Hacker News new | ask | show | jobs
Show HN: Claude Code vs. Codex Global Usage Leaderboard (costhawk.ai)
14 points by optimizethis 36 days ago
4 comments

Great example of how AI can make something look decent, but when you're actually trying to grasp what it tries to communicate, you can really notice the lack of a human touch.
Not just a problem communicating but a lack of a problem here.

Why do I want to publish my team's ai usage with this random site? It's all risk and no upside.

It's an anonymous leaderboard, not sure you read the site: https://costhawk.ai/leaderboard
Tried it just now and really like the concept. It's a lot more than a private leaderboard. Going to onboard the rest of our team and setup a leaderboard for our next product sprint.
Especially when, if you’re an org that is heavy users of AI… you could just vibe code your own leaderboard and not have to pay monthly?

Presumably you will not be using this product long term, so why bother paying for something rather than just vibing a quick telemetry script and dashboard?

Most teams don't want to manage more platforms internally and have enough issues keeping up with their own product. At least that's what I'm hearing from CTOs. CEOs also don't want their CTO managing a bunch of internal tools, they want them delivering customer value.
Pay to publish your team's ai usage*
the human touch from human curation of the ai output, which tends to be sorely lacking.
i really would love to know where the data is coming from
founder here and wanted to clear a couple of things up:

by default the tool only reads usage telemetry that Claude Code, Codex, and Cursor already write locally to your own machine. token counts, dollar amounts, timestamps. no prompts, no code, no file contents.

if you're on a team, your org gets a private internal scoreboard scoped to your org. real names, broken down by team and project, visible only to people inside your own company. this is the actual product. an internal cost view for the person who owns the AI budget.

there is a separate public leaderboard for the broader community. pseudonyms like "CyberHawk", country flag, token count. no real names, no company, no team. on by default, one click to turn off in settings.

Thanks for the clarification, that makes a ton of sense. Love the ideea.
From you signing up and giving them your data
Good to hear. We doing something similar at https://thrumlabs.com/ but with rewards and private leaderboards
where that usage data coming from?
You sign up with them and they track your usage. Why would anyone want to do this?

It’s slop all the way down.

It’s not really for individuals… check the use cases tab.

It’s for enterprise to see who’s using AI and track token usage / spend.

Lots of individuals use it to track their AI spend. Everything is private and internal. Most companies track token usage anyways right now. It's a great tool, we use it across our entire team and love it. What's awesome is some "non-technical" users will surprise you with their output if given a chance.
So it’s heavily self selecting and pointless.
If you think tracking AI usage and connecting it back to outcomes and ROI is pointless, you are likely going to be without a job soon. See commentary from Cloudflare, ClickUp, Meta, etc.
If their tracking tool is open, why not?