# Tenki DevRel Strategy

**A Developer Relations program for [Tenki](https://tenki.cloud) — a product of Luxor Technology**

Prepared by opencolin · Draft v1 (outside-in) · 2026-06-08

---

## Cover note

This is a 12-month DevRel proposal for Tenki — the developer platform that lets teams **review code, run CI, and run AI agents** on one Firecracker-microVM substrate. It is an **outside-in v1**, built from public research (tenki.cloud, the Luxor culture doc, the changelog, the competitive field) and a DevRel operating system proven across developer-tooling and AI companies. A **v2 would fold in a working session with the founders.**

**Tenki's reality today (mid-2026):** three products — **Runners** (drop-in GitHub Actions replacement, the cash wedge), **Code Reviewer** (AI PR review; leading recall, trailing precision), and **Sandbox** (disposable microVMs for AI agents, the agent-era bet) — plus a **compute marketplace** (`/compute`, `/hardware`) that is the consumer-facing edge of Luxor's "compute as a commodity" thesis. ~Weekly shipping cadence (81 releases since Apr 2025). Early traction is real but **warm/crypto-network-heavy**; there's **no developer community, no content engine, an empty blog, and almost no public logos.** That gap is the entire DevRel opportunity.

**Parent:** Luxor Technology — Seattle, founded 2017, ~109 people, **profitable**, SOC 1 + SOC 2 Type II, >$1B/yr trading volume. **Founders/leaders to know:** Nick Hansen (CEO), Eddie Wang (CTO, runs Tenki eng), Ethan Vera (COO), Guzman Pintos (CPO, on Tenki), Lauren Lin (Head of Hardware & Software), Hayssem Elsayed (Tenki Product Lead).

**The role this maps to:** Tenki's open **Head of Developer Relations** (SF, full-time as posted) — *"a developer who also happens to be a phenomenal communicator… build in public, public benchmarks against peer tools, OSS PRs with prompt history."* This plan is engineered to be the leave-behind that **does the job before being hired into it.**

**Engagement (preview, detailed in §10):** three paths — **(A) full-time**, **(B) 90-day paid pilot → convert**, **(C) fractional** — so Tenki can de-risk a founding hire the way their own culture prefers (two-way-door decisions, results over process).

---

## 1. Why Tenki, why now

Tenki's thesis, in the company's own words:

> *"Compute is the new digital oil… our goal is to become the leaders in harnessing, refining, and trading compute — unlocking a new asset class."* — Tenki, About
>
> *"AI and HPC are the hot topics right now and it makes sense that Bitcoin miners — a group that understands how to leverage compute and energy — would want to diversify."* — Ethan Vera, COO

That thesis is the wedge. Six reinforcing strengths:

- **Lowest-friction migration in the category.** One `runs-on` change + a Migration Wizard that opens the PR for you. Nobody makes switching cheaper.
- **Owned compute = structural cost moat.** Tenki rides Luxor's captive bare-metal (mining + the new GPU/AI build-out) instead of paying retail cloud margins like every VC-funded competitor.
- **Three products, one substrate.** Runners + Code Reviewer + Sandbox all run on Firecracker microVMs. In the agent era they converge into one thing: *where an autonomous agent writes, runs, and gets its code reviewed in isolation.* No competitor owns that whole loop.
- **A real (if noisy) benchmark.** Code Reviewer leads recall at 68.9% — ~2× the next best. The honest gap (precision 29.9%) is itself the most interesting unsolved problem in the category, and a content goldmine.
- **Enterprise-trust inheritance.** Profitable parent, SOC 1/2 Type II, audited financials — rare for a dev-tools startup.
- **Shipping velocity.** ~1 release/week from a tiny team — a story in itself, and proof the team can keep pace.

**But the lead is perishable.** Blacksmith ($13.5M, GV) and Depot ($10M Series A) are better-capitalized pure-plays with louder brands. GitHub just cut hosted-runner prices ~39%, compressing the headline savings. CodeRabbit owns AI-review adoption mindshare. E2B/Modal/Daytona are racing for agent-sandbox mindshare. **Nick Hansen's own framing applies — "we need to make sure we are the winner or one of the winners."**

**DevRel is the lever.** Tenki doesn't need more engineering hours — it needs **a developer brand, public trust (especially on the precision/honesty front), named references, and a defining agent-era narrative** before better-funded incumbents define the category for it.

