Sam 💛
@samhh.com
Human of a few decades. Loves animals and functional programming and pastel colours.
https://samhh.com
https://sr.ht/~samhh
https://github.com/samhh
https://samhh.com
https://sr.ht/~samhh
https://github.com/samhh
Why can’t Thread devices seemingly ever be configured beyond the Matter spec without a proprietary hub (e.g. Aqara)? Shouldn’t it be viable on phones with Thread radios?
Also why don’t Nanoleaf (incl/ WiFi) provide Matter firmware updates via your hub?
Also why don’t Nanoleaf (incl/ WiFi) provide Matter firmware updates via your hub?
November 14, 2025 at 8:48 AM
Why can’t Thread devices seemingly ever be configured beyond the Matter spec without a proprietary hub (e.g. Aqara)? Shouldn’t it be viable on phones with Thread radios?
Also why don’t Nanoleaf (incl/ WiFi) provide Matter firmware updates via your hub?
Also why don’t Nanoleaf (incl/ WiFi) provide Matter firmware updates via your hub?
@shew.dev Is there a way in Turborepo to make a task ignore all arguments? An example use case is that I have unit tests that dependsOn some expensive codegen. The unit tests may be run with a file filter, but the codegen and its cache key should be unaffected.
November 12, 2025 at 11:38 AM
@shew.dev Is there a way in Turborepo to make a task ignore all arguments? An example use case is that I have unit tests that dependsOn some expensive codegen. The unit tests may be run with a file filter, but the codegen and its cache key should be unaffected.
Pluribus is the best thing on TV right now, watch it immediately.
November 8, 2025 at 4:14 PM
Pluribus is the best thing on TV right now, watch it immediately.
Are there any solutions for orchestrating ports in a monorepo?
I’d like to avoid hardcoding them since that always leads to conflicts, and I don’t want consumers to have to pick a bunch themselves when they clone.
Maybe sockets, pipes, a daemon? Ideally Turborepo could solve for this.
I’d like to avoid hardcoding them since that always leads to conflicts, and I don’t want consumers to have to pick a bunch themselves when they clone.
Maybe sockets, pipes, a daemon? Ideally Turborepo could solve for this.
November 7, 2025 at 8:36 AM
Are there any solutions for orchestrating ports in a monorepo?
I’d like to avoid hardcoding them since that always leads to conflicts, and I don’t want consumers to have to pick a bunch themselves when they clone.
Maybe sockets, pipes, a daemon? Ideally Turborepo could solve for this.
I’d like to avoid hardcoding them since that always leads to conflicts, and I don’t want consumers to have to pick a bunch themselves when they clone.
Maybe sockets, pipes, a daemon? Ideally Turborepo could solve for this.
Reposted by Sam 💛
I wrote a blog post about Flirt, a new code review tool I started working on:
blog.buenzli.dev/announcing-d...
blog.buenzli.dev/announcing-d...
Announcing Development on Flirt
blog.buenzli.dev
November 5, 2025 at 7:01 PM
I wrote a blog post about Flirt, a new code review tool I started working on:
blog.buenzli.dev/announcing-d...
blog.buenzli.dev/announcing-d...
Noticing that apologetics for ML/AI getting stuff wrong because a human conceivably might in a similar way is common. Next step out of the hype bubble will be to stop comparing the two because only one of these two actually “thinks”, and any overlap is coincidental.
October 21, 2025 at 7:47 AM
Noticing that apologetics for ML/AI getting stuff wrong because a human conceivably might in a similar way is common. Next step out of the hype bubble will be to stop comparing the two because only one of these two actually “thinks”, and any overlap is coincidental.
That moment when you’re frustrated and convinced there’s a bug in a library you’re using so you create a reduced test case, and… it’s working perfectly 🤨
October 17, 2025 at 7:39 PM
That moment when you’re frustrated and convinced there’s a bug in a library you’re using so you create a reduced test case, and… it’s working perfectly 🤨
Would like a BCC feature on GitHub for subscribing/notifying colleagues of issues I've raised or found that relate to something we've been pairing on.
