KC

CEO, co-founder @fraudmarc @Techstars '17

[KC @KColemanGT]

crypto hw wallets & secure enclaves for mailboxes so private they make Protonmail look like Gmail. our newest project https://t.co/byEkJ3clJV

[KC @KColemanGT]

@leaveintosh @pilcrowonpaper You might like https://t.co/08jb7IIe3I. Gives you a url that your distro's `clevis` binds to a LUKS keyslot. Then your server's IP can automatically unlock to boot. No special sw on your host & no keys leave your host.

I'm one of the maintainers.

[KC @KColemanGT]

PSA: Hetzner servers aren't encrypted

Here's how you can finally deploy automatically encrypted Hetzner servers

Plus boot without dropbear using your distro's built-in functionality

[KC @KColemanGT]

1. Government reps at internet working groups
2. Redditors

Have you ever met an anti-encryption person who isn't in one or more of these groups?

[KC @KColemanGT]

PSA: Hetzner servers aren't encrypted

Here's how you can finally deploy automatically encrypted Hetzner servers

Plus boot without dropbear using your distro's built-in functionality

[KC @KColemanGT]

@NickAbraham12 We should chat about smarter sending

[KC @KColemanGT]

@bee_human_ @MatthewBerman @cursor_ai Cursor sends different prompts to different models

2.5 doesn't work directly in agent mode. Fails to edit files, etc. Alias o3-mini to call 2.5 with the stronger agent flow prompt and it comes to life

[KC @KColemanGT]

@cursor_ai agent finally gets @GroqInc speed in llm-router v0.2.0

https://t.co/ZSwdWfjtBY

[KC @KColemanGT]

Just got @cursor_ai agent mode running locally on @ollama 💻rip🪫

[KC @KColemanGT]

@alexellisuk @cursor_ai @ollama @AIatMeta @OpenAI @inletsdev New version has improved API key handling and easier setup.

Cursor randomly tests connections with your selected models, so uncheck any models not supported by the providers configured in your config.json.

https://t.co/ReP6Y1YCNN

[KC @KColemanGT]

built LLM-router for @cursor_ai to support more LLM backends, including locally served models.

switching between @ollama phi3, @AIatMeta llama3, and @OpenAI

open source it?

[KC @KColemanGT]

@cursor_ai @ollama @AIatMeta @OpenAI https://t.co/ReP6Y1ZaDl

[KC @KColemanGT]

@__tosh @ollama @continuedev @SourcegraphCody @sourcegraph I made llm-router to use local ollama models w/ cursor. ollama/phi3 is impressive for ⌘+K smart edits.

https://t.co/ReP6Y1YCNN

[KC @KColemanGT]

@ollama Nice. This thing cooks in @cursor_ai with llm-router

https://t.co/ReP6Y1YCNN

[KC @KColemanGT]

@__tosh @ollama @continuedev @SourcegraphCody @sourcegraph did you stick with either of those or cursor?

[KC @KColemanGT]

@mathdroid @cursor_ai ollama + cursor + llm-router if you want to switch between models faster

[KC @KColemanGT]

@ivanfioravanti @DrTBehrens @erhartford llm-router brings ollama into Cursor for local inference

[KC @KColemanGT]

@StormSoulTV @fayazara Add local @ollama models to your @cursor_ai rotation with llm-router. It's a bit like openrouter but you run locally. Just tagged the first release a moment ago.

https://t.co/ReP6Y1YCNN

[KC @KColemanGT]

Surprising benchmark results from the newest gpt 4 turbo. Nice work, @paulgauthier
https://t.co/lpQvwffpZ7