## 2. The positioning narrative — Tenki in one paragraph

> Your CI, your code review, and your AI agents all need the same thing: fast, isolated, cheap compute you can trust. Tenki gives you all three on one substrate — drop-in GitHub Actions runners that are 30% faster and up to 60% cheaper, an AI reviewer that catches more real bugs than anyone, and disposable microVMs where your agents can write and run code safely. One `runs-on` change to start. Built on compute Luxor owns, so the economics work. The agent era runs on compute you can trust — Tenki is where it runs.

This is the through-line for every post, talk, benchmark, and customer pitch.

---

## Field notes — Sandbox positioning, onboarding, and the long-running-agent wedge

*Sharpened from hands-on use of the product and a read of the live sandbox market. These are the threads I'd pressure-test with the team in week one — not conclusions, starting points.*

**1. The wedge I'd build the brand on: long-running, self-improving agents.**
E2B, Modal, and Daytona optimize for *short-running* sandboxes — spin up, run a task, tear down. But the fast-moving frontier in 2026 is **long-running and self-improving agents** that persist, iterate, and improve across many steps and sessions. Tenki's persistent volumes, snapshot/restore, and per-task microVMs are positioned for exactly that. Rather than fight the incumbents on short-task execution, **Tenki should claim the category no one else is claiming: the sandbox for long-running, self-improving agents.**

**2. First five minutes is the battleground — and the native ADE is the asset.**
I installed the Sandbox ADE desktop app. The native-app approach is a real differentiator — download, drop in your API keys, connect to your sandboxes — versus the hours it takes to self-host a Daytona. But I hit a snag on first deploy and couldn't complete a run. That's the highest-leverage DevRel finding there is: a **"First Five Minutes" audit should be the very first deliverable** — get time-to-first-successful-sandbox under five minutes, with the native ADE as the thing to polish. DevRel as the fresh-eyes user, feeding product.

**3. Where Tenki wins on Sandbox (competitive map):**
- **Most "sandboxes" are afterthoughts** of a broader cloud business (DigitalOcean-style) — one feature, not the brand. Tenki can own *sandbox-first* mindshare.
- **Daytona** is winning visibility, but it's **expensive** and **hard to self-host** (the OSS is painful to run; you're funneled to paid cloud). Tenki wins by delivering strong value at a more reasonable price *and* genuinely easy local/native setup.
- **E2B / Modal** have strong communities but a short-running-agent focus — leaving the long-running wedge open.
- Net edge: **sandbox-first brand + easy native UX + long-running-agent positioning + the integrated runners→review→sandbox loop + owned-compute price.**

**4. Community acquisition: capture adjacent communities via multipliers.**
The fastest path to a community isn't building from zero — it's capturing adjacent ones: E2B/Modal/Daytona users, and the agent-harness crowds around **Cursor, Codex, and Claude Code** ("if you're not playing well with others, you don't win"). The move is to find the **multipliers** — event organizers, hacker-house operators, accelerator leads, top builders — who bring those developers over, and to run Codex / Claude Code meetups that capture their communities.

**5. Hackathons as lead-gen, not prototype-gen.**
Every event needs one explicit goal — attention, qualified leads, *or* prototypes — and those are different events. The mechanics that compound: put your **customer targets in the judge seats** (so the event sources enterprise pipeline), seed in VCs, **co-sponsor with partners** to share cost and reach, and run a **branded event site + global map** as the always-on hub. Online content creates a more viral flywheel than in-person alone — so every event should spin off content.

---

## 3. North Star

> **Weekly active workspaces running ≥2 Tenki products** (e.g., a green CI run *and* an enabled Code Reviewer, or runner + Sandbox).

One metric that captures PMF (installed), activation (a successful run/review), and stickiness + the multi-product strategy (≥2). Every sub-metric rolls up to it. (Single-product activation is the leading indicator; multi-product is the moat.)

