I don't see any evidence that it can build a business

what's important is where big enterprise point their budgets. and there we agree, it's not to chinese models (as hopeful as I am that we can avoid a duopoly and great local models)

on the other hand if it's just a local coding/"use my computer" agent, i highly doubt the effort in maintaining different prompts is worth any gain in performance

edit: which I doubt btw. investors are going to want to dump the bags very soon

So I rebuilt the board I was used to. Icebox, backlog, current iteration, done. Stories are 1, 2, 4 or 8 points, iterations fill to your velocity, and a story goes started, finished, delivered, then accepted or rejected.

I love the simple opinionated board and tried to recreate it. Have been using it for a year and thought why not open source it.

Hosting isn't ideal for open source, but it's my cheap/free stack. Cloudflare Workers + D1, free tier is enough. All MIT, hosted at tracker.usero.io if you don't want to deploy anything.

There's an MCP server too, that's how I point Claude Code at it.

Comments URL: https://news.ycombinator.com/item?id=49676874

Points: 1

# Comments: 1

Unless it's against the terms of service of the private corporation

Why not? They can already take an uber. It's questionable which is safer

See this is only an issue because that is a plausible situation. The actual problem is on the police/cultural side, not the waymo side. No one should even have to think about this when calling the cops. And I can't think of 1 other western society where they do

The problem is, orgs who see themselves as big data orgs want to act that way, even if they're medium data. "But we'll need it when we grow", "we need to know the state of the art tools"

To-date there are 36 tools (signup, clients, feedback, clusters, forms and theming, user tests, AI test runs). E.g. `list_clusters` returns clusters biggest first with up to three verbatim user quotes each. `request_ai_pr` has Usero write and open the pull request server-side on the connected repo.

I used to hate MCP because of the bloat it gave your context, but since 2.1 Claude Code defers MCP tools, so the model sees tool names only until it loads one.

Here's a full example transcript, with all tool calls: https://usero.io/blog/feedback-mcp-server#full-transcript

Pricing: free tier is $0 with 5 AI PRs a month, Pro is $39 for 50, Team is $89 for 200. MCP is on every tier. Blog post: https://usero.io/blog/feedback-mcp-server.

I'm building this because I think as building features gets easier, it becomes even more important for startups to listen to their users and focus on high-quality product decisions. My goal is making those product decisions as easy as possible, helping companies be UserObsessed.

I'd love to hear your thoughts.

Comments URL: https://news.ycombinator.com/item?id=49658898

Points: 1

# Comments: 1