IdeaWell

Project ideas from Hacker News discussions.

GitHub Copilot code review will start consuming GitHub Actions minutes

Original Article

Hacker News Discussion

📝 Discussion Summary (Click to expand)

1️⃣ AI inference is being subsidized and will be priced higher

“Expect to see more of these kinds of announcements as companies need to start showing returns on their AI investments. It’s hard to say how subsidized the current AI products are… but we’re definitely getting a free lunch at VC’s expense the moment.” — AlexandrB

2️⃣ GitHub Copilot code‑review uses Actions minutes + token credits, creating double‑billing worries

“I’m assuming the running of the model is consuming the tokens, and the client coordinating and orchestrating the calls to the model to perform the review is happening in an action runner, thus using action minutes.” — zdragnar

3️⃣ Vendors are moving toward price hikes – a “rug‑pull” or bait‑and‑switch

“Just you wait, this is a gentle tug to test how hard they can pull when the time comes.” — 2ndorderthought

4️⃣ Open‑source models provide a real competitive moat and can be profitably run without heavy subsidies

“I don’t know about highly since they have no moat even more than Anthropic and OpenAI have no moat. Anyone with a few hundred thousand dollars or sufficient free GPUs can compete with them. So running an open model should earn a market‑rate margin.” — semiquaver

🚀 Project Ideas

TokenWatcher

Summary

  • Browser extension + CLI that monitors real‑time token consumption and cost across GitHub Copilot, Claude, OpenAI, and Anthropic APIs.
  • Shows per‑action pricing, budget alerts, and cumulative spend to prevent surprise bills.

Details

Key Value
Target Audience Individual developers and small teams using paid AI coding assistants.
Core Feature Real‑time cost overlay, budget caps, and per‑PR cost breakdown.
Tech Stack Chrome/Firefox extension (TypeScript), Node.js CLI, JSON API aggregator, React dashboard.
Difficulty Medium
Monetization Revenue-ready: subscription tier “$5/mo per user” for premium alerts and historical analytics.

Notes

  • HN users repeatedly lament hidden token costs (“Fck you. Pay me.”) and double‑billing concerns.
  • Could integrate directly with GitHub Actions logs to surface hidden AI‑credit usage, addressing the “double billing” frustration.

LocalInferenceCI

Summary

  • Open‑source Docker‑based CI runner that executes LLM inference locally on cheap GPU/CPU hardware, eliminating reliance on GitHub Actions minutes and external API fees.
  • Provides a self‑hosted “code‑review bot” that runs inside the repo’s CI pipeline.

Details

Key Value
Target Audience Solo developers, open‑source maintainers, and small teams seeking affordable code‑review automation.
Core Feature One‑click deployment of a local LLM (e.g., Ollama, GPT‑Q) as a CI job that reviews PRs and returns comments.
Tech Stack Docker Compose, Python FastAPI, Ollama + ggml‑based models, GitHub Webhook integration.
Difficulty High
Monetization Revenue-ready: “$9/mo per hosted runner instance” with optional managed hosting.

Notes

  • Commenters note “running for x minutes” while reviews happen and desire cheaper alternatives to GitHub Actions (“If only there was a place that was naturally cold…”).
  • Could be bundled with self‑hosted runners on old gaming rigs, directly answering the “self‑host cheap CI” demand.

CopilotCostBreakdown#Summary

  • Web dashboard that ingests GitHub Actions logs and token usage data to automatically compute the exact monetary cost of each Copilot code‑review action, exposing the double‑billing of minutes + AI credits.
  • Generates exportable reports for auditors and budget planners.

Details

Key Value
Target Audience Enterprise users and finance‑oriented dev teams using GitHub Copilot for code review.
Core Feature Automated cost calculation per PR, visual heat‑maps of spending, and alerts when thresholds are crossed.
Tech Stack Python backend, PostgreSQL, Grafana‑style front‑end, GitHub API authentication.
Difficulty Medium
Monetization Revenue-ready: “$2/user/mo for advanced analytics package”.

Notes

  • HN threads repeatedly ask “Why would anyone knowingly pay twice?” and discuss the rug‑pull nature of token pricing.
  • Directly surfaces the hidden cost that users currently have to manually calculate, satisfying demand for transparency.

LambdaLLM Marketplace

Summary

  • Centralized marketplace that aggregates inference endpoints from multiple AI providers (OpenAI, Anthropic, Cohere, open‑weight hosts) and automatically selects the cheapest suitable model for a given request.
  • Offers programmable pricing tiers and SLA guarantees.

Details

Key Value
Target Audience Developers building AI‑enhanced apps who want predictable, low‑cost token usage without vendor lock‑in.
Core Feature Dynamic routing / fallback logic, cost‑based routing, and usage‑based billing API.
Tech Stack Node.js serverless functions, Redis caching, multi‑provider adapters, GraphQL gateway.
Difficulty High
Monetization Revenue-ready: “$0.001 per 1k tokens + 5% platform fee”.

Notes

  • Multiple comments highlight “subsidies won’t last” and the need for “real prices” (“When subsidies stop they’ll jack up prices”). - Provides the market mechanism users crave to avoid being locked into a single provider’s escalating costs.

Read Later