**12-month targets**

| Metric | Today (Jun 2026) | T+12 mo |
|---|---|---|
| Signed-up workspaces | instrument | — |
| Weekly active workspaces (≥1 successful run/review) | instrument | 3,000 |
| % running ≥2 products by day 30 | instrument | 40% |
| Time-to-first-successful-run (the "first 5 minutes") | instrument | < 10 min median |
| Named customer references | ~0 public | 20 |
| Quotable public logos | a few (warm) | 8 |
| Active community engineers (Discord) | 0 | 2,500 |
| Newsletter subscribers | 0 | 8,000 |
| Build-in-public streams | 0 | 40+/yr |
| Public benchmarks shipped | 1 | 6–8 |
| Paid Team/Enterprise accounts | instrument | 150 |
| Conference talks / flagship events | 0 | 8–10 |

---

## 4. The seven initiatives

A proven 7-initiative template, tuned to Tenki's wedge. Each attaches to a SWOT lever (see `swot-analysis` appendix; codes referenced inline).

### Initiative 1 — The Tenki Benchmark Series (the trust engine)
**Attacks:** W2 (precision/trust gap), S5 (recall lead), T7 (AI-review credibility)

The defining content asset. Quarterly + per-model-release benchmarks, **always reproducible (open harness)**:
- **AI Code Review:** recall *and* precision vs CodeRabbit, Greptile, Copilot, Cursor Bugbot, Devin — on bugs **not** in Tenki's own benchmark repos. Publish where Tenki *loses*; that honesty is the credibility (and it's literally a Luxor value: "honesty over kindness").
- **Runners:** speed + $/run vs GitHub-hosted, Depot, Blacksmith, Namespace, with a public cost calculator.
- **Sandbox:** boot latency + cost vs E2B, Modal, Daytona.

Public leaderboard at `tenki.cloud/benchmarks`. **We own the metric definitions before competitors do.** *Drop #1 is the AI-reviewer teardown already scaffolded in `/teardown` — Tenki vs CodeRabbit vs Greptile on real OSS bugs.*

**🏁 TenkiBench — the flagship: the agent × open-model matrix.** The map the category is missing: score every coding agent (Claude Code, Codex, Aider, Goose, ForgeCode…) against every open model (Qwen3-Coder, GLM, Kimi-K2, DeepSeek, Llama…) on real coding tasks, then slice it both ways — *pick a model → which agent gets the most out of it; pick an agent → which open model it shines on.* The pairing matters more than either pick alone, and no neutral benchmark maps it (SWE-bench / terminal-bench rank models alone). It runs **end-to-end on the stack**: **Runners** orchestrate a daily sweep (schedule → fan out → score → publish), **Sandbox** runs each agent×model cell in its own microVM in parallel (where the ~25-agent dockingstation harness lives), and **Luxor `/compute`** serves the open models on GPU. Then the product turn — **bring-your-own-evals**: developers point their own task suite at the matrix and run it on Tenki, turning the leaderboard into a hosted eval surface that pulls devs in as users. Neutral by design (Tenki ships infra, not a coding agent), so we publish the upsets straight. Phasing: fixed matrix weekly → daily → self-serve.

### Initiative 2 — Agent-Era Reference Architectures (Sandbox + BYO-agent)
**Attacks:** S6 (Sandbox), O2 (agent explosion), O4 (CI-convergence), T4 (E2B/Modal mindshare)

Tenki's agent story needs canonical, maintained references:
- **Claude Code + Tenki Sandbox** — co-marketed with Anthropic.
- **Codex + Sandbox** — co-marketed with OpenAI.
- **Cursor / Devin / Cline / Aider + Tenki** — quickstart per harness.
- **"Give your agent root without giving it yours"** — the security-isolation pattern, with microVM-vs-container explainer.

Each ships a GitHub repo + 5-min video + written guide + a **cost/throughput calculator** (input: agent runs/week → output: $ + isolation guarantees vs DIY). Doubles as the "build in public running Tenki against a real OSS bug" cadence the JD asks for.

