viceconsole

That's why I tend to prefer services that are single go binaries, for example.

For complex stuff like Immich, I'll use podman. But these days with uv, even managing python services with systemd is fine.

I use both Restic and Syncthing. I have considered putting encrypted backups on a cloud blob storage, but my preference would be to set up an offsite backup at a relative's house.

Fair, I just never use the web clients. I mainly use the Android phone or TV apps, which do support H.265. The Android TV subtitle support seemed ok, I think I did have issues with the Roku app and subtitles.

They might have to turn on/off some lights using the buttons on the smart switch, yes. Or maybe lose access to Jellyfin for a while. But on whole things have been pretty reliable.

Yes.

I run Adguard on my router, not on my server. I did write a subsequent post about the rest of the stuff I run on my router and my Proxmox cluster, but forgot to mention Adguard.

I considered KeePass+ Synching but tended to run into the same issues as with Owntracks - Android is hostile to always on background apps. Then when it came time to onboard my spouse, KeePass wasn't an option.

Do you not consider systemd services to be declarative/standard?

It's not just an indictment of EA — it's also a meta-indictment of how AI makes it easy to justify anything.

Maybe the EA folks have a ready answer, but as soon as they start taking about long termism, I wonder what crystal ball they have that lets them see the future with such clarity.

Do they weight the probability of trillions of future lives against estimates of species extinction due to, say, global thermonuclear war, or p(doom)? Or even without Armageddon, what if social or political changes mean humans decide to stop reproducing?

It feels like an easy way to justify anything by making one number really large.

My favorite wrong way is when someone completely covers a small item (ramekin, espresso cup) with a larger item (bowl, food storage container).

I do a ton of DIY work, from vehicle maintenance (eg valve adjustment) to plumbing, plastering, insulation, ventilation, carpentry, electrical, irrigation, etc.

"Hit that pipe with precise amount of force" is a good example. That works on a shop floor where there's room to operate or set up a jig. How is it going to work in tight spaces that aren't generalizable?

> This is exactly the part I’m afraid of. The political response of the displaced labor (which may or may not be violent / revolutionary), and the subsequent counter-response of capital… potentially using the very robots that displaced the jobs to begin with.

I get the fear, but at that point we're basically imagining that Terminator is a documentary. Given that the US has more civilian guns than people, and that private ownership of military grade weaponry is strictly controlled, I don't think street wars between private robot armies and folks in the bread line is a realistic outcome.

I think the "humanoid robot replacing all blue collar jobs" fear understates the amount of real world (i.e. visual, physical, sensory) data and tacit knowledge involved in many such jobs.

Driving shows it can be done, but even there it's taken years of highly directed research and investment, and driving (at least where Waymo operates) is still pretty well defined compared to plumber.

Such data can be gathered but the current methods of RLHF by paying white collar experts, accessing private datasets, etc. don't seem to really get at this.

I'm also continually surprised that people seem to think politics doesn't exist anymore. I find it exceedingly unlikely that massive amounts of people will simply accept this purported widespread unemployment without major political consequences.

Personality and habit. For me, being a graduate student for 7 years (masters then PhD) forced me to live simply. Those habits stuck with me through the next decade of good salaries. My spending went up from graduate student poverty, but I still saved and invested 50% to 70% of gross income. Now I probably have more than a lot of peers who started working well before me, but didn't save as much.

To your point, pay yourself first. Set up automatic investments. Find a fiduciary advisor if you must, but it's really not rocket science.

Netherlands, not Belgium. Though with the quality of US targeting, it's possible they might accidentally invade Belgium instead.

Wireguard is easier to set up and add clients so I did that first, but it does not work over my phone's mobile data connection due to carrier filtering, whereas OpenVPN does. Also doesn't hurt to have two options.

Probably half run natively, and half on podman as systemd quadlets.

I specifically look for efficient projects when I have a choice. Also most of my video media is in 720p. The 8 IP cameras I run in lower resolution as well, though that's mainly to avoid network bandwidth issues.

