Chopped Unc. Emo Millenial. Built flipt.io. Now AI @cloudflare. Latest projects: 🌊 covereader.app/ 🤖 vigilmac.com

NC, USA
Stanford dropped CS329Z today. Engineering AI Agents. compound systems, build RAG/tool use from scratch, then frameworks, then how you actually evaluate this stuff. gonna follow along online. looks pretty neat. cs329z.stanford.edu/
13
185
1,475
82,445
FYI tweeting is optional
lmk if you want a video demo
58
Stanford dropped CS329Z today. Engineering AI Agents. compound systems, build RAG/tool use from scratch, then frameworks, then how you actually evaluate this stuff. gonna follow along online. looks pretty neat. cs329z.stanford.edu/
13
185
1,475
82,445
ill post notes in this thread as i work through the lectures. follow along if youre taking it too. I asked Hermes to build a study plan between this course and CS336 piped.video/playlist?list=PL… if anyone wants it I can put it on GitHub
2
1
13
4,423
smh. almost said 'load-bearing' in real conversation today
2
147
turned hashimoto's whiteboard defense into /whiteboard nitter.net/mitchellh/status/21002… 1. map the repo into docs/map with evidence on every “why” 2. quiz you one question at a time 3. never invents rationale, never edits your code 👇
The "whiteboard defense:" I should be able to pull you aside at any moment and ask you to explain any customer-facing system you've shipped. You should be able to clearly explain how it works and defend the decisions you made. This is my benchmark for responsible AI usage. I don't expect line-level familiarity with the code. I don't care if you remember the exact function name or implementation detail. You may not even know it. I don't care. But if I ask "why did you do X instead of Y?", "what happens if this actor behaves maliciously?", "what data structure did you use here and why?", or "where does this fail?" you should be able to answer confidently. For PoCs, demos, experiments, whatever: I don't care. Generate 100% of it and understand none of it. Speed over quality every time in those specific scenarios. But if you're shipping customer-facing work, you can't be shipping things you don't understand at a high level.
1
1
1,737
thinking about creating a Youtube channel called ‘AI for boomers’ so our parents don’t get scammed
1
180
Legitimately how is Opus 5.5 so fast?
1
220
Vigil just went live on Tiny Startups 🚀 appreciate the listing @ratheejaisal Keeps your Mac awake while Claude Code, Codex, Cursor and other coding agents work, then lets it sleep when they stop. tinystartups.com/startup/vig…
1
3
4
1,271
shipped Vigil. Mac menu bar app that keeps your Mac awake while coding agents work, then lets it sleep when they finish, fail, or wait too long. not “caffeinate and forget.” it watches the agents work. Claude Code, Codex, Cursor, pi, etc $3.99 once vigilmac.com/?utm_source=x&u…
5
5
207
gave my AI agents a door into Cove, my local Mac RSS reader. no chat pane in the app. just MCP on loopback: search the archive, list unread, add a feed from a pasted URL. wrote it up: markphelps.me/posts/giving-a…
1
138
attention is still the bottleneck — not how many agents i can spin up. i can keep three projects open at once. still better for me to pick one and ship it. parallel work feels busy; finishing one thing is what actually moves the needle.
1
2
141
1. one project gets the full attention stack 2. the others wait in a queue, not half-alive tabs 3. ship, then pull the next agents multiply hands. they dont multiply focus.
33
wake up babe. other Jared just dropped Jev Open Source
Kev-0.5B: A tiny open source Jev-like decision model with a TypeSafe-compatible API based on Qwen2.5-0.5B that you can train and run on a MacBook Pro. Model card and weights are available on GitHub github.com/jaredpalmer/kev
1
205