Michael Dwan
@michaeldwan.com
Software designer & engineer @replicate.com. Co-founded Fly.io.
I know right! I’m not ready, I’ll tell you that for free!
October 29, 2025 at 5:05 AM
I know right! I’m not ready, I’ll tell you that for free!
Wow congrats dude! Strong signal for jj. Cant wait to see what you build!
October 22, 2025 at 11:46 PM
Wow congrats dude! Strong signal for jj. Cant wait to see what you build!
that's amazing to hear 💖
October 20, 2025 at 4:17 PM
that's amazing to hear 💖
I'm digging it so far. Barely more friction than a TODO file in the repo, but a lot more powerful.
October 20, 2025 at 4:10 PM
I'm digging it so far. Barely more friction than a TODO file in the repo, but a lot more powerful.
Reminds me of a research paper from earlier this year (sorry, can't find the link) that said it was far easier for LLMs to solve complex math problems by evaluating adhoc python than it was to teach them math directly
September 26, 2025 at 9:25 PM
Reminds me of a research paper from earlier this year (sorry, can't find the link) that said it was far easier for LLMs to solve complex math problems by evaluating adhoc python than it was to teach them math directly
Be the change the world needs, Amos. Be the change!
September 11, 2025 at 10:21 PM
Be the change the world needs, Amos. Be the change!
I feel this. As soon as I hit the character limit I say fuck it and go do something else instead.
July 23, 2025 at 10:19 PM
I feel this. As soon as I hit the character limit I say fuck it and go do something else instead.
The Claude Code extension for VS Code/Cursor is an acceptable middle ground. Prompts to kick it off, diffs to follow along & steer it. Still a long ways off from a "good" dx though.
July 23, 2025 at 10:16 PM
The Claude Code extension for VS Code/Cursor is an acceptable middle ground. Prompts to kick it off, diffs to follow along & steer it. Still a long ways off from a "good" dx though.
The Spiderman pointing meme, but with Pac-Mans
June 6, 2025 at 3:21 AM
The Spiderman pointing meme, but with Pac-Mans
I landed somewhere similar after testing AI code review bots. I wish giving humans super powers was as lucrative as trying to replace them.
bsky.app/profile/mich...
bsky.app/profile/mich...
What I actually want is for humans to review code with less effort. LLMs can help with this! We don’t want summaries, we want context. A diff shows what changed—help us understand why. Highlight what matters. Explain intent. Code reviews should be a conversation.
Been trying out Copilot PR reviews. After ~10, it found a couple typos and one bug—not bad. But calling that a review is disingenuous. Static analysis + engaged humans nailed this ages ago, it was just time-consuming. I wish AI aimed to make good things stupidly easy—not replace them with worse.
May 8, 2025 at 4:27 PM
I landed somewhere similar after testing AI code review bots. I wish giving humans super powers was as lucrative as trying to replace them.
bsky.app/profile/mich...
bsky.app/profile/mich...