Despite the 2C number having been nothing more than wild-ass guesses in scientific papers before being seized on by politicians eager to be seen as doing something -- anything -- to address the existential crisis of "global warming". At least, until "global warming" began to produce colder weather, warmer weather, dryer weather, wetter weather, more storms, fewer storms... and "global warming" was rebranded as "climate change" and hooked into attribution studies so that any weather that deviated from a continual mild, calm climate could be pointed to as evidence of "climate change".

And if you go to the NOAA website to find the year-by-year graph of atmospheric CO2 level and match the graph against important dates in the fight against CAGW -- the launch of the UNFCCC, COP1, COP3, the Kyoto Protocol Treaty, COP15, COP21, the Paris Agreement, COP28, and COP30 -- you'll see that all the efforts of all the world governments to reduce anthropogenic CO2 emissions have done Sweet Fanny Adams to reverse, or even slow, the rise in atmospheric CO2 levels. But, of course, if you follow the pronouncements of the CAGW adherents, it's still all our fault -- individual identical molecules of CO2 know whether they come from natural or anthropogenic sources, and the CO2 molecules from anthropogenic sources hang around in the atmosphere decades if not centuries longer than the CO2 molecules from natural sources.

Needs to be:

Microsoft Edge: "You are part of the rebel alliance and a traitor. Take it away!"

I have to wonder how many lawsuits would be generated from the "drunk driving detection" false-positive result from the driver swerving to avoid a road hazard incapacitating the vehicle at speed and causing an accident with varying degrees of injury or fatality...

...or sue other people for producing material that more-or-less, if you squint and look at it crosseyed through distortion glasses, might be construed as vaguely resembling something your company produces. (GW having sued a miniature designer for producing STLs of miniatures that 'can be used for playing Warhammer', a description equally applicable to Plastic Army Men or hand-drawn cardboard counters)

I wonder how much of this " 'unprecedented' demand " is Meta deliberately restricting production to whip up more interest by creating an artificial scarcity, and whether we'll quickly see an announcement that 'the first production run has sold out', and prospective buyers will have to wait until the second or subsequent production runs, fanning FOMO to drive more interest.

Depending on how deeply you drink the conspiracy-theory Kool-Aid, this is the preparation for Microsoft releasing a CD album and track subscription service that's an add-on to the new Windows Copilot Media Player, which will, with no way to disable the feature, automatically suggest other albums you might like based on the CD you have in the player (the Copilot-driven suggestions will be part of the base app, not the add-on; the base app needs to upload your CD usage to Microsoft for them to monetize your music interests to other companies). As part of this, Windows Copilot Media Player will require an active internet connection, and will not play music if it can't reach the Microsoft servers.

The Tesla semi is supposed to have a 900kWh battery, so even at the full 1.2MW delivery, it will still take 45 minutes to fully charge its battery. Assuming that, to preserve battery life, it's charged only to 80%, and recharged when between 10% and 20%, that's a half hour at the charging station. A large truck stop on an interstate can have 15 to 20 fuel pumps, with a fillup taking 10-15 minutes; that makes a recharge 2x - 3x the time to fill up with diesel (unless the jerk in front of you has gone in to take a dump or a shower while their truck is sitting at the pump). Now assume that the electric truck stop has 20 charging stations and a 10MW power feed; if it's full up, you're only getting 500kW; that pushes your charging time to 60-75 minutes.

And I have to wonder what the net efficiency of the process is -- i.e., if you're getting 200MWh of energy out of a fully-charged system, how much more than 200MWh did you have to put in to charge it?

And 10 hours of storage is enough to flatten some of the day/night cycle of solar power, but it doesn't do much for protracted low-production periods -- if it's overcast for three days, and the solar farms are producing 10% of their normal power, that's 72 hours where the pumped CO2 storage has to provide power, so a plant 10x the size would only cover the base 20MW with a cushion for another day, and if someone builds a power-hungry AI data center nearby in the expectation that they'll be getting reliable power, they can drain the storage pretty quickly.

I'm not sure that it's possible to come up with a way to throw AI systems out the back of a spaceship fast enough to make them viable as a propulsion system, no matter how much better they would be used that way.

They'll need something to keep the lights on during a dunkelflaute. Oh, wait; with Millibrain's wisdom, all the subjects can just sit in the dark when the renewables crash and take the entire country's grid with it -- it's the AI datacenters that are important enough to get continuous power.

And the ongoing implementation of misfeatures that just add instability to the OS. These "StartMenuExperienceHost" failures? WTH is the start menu 'experience' and why would I want it? Just give me the fscking Start Menu and have it do what its name implies without any bells and whistles; if I wanted some AI-curated selection of what the system thinks I want to do, I could just dump a pile of fewmets on my keyboard and get the 'experience' directly.

A thought just occurred to me... could Microsoft relying more and more on AI-generated code explain some of the increasing enshittification of Windows? And Microsoft execs asking AI to tell them what new 'features' to add to Windows account for most of the rest?

If there is a limit on the size of the config file used to identify automated traffic, then the process for updating it needs to save the updated file in temporary storage and sanity-check it; if it fails the sanity check, it should throw an error and reject the update. If it passes the sanity check, the old file should be backed up, and only then should the updated file be moved to replace the old one. That way, if there's an identifiably out-of-band update, it doesn't stomp on the current good data, and even if it passes the initial check, the previous iteration of the file is immediately recoverable to restore service. And there should be a utility that copies -- not moves -- the backup file back to the main file.