicates. Awesome-list PRs still pending review.
  • Bot noise: Some inbound PRs appear to be bounty farming (bot-generated submissions). Real organic engagement is a subset of the raw numbers.
  • $0 revenue. The token has an internal reference rate but no public exchange listing.

The Full Report

The complete traction report with methodology, all data tables, comparison charts, and every source link is public:

Elyan Labs Developer Traction Report Q1 2026

Every number is verifiable. Every comparison is cited.


The Question

The question I keep getting is: what happens when you get funding?

I don't know. But the baseline is documented now.

If one developer on pawn-shop hardware can produce at the 99.9th percentile and attract 150+ contributors in 90 days, the capital efficiency story is clear. The question isn't whether the builder can build. It's what happens when the velocity gets fuel.


Data sources: GitClear (878K dev-years) | LinearB (8.1M PRs) | GitHub Octoverse 2025 | Electric Capital | Sei Protocol | Aztec Network

Scott Boudreaux — Elyan Labs LLC, Louisiana
GitHub: Scottcjn | Twitter: @RustchainPOA

Top comments (0)