I make open source dev tools at entire.io like mise, aube, fnox, hk, mbx, pitchfork. Previously I wrote the heroku cli and oclif. ex-amazon/meta.
Dallas, TX
- Tweets 1,750
- Following 1,267
- Followers 6,847
- Likes 1,480
this was awesome @dhh
Here's my conversation with @DHH. He is back for round 2! It was an epic fun 5 hour conversation about the future of programming, AI, Linux, and human civilization.
It's here on X in full and is up everywhere else (see comment).
Timestamps:
0:00 - Episode highlight
1:27 - Introduction
2:56 - Programming with AI agents
18:14 - How software will change
27:30 - AI impact on open source
37:21 - Building Omarchy Linux distro
47:05 - Vibe coding vs agentic engineering
1:00:06 - The end of manual programming
1:10:24 - Advice for programmers
1:22:30 - Surviving Internet Hate
1:31:46 - Programming setup for AI Agents
1:44:11 - Obsessing about speed
2:07:06 - Voice prompting vs typing
2:21:05 - Best AI coding models
2:37:55 - Best AI coding harnesses
2:50:57 - AI video generation and filmmaking
3:10:28 - Fatherhood
3:38:35 - Linux will win the desktop
3:49:51 - PewDiePie
3:59:24 - Future of programming
4:22:17 - Politics and immigration
4:53:54 - Longevity, over-optimization, and fear of death
5:05:38 - Eternal recurrence and future of human civization
13
19
499
47,775
a snappy screensaver absolutely makes an operating system better.
preview should be fast. it should use minimal battery. it should close *extremely* quickly.
So you Omarchy nerds got millions USD of donations and work on optimizing the screensaver and enthusiastically join the propganda of Anthropic?
And isn't this a sign of being a puppet of AI model vendors?
This ain't making Linux Desktop great.
29
2,702
jdx retweeted
🔴 Live today at 4pm CET with @jdxcode, creator of mise 🛠️
piped.video/watch?v=aPg44O9i…
1
2
1,767
one of the problems with rust dev that mr-boxington fixes (or at least mitigates) is a bunch of agents running and hitting OOM while also permitting cargo builds to run in parallel (even in the same worktree). it has a scheduler that hands out permits to keep cpu/memory under control.
it's not perfect and you can definitely still hit OOM issues. one reason for this is it monitors cargo processes but your agents may be running tests or doing other work that is outside of its scheduler.
in the next release I'm adding optional support for `cargo test` executions to run within the scheduler. it does this by running rusage and recording how much cpu/memory the tests take.
anyhow, if you use mr-boxington, definitely send some ideas on how we could continue to improve this.
3
2
28
3,240
jdx retweeted
Mr. Boxington by @jdxcode has saved the day.
Before using it my machine was getting CRUSHED having multiple agents writing Rust. I would get alerts constantly of the HDD being full. Problem solved!
1
1
3
465
jdx retweeted
btw, thank you @jdxcode for mbx, it brought my compile times down ~40%
the rust compiler is too slow for the 1000 tok/sec models
the model will spend 95% of development time just waiting for the compiler!
1
1
6
2,236
mise.lock now automatically generates aube/uv lockfiles as external files:
[[tools."pypi:black"]]
version = "24.10.0"
backend = "pypi:black"
uv = { path = ".mise/locks/pypi-black/24.10.0", digest = "sha256:…" }
this not only locks the full dependency tree but i'm using the idiomatic lockfile here so that security scanners don't need to understand some new toml syntax for all the transitives, they can just read aube-lock.yml/uv.lock like normal
(aube-lock.yml is just a renamed pnpm-lock.yml btw)
3
28
2,599
another boneheaded "perf doesn't matter" take.
not only have I made a career out of building package managers that run far faster than this, if things like `mise install` regress to even 100ms I hear complaints from my users. Happened 3 days ago in fact.
also the fact that mise can be so fast means that things like `mise x` are even possible. Nobody would use it if it added 500ms of overhead.
No one cares. Anything below 500 ms for a package manager is decent and if you think you need something faster than that, you are probably lying to yourself.
5
79
7,244