Off-and-on trying out an account over at @tal@oleo.cafe due to scraping bots bogging down lemmy.today to the point of near-unusability.
You almost always want to pair lsof with -n, as by default, it (slowly) lists a bunch of things that require name resolution.
$ time lsof >/dev/null
real 2m12.352s
user 0m0.499s
sys 0m1.217s
$ time lsof -n >/dev/null
real 0m1.600s
user 0m0.523s
sys 0m1.029s
$
You can also still access the file as long as there’s a process that still has it open. I have, in the past, “undeleted” a file or two doing that.
$ echo foo > bar
$ tail -f bar
foo
In another terminal:
$ pidof tail
1525534
$ ls -l /proc/1525534/fd|grep bar
lr-x------ 1 tal tal 64 Jun 16 06:42 3 -> /home/tal/bar
$ rm bar
$ ls -l /proc/1525534/fd|grep bar
lr-x------ 1 tal tal 64 Jun 16 06:42 3 -> /home/tal/bar (deleted)
$ cat /proc/1525534/fd/3
foo
$ cat /proc/1525534/fd/3 > bar-recovered
$ cat bar-recovered
foo
$
That is, the /proc entry for tail’s file descriptor 3 there looks kinda like a symlink, but the kernel doesn’t actually make it behave in quite the same way as a normal symlink.
That being said, getting back to the original point about unlinking not being able to remove the directory entry…it won’t sit there blocking you from putting a new directory entry there with the same name, the way Windows file semantics mandate.
EDIT: Also, what rm removes is the directory entry rather than the inode. The inode sticks around as long as the file data is there. You can have multiple directory entries for an inode, or none at all, but file data will have an inode associated with it.
$ touch a
$ sudo ln -T a b
$ stat -c %i a
216538023
$ stat -c %i b
216538023
Same inode, different directory entries.
https://en.wikipedia.org/wiki/Inode
inode persistence and unlinked files
An inode may have no links. An inode without links represents a file with no remaining directory entries or paths leading to it in the filesystem. A file that has been deleted or lacks directory entries pointing to it is termed an ‘unlinked’ file.
Such files are removed from the filesystem, freeing the occupied disk space for reuse. An inode without links remains in the filesystem until the resources (disk space and blocks) freed by the unlinked file are deallocated or the file system is modified.
Although an unlinked file becomes invisible in the filesystem, its deletion is deferred until all processes with access to the file have finished using it, including executable files which are implicitly held open by the processes executing them.
There have been cars that can manage it.
https://rarehistoricalphotos.com/amphicar-president-johnson-1960s/
The Amphicar was manufactured in West Germany and first made an appearance in America at the 1961 New York Auto Show.
Originally designed by Hanns Trippel, these amphibious cars hit the US market from 1961 to 1967. Most notable for their ability to cruise on land and operate in the water, these cars were a fun new invention in the car world for that time.
One of the most famous owners of the Amphicar was President Lyndon Johnson. Those leisurely rides created the opportunity for a clever prank that Johnson enjoyed playing on his passengers.
It involved barrelling the vehicle down a hill into a lake and exclaiming to his passengers that the brakes had malfunctioned, while they had no idea the car was designed to float in water.
Police arrested Jimmy Jack McDaniel, 70, after he and two visitors from Germany to whom he was giving a ride became marooned near the Katie’s Woods Park Boat Ramp around eight o’clock at night.
Making sure that they get the full tourism experience, I see.
https://en.wikipedia.org/wiki/Business_Insider
In 2023, Business Insider shifted its organizational model, adding multiple artificial intelligence (AI) products in 2024, and reducing its staff by nearly 40% between April 2023 and May 2025.[7]
I suspect that the author is more likely to be impacted than most of the people involved.
Journalism’s been having a rough time for some decades from technological change, though that predates AI as we know it today.
First — in the US, not sure about everywhere else — there was a shift away from local news towards focusing on national news. You don’t need as many journalists to cover a limited number of national stories. IIRC, that started before widespread Internet adoption, but the Internet accelerated it a lot:
https://theharvardpoliticalreview.com/local-news-democracy-risk/
The appearance of news deserts across counties and communities in the U.S. has been a widespread phenomenon in recent years. But why? In an interview with the HPR, Jeremy Meserve, the Staff Producer and Archivist for the Belmont Media Center, pointed to the over-corporatization of media consumption as a cause of the decline in quantity and quality of local journalism.
The rise of social media platforms like Facebook, Twitter, and Instagram has fundamentally changed how people consume media. Social media platforms like these have spelled the end for many local newspapers as people have shifted their media consumption priorities to more convenient options. Meserve believes that part of the downfall of local newspapers had to do with the old business model, where many local papers were free. So when social media emerged, people stopped reading, as social media platforms provided faster and equally free media. As a result of this, newspapers lost their audience and their benefactors which led to that old business model being unsustainable.