I wonder, if the world used @tangled.org, if there'd be a cute way to piggyback on ATProto for niche extras like this.
I wonder, if the world used @tangled.org, if there'd be a cute way to piggyback on ATProto for niche extras like this.
October 9, 2025 at 2:56 PM
Would like a BCC feature on GitHub for subscribing/notifying colleagues of issues I've raised or found that relate to something we've been pairing on.
I wonder, if the world used @tangled.org, if there'd be a cute way to piggyback on ATProto for niche extras like this.
I wonder, if the world used @tangled.org, if there'd be a cute way to piggyback on ATProto for niche extras like this.
Is there any indication when @rspack.dev/Rsbuild might get RSC support? Notably absent from rsc.krasimirtsonev.com, the React Router docs suggest either Vite or Parcel instead, and of course alternatively there's Next's Turbopack.
React Server Components support
Can I use React Server Components (RSCs) today. A project that shows the support of RSCs across the frameworks and libraries.
rsc.krasimirtsonev.com
September 24, 2025 at 1:49 PM
Is there any indication when @rspack.dev/Rsbuild might get RSC support? Notably absent from rsc.krasimirtsonev.com, the React Router docs suggest either Vite or Parcel instead, and of course alternatively there's Next's Turbopack.
The broadly false promise of LLM productivity gains in software engineering reminds me of going deep on something like Vim. It feels more productive than it actually is.
September 21, 2025 at 5:04 PM
The broadly false promise of LLM productivity gains in software engineering reminds me of going deep on something like Vim. It feels more productive than it actually is.
Finally reading about React’s Activity API and it’s very promising. The docs do make me long for a stronger intuition for useEffect vs useLayoutEffect though.
September 20, 2025 at 8:01 AM
Finally reading about React’s Activity API and it’s very promising. The docs do make me long for a stronger intuition for useEffect vs useLayoutEffect though.
My Thread network issues have largely resolved since enabling IGMP Snooping. Some people will tell you to do the opposite. 🤷
September 20, 2025 at 7:26 AM
My Thread network issues have largely resolved since enabling IGMP Snooping. Some people will tell you to do the opposite. 🤷
AirPods Pro 3 fit my ears much better, I can finally shake my head (or merely grin) without them falling out! YMMV.
September 20, 2025 at 7:22 AM
AirPods Pro 3 fit my ears much better, I can finally shake my head (or merely grin) without them falling out! YMMV.
Patiently waiting for pnpm support in tsgo… any day now…
September 20, 2025 at 7:18 AM
Patiently waiting for pnpm support in tsgo… any day now…
iPhone Air with Apple’s networking chip has slightly worse Bluetooth range than my 15 Pro had, more like the M4 iPad Pro.
Every morning I get up and go make a coffee on the other side of the house with ANC (espresso machines can be loud) so this is a pretty consistent test.
Every morning I get up and go make a coffee on the other side of the house with ANC (espresso machines can be loud) so this is a pretty consistent test.
September 20, 2025 at 7:09 AM
iPhone Air with Apple’s networking chip has slightly worse Bluetooth range than my 15 Pro had, more like the M4 iPad Pro.
Every morning I get up and go make a coffee on the other side of the house with ANC (espresso machines can be loud) so this is a pretty consistent test.
Every morning I get up and go make a coffee on the other side of the house with ANC (espresso machines can be loud) so this is a pretty consistent test.
Did Apple’s operating systems always set beta to off in settings once a RC build goes stable? Don’t remember this convenience last year.
September 15, 2025 at 6:41 PM
Did Apple’s operating systems always set beta to off in settings once a RC build goes stable? Don’t remember this convenience last year.
I’ve noticed quite starkly since I stopped using different wildcard emails everywhere that I’m being tracked across the internet.
