physicles

This piqued my curiosity. Are you the author of the paper?

For this principle to become widely known, it needs to be communicated in a more succinct way. 400 pages is too much to ask people to invest.

Even so, as far as I understand the gist of what the paper says, the principle helps make explicit some intuition I’ve had about design for a long time.

I’m not one for conspiracy theories, but since SpaceX is the only launch services provider that could actually put one of these in orbit, this smells a lot like hyperloop to me — an unserious proposal that serves as a distraction and furthers Musk’s aims, and benefits anyone who can get close enough to the piles of cash that VCs will drop on this.

You know what’s easier and cheaper than putting a data center in space? Putting one literally anywhere else other than space.

> sometimes the people you are working for are incapable of reasoning about planning artefacts or understanding how the system will look or operate simply from a document

I’m wrestling with this now. Over my career I’ve seen a strong correlation between good writers and good software engineers, but not everyone fits this mold. Shorter cycles and more chances for communication and feedback are helpful here.

I’m currently fighting the “don’t use Gemini to write internal documents” war at my company. It’ll be long and hard, but I think I’ll eventually prevail.

Every time someone throws a document written by AI at me, it feels so disrespectful.

> "The single biggest problem in communication is the illusion that it has taken place."

Love this! Corollary: when you have too many meetings, that’s easy to notice. When you don’t have enough meetings, that’s harder to notice.

I’m in the process of carefully adding meetings and process to our small team of 6 (we had a PM from a large company drop in a few years ago and haphazardly add a bunch of process, and it didn’t really help).

We’re fully remote and have a daily huddle and, on average, 1 hour of meetings a week. It turns out this isn’t enough. So far, each bit of communication we’ve added has resulted in better outcomes and higher morale because we feel more like a team.

I agree that the days when “everyone” watched the same show are done. But if you can find a small group to watch a show with (better in person), then there are better shows available for that experience these last several years, even if the average quality has gone down.

What are some of your favorite shared experiences to replace tv?

We’re in the middle of this right now. Go makes this easier: there’s a go CLI command that you can use to list a package’s dependencies, which can be cross-referenced with recent git changes. (duplicating the dependency graph in another build tool is a non-starter for me) But there are corner cases that we’re currently working through.

This, and if you want build + deploy that’s faster than doing it manually from your dev machine, you pay $$$ for either something like Depot, or a beefy VM to host CI.

A bit more work on those dependency corner cases, along with an auto-sleeping VM, should let us achieve nirvana. But it’s not like we have a lot of spare time on our small team.

Yep, I have a cron that does git add . && git commit -m “daily commit”. Haven’t touched it in a couple years.

Piggybacking on this… do all nvidia cards have the same issue with Linux drivers, where the fan won’t ever go below 30%? I have a 3090 on Ubuntu 24 and hours of googling netted nothing that worked.

Yeah this happened to me at least once, and I had to spend several hours with low-level recovery software to get my files back. This was the catalyst that finally got me to ditch windows for good 7 years ago.

Linux grew its share by 14% in one month? Amazing.

Which frame do you use? I’ve been looking for this for a long time, but I wasn’t aware color e-ink was good enough for this yet.

Agreed, and Carmack as always was ahead of the curve here.

In case anyone here hasn’t seen it, here’s his famous essay, Functional Programming in C++:

http://sevangelatos.com/john-carmack-on/

He addresses your point:

> I do believe that there is real value in pursuing functional programming, but it would be irresponsible to exhort everyone to abandon their C++ compilers and start coding in Lisp, Haskell, or, to be blunt, any other fringe language. To the eternal chagrin of language designers, there are plenty of externalities that can overwhelm the benefits of a language, and game development has more than most fields.

I wrote some tooling to help debug sql queries with many CTEs. It parses the sql, finds all the CTEs, and prints the result of each CTE formatted as csv. If the .sql file changes on disk, it reruns the query and tells you which CTEs’ output changed. Saved me hours in debugging.

Why do you need a warm standby for production? Do you need >= 3 nines?

Our staging environment has its own instance that is rebuilt from prod, with pii removed, every day outside working hours (this normally takes about 15 minutes). It’s fantastic for testing migrations, and is easy to support compared with a warm standby.

What do you propose instead?

Interesting idea! But the PC it’s attached to is a mini PC, so I don’t have motherboard access.

Windows as a platform lost the war many years ago when the web (and then mobile) took over, and is no longer a key part of Microsoft’s business since the CEO switch in 2014.