Second, Google basically took over the ad market that a substantial amount of journalism relied on for revenue. Sure, some money came from subscriptions, but a lot of magazines and newspapers relied on their ability to put ads in front of a broad demographic’s eyeballs. You don’t want to pay a newspaper for relatively untargeted ads when you can pay Google, which can hit exactly the demographic that you want to advertise to.
Third, my understanding is that some stuff — like “business news” articles, where one just wants a summary of earnings reports or someone talking about the general movement of stocks and a vaguely-plausible explanation attached — became largely automatically generated some time back. This predates the LLM boom as well:
searches for an example
https://www.ap.org/the-definitive-source/announcements/automated-earnings-stories-multiply/
The Associated Press, working with Automated Insights and Zacks Investment Research, is now automatically generating more than 3,000 stories about U.S. corporate earnings each quarter, a tenfold increase over what AP reporters and editors created previously. Here, Assistant Business Editor Philana Patterson, who has been overseeing the rollout of this process in the newsroom, gives an update on AP’s automation efforts that began last summer.
That might sound like something happening today, but…that’s a story from June 2015, over a decade ago.
roughly 3,800 internal repositories
I suppose that part of the moral here is to compartmentalize information internal to a company. Like, if you’re not on the team working on X, then you probably shouldn’t have repository access to X.
That’s an interesting thought.
thinks
Tax revenue would be less-frequent, and there might be potential to create a misincentive to encourage people to unsafely drive on threadbare tires longer than they otherwise would. But I could see that being done.
But is $130 actually fair?
Well, a flat fee doesn’t take into account vehicle weight or annual mileage, which the gas tax more-or-less does. And the road maintenance cost is a function of those two things. A flat fee would penalize drivers of infrequently-driven small vehicles.
But…I suppose that gathering that data would also add some privacy concerns and costs, like the government needing to record how many miles your vehicle has traveled in a year.
EDIT: The really obnoxious thing is that everyone else is grabbing movement data on vehicles to make money off. Automakers via integrated cell radios. ALPR network operators. I assume that charging station operators do too — fast DC connections like NACS transmit the vehicle’s VIN, and I’d be very surprised if charging companies aren’t monetizing that data.
Gotcha. Yeah, the stuff in fstab is just a convenience; it’s equivalent to running a bunch of mount commands at boot. You might be able to just run “mount” again without the '-o remount" option. I was just listing that in case you were seeing some kind of errors in trying to manually mount it.
That’s a neat tidbit.
https://en.wikipedia.org/wiki/London_Underground_cooling
The heat in the tunnels is largely generated by the trains, with a small amount coming from station equipment and passengers. Around 79% is absorbed by the tunnels’ walls, 10% is removed by ventilation, and the other 11% remains in the tunnels.[3]
Temperatures on the Underground have slowly increased as the clay around the tunnels has warmed up; in the early days of the Underground it was advertised as a place to keep cool on hot days. However, over time the temperature has slowly risen as the heat sink formed by the clay has reached its thermal capacity. When the tunnels were built the clay temperature was around 14 °C (57 °F); this has now risen to 19–26 °C (66–79 °F) and air temperatures in the tunnels now reach as high as 30 °C (86 °F).[3][4][5]
Nah, that’s good. What I mean is, if it’s in /etc/fstab, it should be possible to manually mount it without a reboot. Have you tried manually remounting it after power comes back?
Does anyone know how I can resolve this issue?
I don’t know why you’d need a reboot to remount the thing. Are you just not familiar with how to add something to /etc/fstab and mount it manually and are relying on some kind of auto-mounting system that only happens to run at boot, or is it giving some kind of error?
If an error, what happens when you do:
$ sudo mount -o remount /mnt/the-mount-point
?
I was going to mention that, but also that while I’m sure that it’d handle power loss while it’s unmounted the filesystem I don’t know what happens if the backing storage goes away while it’s in use.
why they’re letting so much heat energy out into the atmosphere
Surely at least some of that heat could be tuned back into electricity.
To harness useful energy from heat, you have to let heat flow from hotter areas to colder areas, to permit entropy to increase.
https://en.wikipedia.org/wiki/Entropy
Entropy is central to the second law of thermodynamics, which states that the entropy of an isolated system left to spontaneous evolution cannot decrease with time. As a result, isolated systems evolve toward thermodynamic equilibrium, where the entropy is highest. “High” entropy means that energy is more disordered or dispersed, while “low” entropy means that energy is more ordered or concentrated.
They might be able to harness energy from the flow from warmer to cooler areas, but whether or not they do that, at the end of the day, they have to let the heat go, just like a power plant that uses water-evaporation-assisted cooling. If they’re near the ocean, they can maybe stick it into the water instead of the air, and maybe to some degree, you can stick heat into groundwater. But they can’t just take a unit of heat and convert it into a unit of useful work and not have that unit of waste heat.
