Happy Friday Eve! In today's edition of the Leaderboard: a site to report companies that don't get back to you about job applications, a context-aware app that teaches you keyboard shortcuts, and a tool to fine-tune your LLMs. Let's dive in.
Hold em accountable
Ghostedd:Â Anonymously report companies that ghosted your job application.
Ever gone through multiple rounds of interviews only to never hear back from the employer? Ghostedd lets you call them out and warn others. It’s basically Glassdoor for jilted job applicants — the vindictive complement to the new LinkedIn #DesperateForWork banner. If the Ghostedd team can ensure strict verification of user posts, I can see the site taking off. The format is familiar and the content taps into some of our strongest human impulses (getting revenge and gossiping).
The OG productivity hack
Reiden AI:Â A context-aware keyboard shortcut copilot.
I’m a huge keyboard shortcut fan. In fact it’s a huge reason why I’m still hooked on the Superhuman craze. My only issue is that shortcuts aren’t usually universal so there’s sometimes a period of learning before you can fully become one with a new app. That’s why I like Reiden. It uses computer vision to help you learn keyboard shortcuts in real time. It’s context-aware so it knows what app you’re using and how you can save time.
Fine tune those models
LLMWare: Fine-tunes and deploys small language models privately or locally for enterprise.
LLMWare seems like a good niche bet in the game of the scaling laws of LLMs. While most startups chase bigger models for next-gen capabilities, there’s a sweet spot for smaller, more efficient models in simpler tasks. The important things, though? 1) Make sure you have really solid eval for your use cases. If not, stick with the larger models and cross your fingers. 2) Latency is a better reason for smaller models. Worry about cost savings after you’ve proven it actually works — accuracy is key to that. 3) Spend at least 2 days to write up some abstractions for your use case, before using ANY framework. (You’ll thank me later).
Share agents, not your machine

CoChat connects your local OpenClaw to a shared workspace so your team can run agents together, review outputs side-by-side, and iterate in real time — no SSH access to your laptop required.
Bring every instance into one hub: local OpenClaw, KiloClaw, multiple machines. Same agents, shared context, zero stepping on toes. Don’t want it running on personal devices? Deploy managed, containerized OpenClaw instances with real access controls.
Claude, GPT, Gemini — switch models mid-conversation and compare outputs in one thread.
Monday through Friday
Our ultra-fast Daily: Three takes on new products. Yesterday’s top ten launches. That’s it.