### Initiative 3 — The 2-Minute Migration Engine
**Attacks:** S1 (drop-in), O1 (GitHub pricing backlash), W6 (margin pressure), T2/T6 (commoditization)

The runner growth motor:
- **Migration content + Migration Wizard storytelling** — "I moved a real monorepo's CI to Tenki in 2 minutes, here's the diff and the bill."
- **Honest comparison guides** — Tenki vs GitHub Actions, Depot, Blacksmith, Namespace (own the evaluator's search).
- **Public cost calculator** — model your GitHub Actions bill on Tenki; lean into the GitHub-pricing-change moment.
- **First-five-minutes overhaul** — partner with eng/product to crush time-to-first-green-run (a North Star input).

### Initiative 4 — Community + Content Engine
**Attacks:** W3 (no community/brand), W8 (unknown)

The operating system that's currently missing entirely:
- **Discord** — `#runners`, `#code-reviewer`, `#sandbox`, `#claude-code`, `#codex`, `#showcase`, `#benchmarks`.
- **Weekly office hours** + **weekly build-in-public stream** (run Tenki against a live OSS bug — JD requirement).
- **Content cadence:** 1 Tenki-team deep post / 2 weeks; 1 "How I ship with Tenki" community post / week; biweekly newsletter; weekly YouTube (migration + Sandbox walkthroughs); coordinated HN/Reddit/X launches.
- **Podcast guesting:** Latent Space, AI Engineer, ChangeLog, Software Engineering Daily, Devtools.fm, Practical AI.
- **Seeded distribution:** cross-pollinate from **dabl.club's 86k developers** (attributed via UTMs) — instant top-of-funnel most startups spend a year building.

### Initiative 5 — Design Partner Cohort (manufacture the references)
**Attacks:** W1 (no public customers), W8

12-week cohorts · 8–10 companies · 4 batches/yr. White-glove from the Tenki team:
- Dedicated Slack channel, weekly founder office hours, custom onboarding for their agents/CI, early access + lifetime discount.
- **In exchange:** published case study with quantified ROI ($ saved, build-time saved, bugs caught), quotable quote, co-presented talk, logo on `tenki.cloud/customers`.
- **Target list:** AI-native teams already on Claude Code/Cursor and likely to endorse — Vercel, Linear, Replit, Sourcegraph, Browserbase, Modal, Cognition, plus Luxor's existing network (Blockware, Citrea, Sia, etc.).
- **Quarterly public Demo Day livestream** — manufactures social proof at scale (YC playbook).

### Initiative 6 — Tenki for Open Source + Champions + Hackathons
**Attacks:** W1, W3, W5 (not OSS → no contribution flywheel), O7

Tenki isn't open-source, so we substitute the OSS flywheel with three mechanisms:
- **Tenki for Open Source** — free Runners + Code Reviewer for public repos. Every OSS maintainer who adopts is a public reference and a distribution surface.
- **Champions program** — recognize + reward sustained community contributors (content, examples, support).
- **"Agent Camp × Tenki" hackathon series** — leveraging Colin's event engine: devs build on Sandbox; winners get credits + spotlight. Plus **OSS PRs where Tenki was the assistant** (diff + prompt history attached — JD requirement).

### Initiative 7 — Strategic Channel & Ecosystem Integrations
**Attacks:** O6 (co-marketing), T2/T3/T4 (incumbent threats → distribution)

Convert ecosystem threats into channels:
- **GitHub Marketplace** — first-class listing for Runners + Code Reviewer.
- **Anthropic (Claude Code) + OpenAI (Codex) + Cursor** co-marketing via the Sandbox reference architectures.
- **AI-agent registries / directories** — featured Sandbox placement.
- **The Luxor compute tie-in** — enterprise/compute story for the `/compute` + `/hardware` audience.
- **Alliances:** Vercel (deploy/CI templates), and the AI-coding ecosystem broadly.

---

## 5. Phased 12-month plan

Three phases, each ending in a marketing moment.

