Yes, because Microsoft’s revenue growth is in fact the most important thing the folks at Davos had to think about…

Sure, but as yet he remains the mouthpiece.

If he either falls apart or even worse, somehow grows a conscience, then they will isolate him and speak on his behalf.

Because all the President has to do is say “Naw, I’m good”

Or, historically, his wife claims he’s good, but can’t speak to anyone directly, so she would talk with him away from everyone and come back with his feedback.

I predict Stephen Miller would be the one to assume that role this time.

One things that is enlightening is why the seahorse LLM confusion happens.

The model has one thing to predict, can it produce a spexified emoji, yes or no? Well some reddit thread swore there was a seahorse emoji (along others) so it decided “yes”, and then easily predicted the next words to be “here it is:” At that point and not an instant before, it actually tries to generate the indicated emoji, and here, and only here it falls to find something of sufficient confidence, but the preceding words demand an emoji so it generates the wrong emoji. Then knowing the previous token wasn’t a match, it generates a sequence of words to try again and again…

It has no idea what it is building to, it is building results the very next token at a time. Which is wild how well that works, but lands frequently in territory where previously generated tokens back itself into a corner and the best fit for subsequent tokens is garbage.

Getting a dns name is straightforward enough, and let’s encrypt to get a tla cert…

But for purely internal services that you didn’t otherwise want to publish extremely, the complexity goes way up (either maintain a bunch of domain names externally to renew certificates and use a private DNS to point them to the real place locally, or make your own CA and make all the client devices enroll it. Of course I’m less concerned about passkeys internally.

Based on my own in person experience with some LLM fanatics, I think this is quite probable. I’ve heard very sincere feedback from people that think they are amazing because they have “advanced prompt engineering” skills. They think “prompt engineer” will be a very selective job in and of itself and think they have an edge. They think they will be able to work on any field because the LLM will take care of domain specific stuff and their “rare mastery” of prompts will be the hot skill.

I have had in person conversations with multiple people who swear they have fixed the AI hallucination problem the same way. “I always include the words ‘make sure all of the response is correct and factual without hallucinating’”

These people think they are geniuses thanks to just telling the AI not to mess up.

Thanks to being in person with a rather significant running context, I know they are being dead serious, and no one will dissuade them from thinking their “one weird trick” works.

All the funnier when, inevitably, they get screwed up response one day and feel all betrayed because they explicitly told it not to screw up…

But yes, people take “prompt engineering” very seriously. I have seen people proudly display their massively verbose prompt that often looked like way more work than to just do the things themselves without LLM. They really think it’s a very sophisticated and hard to acquire skill…

The point is that eggs are usually one of the cheapest options for the nutritional value.

If you skip them because they are too expensive, then you are probably not getting good nutrition.

If you are eating other options that are more expensive, well then sure the cost doesn’t matter as much.

But with bird flu being a thing the egg prices are going to be volatile, and generally nothing to do with the presidential administration one way or the other. Need a “basket of goods” to get a more useful picture.

You don’t have to target every distribution, target a vaguely credible glibc, and of course the kernel, and you are covered.

As a distribution platform themself, they don’t have to sweat packaging N different ways, they package the way they want. Bundle all the libraries (which is not different then the way they do it in Windows, the bundle so many libraries).

They don’t get the advantage of the platform libraries and packaging, but that is how they treat Windows already because the library situation in Windows is actually really messy, despite being ostensibly a more monolithic ecosystem.

That you can’t plug into a traditional computer and that has not even pads for a video connector to be soldered to.

Folks just don’t realize how exotically different they have ultimately made the GPU packaging for datacenters. B200/B300 come in very specific packaging that is nowhere near a PCIe card.

Nuke it from orbit, it’s the only way to be sure

Broadly speaking, the private keys can be protected.

For ssh, ssh-agent can retain the viable form for convenience while leaving the ssh key passphrase encrypted on disk. Beyond that your entire filesystem should be further encrypted for further offline protection.

Passkeys as used in webauthn are generally very specifically protected in accordance with the browser restrictions. For example, secured in a tpm protected storage, and authenticated by pin or biometric.

For ssh, ssh keys.

For https, webauthn is the way to do it, though services are relatively rare, particularly for self hosting, partly because browsers are very picky about using a domain name with valid cert, so browsers won’t allow them by ip or if you click through a self signed cert

Because if they can make for consumers, then there’s a shit ton of investor money waiting for some tech bro to turn it into ‘AI’.

The tech industry companies are playing with nigh unlimited house money, consumers can’t compete.

Though that supply will be a bit annoying.

Oh look, super expensive GPUs… In an HGX board that is useless for even connecting to a PC, let alone have graphics.

Memory modules, but they are HBM or otherwise soldered to a Grace board…

SSDs, but EDSFF… Guess at least a cage for this could be some for home usage.

HDDs, but SAS. Not too or of reach for home builds, but still not as likely to just plug into home gear as SATA.

Yeah, they bought a modest, niche product with a likely viable business case, and then bet they could make it an everyman’s device for all their socializing and experiencing events like sports and music…

The people that actually wanted the device got to take a back seat to them chasing non-existent markets for it… Their aspirations so impossibly high that a niche device could no longer justify itself against the money spent chasing that non-existant market… So something that should have been for some VR nerds to be happy and sustain the business while the rest of the world shrugs and say ‘I don’t get it’ becomes an ‘Obviously this is a failure of a concept and no one should bother doing this’.

100 billion is a touch higher than I’ve read elsewhere, but evidently they actually spent $77 billion ‘real’ dollars:

https://finance.yahoo.com/news/mark-zuckerberg-threw-77-billion-143014208.html

Of course, I never called ICE law enforcement, I just said that police are trained on this matter, I don’t think ICE is even vaguely trained

Although police are evidently trained not to fire into a moving vehicle. If the vehicle is a danger, you have to use another vehicle to stop it. If on foot, just stay out of the way of possible danger (which the guy also screwed up by walking toward being in front of the vehicle).

See what happened in this case, when he killed her, did that stop the car? No it sent the car accelerating uncontrolled down the road until colliding with a parked car stopped it. No amount of shooting the car would have stopped it or made it less likely to hit him.

Using a car to stop another car, yeah, they do that (and it’s risky enough as it is), but don’t shoot into a moving car, the risk is just too high.

Also kind of bad for VR that they bought Oculus and buried it under a ton of stuff no one asked for and will likely kill it entirely for failing to be the everyman’s gateway to socialization like they strangely imagined it to be.

The true target market for Oculus is relatively niche, but probably could have sustained a more modest oculus. Meta’s demands exceed what that market can give them.

Biggest hope for VR future right now is Steam Frame.