A search on one site on one device routinely leads to an ad somewhere else on another site for the same thing. This didn’t used to happen.
A search on one site on one device routinely leads to an ad somewhere else on another site for the same thing. This didn’t used to happen.
September 14, 2025 at 8:01 AM
I’ve noticed quite starkly since I stopped using different wildcard emails everywhere that I’m being tracked across the internet.
A search on one site on one device routinely leads to an ad somewhere else on another site for the same thing. This didn’t used to happen.
A search on one site on one device routinely leads to an ad somewhere else on another site for the same thing. This didn’t used to happen.
Ego Death at a Bachelorette Party 💖
August 28, 2025 at 8:03 PM
Ego Death at a Bachelorette Party 💖
My favourite least favourite thing to write in commit messages: “for some reason”
August 27, 2025 at 6:29 AM
My favourite least favourite thing to write in commit messages: “for some reason”
With Android requiring verification even for side loading it’s clear that technolibertarianism is finally dead.
We can either embrace a leftist solution in which we give power to the state, or that power will be left trillion dollar corps.
We can either embrace a leftist solution in which we give power to the state, or that power will be left trillion dollar corps.
August 26, 2025 at 7:52 AM
With Android requiring verification even for side loading it’s clear that technolibertarianism is finally dead.
We can either embrace a leftist solution in which we give power to the state, or that power will be left trillion dollar corps.
We can either embrace a leftist solution in which we give power to the state, or that power will be left trillion dollar corps.
Many of us are stuck with GitHub but pine for change-based reviews (jj!) they're not going to implement while they chase the AI unicorn.
Here's an idea: A UI layer, like Graphite, for review. This layer has a bot that tracks force pushes in a comment, so it can associate change-id headers. Viable?
Here's an idea: A UI layer, like Graphite, for review. This layer has a bot that tracks force pushes in a comment, so it can associate change-id headers. Viable?
August 12, 2025 at 1:08 PM
Many of us are stuck with GitHub but pine for change-based reviews (jj!) they're not going to implement while they chase the AI unicorn.
Here's an idea: A UI layer, like Graphite, for review. This layer has a bot that tracks force pushes in a comment, so it can associate change-id headers. Viable?
Here's an idea: A UI layer, like Graphite, for review. This layer has a bot that tracks force pushes in a comment, so it can associate change-id headers. Viable?
PHP gets a pipe operator, meanwhile the JS proposal is stuck in committee hell. 💀
August 5, 2025 at 9:10 AM
PHP gets a pipe operator, meanwhile the JS proposal is stuck in committee hell. 💀
Liquid Glass is like a continuous microstim with its pretty refractions everywhere.
Maybe a hot take but form for its own sake, and even to the mild detriment of function, has value too. This can also be applied for example to architecture.
Maybe a hot take but form for its own sake, and even to the mild detriment of function, has value too. This can also be applied for example to architecture.
July 28, 2025 at 7:30 PM
Liquid Glass is like a continuous microstim with its pretty refractions everywhere.
Maybe a hot take but form for its own sake, and even to the mild detriment of function, has value too. This can also be applied for example to architecture.
Maybe a hot take but form for its own sake, and even to the mild detriment of function, has value too. This can also be applied for example to architecture.
It’d be funny if the thing that finally forces GitHub to improve its code review experience is Microsoft’s attempt to force gen AI into everything, which - ignoring its merits or lack thereof - increases the importance of code review for vibed code.
July 19, 2025 at 7:07 PM
It’d be funny if the thing that finally forces GitHub to improve its code review experience is Microsoft’s attempt to force gen AI into everything, which - ignoring its merits or lack thereof - increases the importance of code review for vibed code.
Impossible challenge: design a chess app with an intuitive castling interface.
July 16, 2025 at 9:19 AM
Impossible challenge: design a chess app with an intuitive castling interface.