### Phase 1 — Foundations (Mo 1–3)
- Instrument the funnel + define the DevRel scorecard with product.
- First-five-minutes overhaul; top onboarding frictions fixed/PR'd.
- Discord + weekly office hours + weekly build-in-public stream live.
- **Benchmark Drop #1** (the AI-reviewer teardown).
- Migration comparison guides + cost calculator shipped.
- Cohort #1 applications open; dabl.club cross-pollination begins.
- **Milestone:** community seeded (500+ engineers), funnel instrumented, 1 benchmark live, time-to-first-run baselined + improving.

### Phase 2 — Agent-era push (Mo 4–6)
- Sandbox reference architectures: Claude Code, Codex, Cursor.
- Anthropic/OpenAI co-marketing motion opened.
- **Agent Camp × Tenki hackathon #1.**
- Cohort #1 Demo Day (public livestream, 8 customer stories).
- **Benchmark Drop #2** (runners + sandbox head-to-head).
- Tenki-for-Open-Source launch.
- **Milestone:** 1,500 weekly active workspaces, 8 case studies in flight, agent-era narrative landed.

### Phase 3 — Scale + enterprise (Mo 7–12)
- Flagship event presence: **AI Engineer World's Fair, GitHub Universe, KubeCon.**
- Enterprise references + the Luxor compute/enterprise narrative.
- Champions program; Cohorts #2–3; Benchmarks #3–4.
- First analyst attention.
- **Milestone:** 3,000 weekly active workspaces, 20 named references, 150 paid Team/Enterprise accounts.

---

## 6. How DevRel pre-sells revenue

Tenki monetizes via **usage** (per-minute runners, per-review, per-second sandbox) → **Team/Enterprise plans** → and feeds the **Luxor compute/enterprise** motion. DevRel manufactures the proof at each step:
- **Benchmarks** = the trust that converts trials to paid usage.
- **Reference architectures** = "look what you can build" portfolio for enterprise + the compute audience.
- **Cohort case studies** = sales narratives with quantified ROI.
- **Community → pipeline** = activated workspaces are the top of the sales funnel; track community→paid conversion explicitly.
- **DevRel engineer as bench overflow** (later): paid integration/onboarding engagements turn DevRel into partial revenue and force a tight product-feedback loop.

---

## 7. Competitive positioning (the three battlecards)

- **Code Reviewer →** *trust, not just recall.* Lead with reproducible benchmarks and radical honesty about precision; "merge decisions, not nitpicks." Beats CodeRabbit's adoption and Greptile's context on credibility.
- **Runners →** *the cheapest, lowest-friction switch, on owned metal.* Lean into the GitHub-pricing moment and the one-line migration; the cost moat is Luxor's compute.
- **Sandbox →** *the agent's safe place to run.* microVM isolation > containers for untrusted agent code; co-marketed with the agent harnesses themselves. Beats E2B/Modal on the "part of a whole loop" story.

---

## 8. Content calendar — first 90 days

| Wk | Drop | Channel | Persona |
|---|---|---|---|
| 1 | "I migrated a real monorepo to Tenki in 2 minutes — here's the diff and the bill" | Blog + X | CI owners |
| 2 | **Benchmark Drop #1:** AI reviewers on real OSS bugs (the teardown) | Blog + HN + X | Eng leads |
| 3 | "Give your agent root without giving it yours" — Sandbox + Claude Code | YouTube + Blog | Agent builders |
| 4 | Tenki vs GitHub Actions vs Depot vs Blacksmith (honest comparison + calculator) | Blog (SEO) | Evaluators |
| 5 | Cohort #1 applications open | Landing + email (dabl.club) | AI-native teams |
| 6 | "The precision problem in AI code review" (POV essay) | Blog + X | Eng leads |
| 7 | Codex + Tenki Sandbox reference architecture | GitHub + Blog | Agent builders |
| 8 | Community story #1 ("How I ship with Tenki") | Blog + LinkedIn | All |
| 9 | Weekly stream milestone: live-fix an OSS bug w/ Tenki | YouTube/X live | All |
| 10 | Customer story #1 (cohort reveal) | Blog + LinkedIn | Enterprise |
| 11 | "Runners, review, sandbox: why they're one product" (the convergence thesis) | Blog + HN | Eng leads |
| 12 | Agent Camp × Tenki hackathon announce | Landing + email + X | Agent builders |