Microsoft’s current strategy is to 1) keep a stranglehold on enterprise (so, Office and 365 subscriptions) and 2) make money on workloads regardless of platform (with Azure).

What new software, besides games, targets windows anymore? If you wanted to target windows, what SDK and language would you build on, and how many times will it be replaced with $new_hotness in the next 10 years?

I have a similar experience with my GPU. Got a 3090 and found that nvidia’s linux driver enforces a minimum GPU fan speed of 30% regardless of temperature. That’s unacceptably loud if you’re in the same room. After hours of searching online, flashing the GPU’s bios is apparently the only solution, so I bought a bunch of acoustic foam instead.

I’m still 100% Linux on all devices, but this bit really sucks.

Don’t rule them out.

I was in the same boat, doing most text navigation on the keyboard and fully adapted to the quirky arrangement of home/end/pgup/pgdn of the Microsoft natural keyboard elite I’d been coding on for 23 years.

But I do a lot of nomading, so I really wanted a small keyboard. And I’m a tall guy, so I was looking for a full split (I tried the Kinesis Freestyle2 for a while but didn’t love it).

I bought a ZSA Voyager. The first day I went from ~120wpm to 20. I hated it and had buyer’s remorse. Second week, I was still switching to another keyboard in the afternoons because I was making so many typos and felt so tired. I swapped out most of the switches with higher force ones, and that helped a ton. It took about two months before the keyboard disappeared again and there was only me and the work.

The biggest single win is that backspace is under my right thumb next to space bar, which felt life-changing once I got used to it (if I were still using a normal keyboard, I’d probably re-map right alt to backspace). And while word selection with ctrl+shift+arrow took a bit longer to get used to, now I appreciate not having to move my fingers at all to hit the arrow keys.

Nine months later, you can pry the Voyager from my cold dead hands. It’s probably the last keyboard I’ll ever buy. I do wish it had two more keys (haven’t found good homes for Alt and ~) but overall I’m very happy with where I ended up.

Could you give some specifics? For example, does this look like repeatedly trying things that make you a bit uncomfortable, and CBTing when you have those uncomfortable feelings? Is it helpful to do this with people you already feel safe around?

This is one of the reasons why I’m getting familiar with self-hosting. Local models are improving shockingly fast. I use Gemma3 27B for generating summaries of podcast transcripts, for instance.

Ironically, the hardest level for me was selecting which faces are AI-generated

I bought a Voyager about a year ago. On day 1 I nearly had buyer’s remorse because I topped out at about 20wpm, even though all my keyboards have been ergo since 2001.

After more practice, and after swapping out the key caps, I now think this is possibly the last keyboard I’ll ever buy. I’m having a hard time imagining anything better (though I could use maybe 1 extra thumb key on each side).

Love seeing meaningful stdlib improvements.

I just ran our full suite of a few thousand unit tests with GOEXPERIMENT=jsonv2 and they all passed. (well, one test failed because an error message was changed, but that's on us)

I'm especially a fan of breaking out the syntactic part into into its own jsontext package. It makes a ton of sense, and I could see us implementing a couple parsers on top of that to get better performance where it really matters.

I wish they would take this chance to ditch omitempty in favor of just the newly-added omitzero (which is customizable with IsZero()), to which we'll be switching all our code over Real Soon Now. The two tags are so similar that it takes effort to decide between them.

Is that intended to be a good example? There's still tons of duplication between the environments.

Kustomize eliminates the vast majority of the duplication (i.e. a unique fact about the cluster being expressed in more than one place), it's just the boilerplate that's annoying.

For toy projects, sure. For production, the probability that the LLM would sneak in some subtle typo is just too high.

We have about 30 services running in 4 environments, including dev. I desperately want a better kustomize that removes most of the boilerplate and adds linting (like, every process should have a ram limit, but no cpu limit). I estimate about 75% of the lines of YAML are redundant.

I had this problem on the latest Pop OS LTS (I know, I know). Took about 4 hours to find the magic incantation that would let me install podman >= 4 without breaking anything else.

Could you describe what made those tests effective? I just wrote some tools to write concise tests for some analytics queries, and some principles I stumbled on are:

- input data should be pseudorandom, so the chance of a test being “accidentally correct” is minimized

- you need a way to verify only part of the result set. Or, at the very least, a way to write tests so that if you add a column to the result set, your test doesn’t automatically break

In addition, I added CSV exports so you can verify the results by hand, and hot-reload for queries with CTEs — if you change a .sql file then it will immediately rerun each CTE incrementally and show you which ones’ output changed.