You can, in areas that have a use for heat, make use of that waste heat. For example, district heating can make use of the waste heat from a power plant — you pipe steam or something from the power plant that you want to be cooler to homes that you want to be warmer.
District heating (also known as heat networks) is a system for distributing heat generated in a centralized location through a system of insulated pipes for residential and commercial heating requirements such as space heating and water heating. The heat is often obtained from a cogeneration plant burning fossil fuels or biomass, but heat-only boiler stations, geothermal heating, heat pumps and central solar heating are also used, as well as heat waste from factories and nuclear power electricity generation. District heating plants can provide higher efficiencies and better pollution control than localized boilers. According to some research, district heating with combined heat and power (CHPDH) is the cheapest method of cutting carbon emissions, and has one of the lowest carbon footprints of all fossil generation plants.
If you live somewhere where that works, it’s basically “free” heating from an energy standpoint, which is cool. Much of the US isn’t well-suited to residential district heating, because we tend to have residences in low-density suburban areas that are pretty spread out and where it’s a pain to transport heat around, but we do have some district heating in city cores. Manhattan, which is one area where we do have high density, famously uses steam heating.
Today, Con Edison operates the largest commercial steam system in the world (larger than the next nine combined).[4] The organization within Con Edison responsible for the system’s operation, known as Steam Operations, provides steam service to over 1,700 commercial and residential customers in Manhattan from Battery Park to 96th Street uptown on the west side, and 89th Street on the east side of Manhattan. Roughly 27 billion pounds (12,000,000 t) of steam flow through the system every year.
For that to work, you have to actually have some use for that heating (and you probably only want heating some of the year, unless you’re up in the polar regions or on a mountain or something).
You can also use waste heat to drive industrial processes that require heat, but waste heat from a datacenter isn’t super-hot compared to, say, that from a power plant, so I don’t know how interesting that necessarily is. Lots of chemical processes that might require elevating something to a much higher temperature, but a datacenter — at least using current computing hardware — normally tries to keep temperatures from getting to something like the boiling point of water.
Some greenhouses will also use waste heat (in the case of power plants doing cogeneration, some of the waste carbon dioxide as well) to help boost plant growth.
Well, now we’ve got Trump on the right imposing new regressive taxes:
https://www.hbs.edu/bigs/robert-lawrence-qa-trade-tariffs
And since tariffs are only levied on goods, and since it’s poorer Americans who spend a greater share of their consumption on goods than rich Americans, this is a highly regressive tax.
And Newsom on the left:
https://en.wikipedia.org/wiki/Sales_tax
Because the rate of a sales tax does not change based on a person’s income or wealth, sales taxes are generally considered regressive…Investopedia defines a regressive tax as “[a] tax that takes a larger percentage from low-income people than from high-income people. A regressive tax is generally a tax that is applied uniformly. This means that it hits lower-income individuals harder”.
EDIT: As of 2025, California has the eighth-highest average combined state-local sales tax rate in the US, at 8.802%:
https://taxfoundation.org/data/all/state/sales-tax-rates-2025/
I mean, you can install each in a VM if you want to play with them.
Mistral AI’s CEO says Europe has 2 years to stop becoming America’s AI ‘vassal state’
Also on this page:
https://www.businessinsider.com/elon-musk-xai-explored-collaborating-with-mistral-cursor-2026-4
- Elon Musk’s xAI recently announced a partnership with AI coding startup Cursor.
- The two companies have also explored a three-way partnership with French AI startup Mistral.
- Musk is looking to catch up to Anthropic and OpenAI.
I mean, I suppose “know your audience”. If you’re going to the French legislature for money, you pitch accordingly.
“It will be decided in the next two years,” Mensch said during a Tuesday hearing on digital sovereignty and AI at France’s National Assembly, translated by Business Insider.
I suppose that if you’re another development studio working with Krafton, like InZOI, you’re probably looking with a lot of suspicion at every move from your publisher now.
rsync is a fine tool, but it’s not really for backups. It’s for unidirectionally syncing data, which is part of a backup.
There are tools that use librsync to do a backup system, like rdiff-backup or duplicity.
I have used rdiff-backup in the past, but restic supports deduping data keyed off content rather than filename, which is functionality that I rather like having. I drive a restic job from backupninja, which handles doing things like notifying you if a backup has failed.
Meta’s you kind of expect. They’re cranking down R&D on VR and up R&D on AI in their research wing, and while maybe some people have enough overlap in skillset to retain, a lot are going to be specialized.
I’d also add that I think that’s probably a good move. They’d caught a lot of flak for a long time over the potential returns on VR being limited, and by this point, they’ve lost a lot of money on it. I don’t think that there’s a clear path to a point where their Metaverse stuff takes off.