---

## 9. Events (12-month plan)

| Quarter | Event | Investment | Goal |
|---|---|---|---|
| Q3 2026 | AI Engineer World's Fair | Talk + booth (~$25K) | 500 demos, cohort applications |
| Q3 2026 | **Agent Camp × Tenki** (own event, dabl.club engine) | ~$30K | Sandbox activations at scale |
| Q4 2026 | GitHub Universe | Booth (~$30K) | CI-owner acquisition |
| Q4 2026 | KubeCon NA | Talk (~$5K) | Platform-eng persona |
| Q1 2027 | DevRelCon / Render ATL | Talks (~$10K) | DevRel + platform community |
| Q2 2027 | AI-coding ecosystem events (Cursor/Anthropic dev days) | Sponsor (~$15K) | Agent-builder persona |

Plus dabl.club's existing event cadence (300+ AI events) as a continuous, low-cost distribution channel. Total event budget: **~$115K/yr.**

---

## 10. What opencolin delivers + engagement options

The substance below is identical across options; Tenki picks the container.

### Three engagement options

| Option | Shape | Anchor cost | When to pick |
|---|---|---|---|
| **A · Full-time, all-in** | Wind down the Nebius engagement (~60-day transition) and join as **Head of DevRel** | SF comp + equity | You want exclusive focus from day one |
| **B · 90-day paid pilot → convert** | Defined OKRs + exit criteria; **$35K/mo retainer + $50K program seed**, then convert | ~$155K for 90 days | You (and Luxor) want to de-risk before either commits — the two-way-door path |
| **C · Fractional with OKRs** | Quarterly engagement, ramps as Tenki scales | **$35–50K/mo** retainer + program budget | You want the operating system without a full headcount yet |

**Recommended given the situation:** **B → A.** The role is posted full-time and SF-based, but Tenki's culture *rewards* exactly this framing — *"results over process," "two-way-door decisions," "almost every product started as a replacement for an existing system."* The 90-day pilot **is** this plan's Phase 1, with hard metrics and a Day-90 convert decision. You're not asking them to bet on a hire; you're letting the funnel decide.

**Logistics handled up front (honesty over kindness):**
- **Location:** as a contractor (Option B/C) I work remote from [PT/ET] and come to SF for [cadence]; on conversion (A) we settle FT location. Don't make them ask.
- **Conflict disclosure:** I currently run developer demand-gen for **Nebius**, which overlaps with Luxor's compute business. I'd treat that as a conflict and **wind it down**, leaving them in a strong state. Flagged proactively so there are zero surprises.
- **dabl.club is an asset, not a distraction:** its 86k developers become an attributed acquisition channel feeding **Tenki-owned** community, benchmarks, and events.

### Always-on operating system (monthly retainer)
DevRel strategy + quarterly OKRs · content engine (writing, video) · community ops (Discord, office hours, newsletter) · benchmark publishing + reproducibility infra · cohort program end-to-end · hackathon/OSS program · conference talk machinery · analyst/press · brand/SEO · monthly metrics readout.

### Tooling deployed (opencolin)
*OpenClaw* (an AI DevRel agent for triage/outreach/routing) · *MetricsBoard* (public dashboard) · *Cohort Portal* · *Bounty/Champions Tracker*.

---

## 11. Resource model

### Headcount (12-month)
| Function | FTE | Annual |
|---|---|---|
| Head of DevRel (Colin — or fractional mo 1–3) | 1 | $220K (or retainer) |
| Developer Advocate (community + content) | 1 | $180K |
| DevRel Engineer (reference architectures) | 1 | $200K |
| Technical writer / video (contract) | 0.5 | $80K |
| **Subtotal** | **3.5** | **$680K** |

### Programs
| Program | Annual |
|---|---|
| Design Partner Cohort + Demo Day production | $50K |
| Agent Camp × Tenki + hackathons | $60K |
| Conference sponsorships + travel | $115K |
| Tenki-for-OSS credits + Champions rewards | $40K |
| Content production (video, podcast, paid) | $50K |
| **Subtotal** | **$315K** |

### Startup-friendly ramp (recommended)
| Phase | Spend | Headcount |
|---|---|---|
| Mo 1–3 (Pilot) | $35K/mo + $50K programs | Colin (fractional) |
| Mo 4–6 | convert Colin + $80K programs | + 1 DevRel hire |
| Mo 7–12 | full team + $185K programs | + DevRel Engineer |

**Year-1 startup-friendly total: ~$900K–1.0M** (vs ~$1.6M full-build). Cuts the cohort to 2 batches, halves conference budget.