Honestly I have not run into any memory issues.

Beelink mini PC, I think S12 with N100 and 16 GB RAM, purchased before the rampocalypse...

Immich, Jellyfin, Gonic, Snapcast, Slskd, Audiobookshelf, Transmission, Automatic Ripping Machine, Vaultwarden, Frigate, Home Assistant, Actual Budget, Pocketbase, Vikunja, Forgejo, Readeck, Backrest/Restic, Gotify, Uptime Kuma, Homepage, Caddy, Shared network folder with spouse

I also run Cockpit and Beszel but don't use them often.

Router runs Adguard, Wireguard, and OpenVPN.

I still rip DVDs, a few a month. No streaming services in our household, so DVDs are how we get multilingual versions of things.

And I can rip the special features.

Firefox reader mode displays the full article perfectly.

Academics' primary external recompense is in credit, attribution, and recognition. AI has effectively ended attribution, so I'm not at all surprised people are less interested in sharing and publishing. Especially when anything you publish will just be used to train the next iteration that still won't credit you.

Your three examples historically were successful in achieving widespread usage, and in the case of SQL, shows no sign of disappearing.

Not just Eastern European stores, they do it in France (it's a a French company) and India too. First time I saw it was in India, it was very impressive.

They also have a lot of vertical integration, they manufacture many of the products they sell.

Having actually worked with industrial machinery, the proper UI for this is 1) unimpeded view of machine and 2) big red physical emergency stop button.

If you're relying on focus stealing or popups or notifications to tell you when your equipment is on fire, you're doing it wrong.

In the US at least, many pension funds are not sophisticated, they're small, underfunded, and getting taken for a ride by expensive advisors who promise fantastical returns that will help dig them out of their funding ratio hole. Many would be better off using an S&P 500 index fund for their equity component instead of getting wined and dined into an illiquid, opaque private equity investment.

Telling that among OECD countries, the US is an outlier in having a much lower average funding ratio, and this despite the fantastic performance of the US stock market over the last 15 years.

I find that Chinese manufacturing offers the complete range of quality for every price point. At reasonable prices, the quality meets or exceeds that of other countries. But if you want something cheap and cheaply made, you can find that as well. And at the extreme, you still have counterfeit parts and products.

Probably true, but Trello just had a major multi-day outage that prevented commenting on or moving cards, so I would be hard pressed to recommended it to anyone at this point.

A few years ago I wrote a workforce scheduling program designed to be used by non-programmers. I worked in a restricted environment so couldn't install anything. The whole thing ran on SWIPL's web offering.

Users simply had to change the basic "facts" (who was available on what days, how many people were needed), and the program solved for the various constraints and offered solutions.

It was maybe about 300 lines of Prolog, no complex dependencies. It replaced a pile of Python scripts that required a lot of state, didn't really work, and could only run on a few specific computers.

For regular users, it was relatively easy to understand and change the facts. SWIPL for web also offers a nice "notebook" interface that lets you mix data, code, and markdown / output blocks so the documentation was inline.

> Vibe-coding makes you feel like you have infinite implementation budget. You don't. You have infinite LINE budget (the AI will generate as much code as you want). But you have the same finite complexity budget as always.

This is a special case of a general fundamental point I'm struggling with.

Let's assume AI has reduced the marginal cost of code to zero. So our supply of code is now infinite.

Meanwhile, other critical factors continue to be finite: time in a day, attention, interest, goodwill, paying customers, money, energy.

So how do you choose what to build?

Like a genie, the tools give us the power to ask for whatever we want. And like a genie, it turns out we often don't really know what we want.

I get irritated by Zoom saying I need to update right when I open the app and want to join a call. Or even worse, sometimes I'll have had the app open (checking video and sound) and it won't notify about a required update until I actually go to join a call.

Never understood why they don't propose the update when the call has ended.