---

## 12. Metrics dashboard (`tenki.cloud/metrics` or internal, weekly)

**Acquisition:** signups/wk · Discord members · newsletter subs · dabl.club-attributed referrals.
**Activation:** % configuring ≥1 product in 7 days · time-to-first-successful-run · % running ≥2 products at day 30 (North Star input).
**Engagement:** weekly active workspaces · community posts/wk · streams + benchmarks shipped.
**Retention:** 30/60/90-day workspace retention · products-per-workspace.
**Commercial:** trial→paid conversion · community→pipeline · Team/Enterprise accounts · compute-marketplace referrals.

---

## 13. Risks + mitigations

| Risk | Likelihood | Mitigation |
|---|---|---|
| Better-funded pure-plays (Blacksmith, Depot) outspend on brand | High | Win on cadence + honesty + the 3-product convergence story they can't match |
| GitHub cuts prices / ships its own AI review | High | Don't compete on price alone; lead with owned-compute economics + agent-era story |
| AI-reviewer precision gap erodes trust | High | Make precision the public roadmap; publish honest benchmarks (turn the weakness into the brand) |
| Crypto/Bitcoin-mining association deters some devs | Medium | Lead with the dev products + agent-era; keep the Luxor tie as an economics asset, not the headline |
| CodeRabbit/Cursor own AI-review mindshare | Medium | Reproducible head-to-heads + the "merge decisions not nitpicks" wedge |
| E2B/Modal/Daytona own sandbox mindshare | Medium | Co-market with the agent harnesses; "part of a whole loop" positioning |
| Tenki is unknown ("hadn't heard of it") | High | This is the whole DevRel mandate — brand from zero via benchmarks + community + dabl.club distribution |
| Pricing keeps shifting / margin compressed | Medium | Anchor content on value + economics, not a fragile $-savings headline |

---

## 14. Why Colin

A builder-DevRel with the playbook and the audience. Background: built **Cisco Meraki's developer program** (docs, webinars, partner ecosystem, the original playbook), **Head of DevRel at MetaMask** (multiple L2 launches), and founder of **dabl.club** — my own developer community of **86k subscribers (36–40% open rate)**, funded by Polygon & Coinbase, pivoted hard into **AI agents in early 2025** (Agent Camp, 300+ AI events).

Events I run: [BuilderShip](https://ship.builders/) · [ClawCamp](https://claw.camp) · [Self-Improving AI Agents Hackathon](https://sia-hackathon-showcase.vercel.app/) · [Automate Engineering](https://automate.engineering).

Tenki gets:
- A proven 7-initiative framework adapted to your wedge
- An **86k-developer distribution channel** on day one
- A builder-DevRel who shaped the developer program at Cisco Meraki and led DevRel at MetaMask
- Fractional cost until the function proves out — then convert
- The exact "build in public / public benchmarks / OSS PRs" motion the JD describes, already in motion (see `/teardown`)

---

## 15. Next steps

1. First-round screen (Estefy) → align on which engagement option fits.
2. Founder working session (Eddie / Hayssem / Nick) → sharpen to v2.
3. If pilot: sign 90-day pilot → Week 1 kickoff (instrument funnel, Discord + office hours, Benchmark Drop #1, Migration guides, Cohort #1 prep).
4. If full-time: confirm comp + equity, plan 60-day Nebius wind-down.
5. Quarterly review gate either way.

Contact: colin@lowenberg.com

---

*Appendix (in the accompanying site): full SWOT + vendor matrix, presentation deck, programs (cohort / OSS / hackathon), tutorials (Claude Code + Sandbox, migration), metrics dashboard.*