I'm assuming that when they do one of those distribution plots of the output values (the ones that show clear patterns for pseudo random generators when run for long enough) they can prove that the distribution is totally uniform, and with time as a further axis, every attempt achieves that even distribution in a different sequence. That implies they can account for, or negate the impact of, every potential variable in the system.

Lava lamps (like Cloudflare actually use as part of their RNG, IIRC) might be just as good, but mathematically proving that could be a little more challenging, and there may be any number of corner case effects, such as the temperature on either side of the glass or minute variations in the heating coils, that cause an almost imperceptable bias towards the denser coloured fluid in the lamp being in certain parts of the lamp than others for short periods of time.

As to whether we need this, quite probably not. However, TFS does propose use as a kind of "master clock" to regulate other systems that would be less precise (or random, in this case) on their own. Whether that's more cost effective or practical than just combining multiple sources of randomness together to get a single output data stream I guess will be determined by any users that really, really, need a truly random data stream, and how the realities of a post-quantum world eventually play out. If you are in some kind of situation where an adversary can keep retrying at a suitable rate and only needs to predict/guess the next in sequence correctly once to "win", then perfect randomness over a sufficiently large search area is something you are going to be all over.

See my post above for a bit more detail, but this looks like it could be an SPF include failure. They have included "_spf-ssg-a.microsoft.com" in the SPF, which in turn includes "spf.protection.outlook.com". AFAIK, that's basically the Outlook.com webmail service, so quite possibly at least some, and possibly any, users of that service could impersonate "microsoftonline.com" and get an SPF pass.

If so then yeah, that's *totally* the kind of lack of attention to detail you tend to expect from Microsoft.

It was (and still is), but this is the problem that SPF was designed to solve (as opposed to being the FUSSP some made it out to be). If you have a critical domain that you use for sensitive stuff, like "microsoftonline.com", or any bank's domain, etc, then you need to be very specific on your SPF record's contents and make sure it has an "-all" in there to force a reject for failures, rather than the looser "~all" or (heaven forbid) "+all" which is really only intended for testing. Spammers know this, and seldom waste their time trying to spoof domains that will cause a failure; every domain I have setup SPF with "-all" on has seen Joe-Job bounces drop to zero pretty much overnight. DKIM works slightly differently, but adds another layer to this. Microsoft for sure knows this too and does indeed do both, but that doesn't mean you can't slip up and leave a hole somewhere.

So, taking a quick look, as things stand, the SPF record for "microsoftonline.com" is:

"v=spf1 ip4:216.32.180.228 include:spfa.microsoftonline.com include:spf-exacttarget.microsoftonline.com include:spf-msods.microsoftonline.com include:spf-mfa.microsoftonline.com include:_spf-ssg-a.microsoft.com -all"

They've got the "-all" in there, which is good, but also a whole bunch of "include" directives, including one that refers to ExactTarget a third party MSP, but the one that appears like it could possibly be the problem is the last one. That contains a further include, and in there is "spf.protection.outlook.com". All the includes do have "-all" but, AFAIK, that domain covers the outbound mail gateways for a least some parts of the Outlook.com webmail service, so if the spammers have been able to a suitable account using a server within one of the many IP ranges listed in that include that doesn't properly restrict the domains able to send their mail, then they are good to go.

I think it's more horses for courses, and can also vary considerably between what different demographics, both contemporary and historical, think of as "perfection". Hollywood is largely driven by white western males, so they naturally favour your "20% silicone", although that does seem to be undergoing a gradual change of late, but that's not the case for world cinema as a whole; you'll find far fewer wannabe Barbie Dolls in African cinema, for instance.

From a people portraiture perspective, especially candids, there is also a night and day difference between what a photographer would most typically want to shoot in a studio vs. on the street. The former is very much about some ideal of perfection, with hours spent on makeup and clothing the model(s) and setting up the lighting rigs, before the camera even gets turned on, whereas in the streets and fields, you are totally going to home in the people with the most interesting features, and those often tend to be very much the definition of imperfection. You are actively looking for the aged faces with more lines than a metro map, more piercings/tats than Vogue would likely ever consider acceptable, and anything else that really tells a story about the kind of life the viewer of the resultant image might imagine them to live. For the right images, there is absolutely value that can be measured in both clicks and dollars there too.

Also, why limit it to women implanting silicone to comply with some visual aspirational idea of perfection being forced on them by men (mostly), media, and entirely unrealistically proportioned dolls? Have you seen the lengths some men are going to as part of the "looksmaxxxing" fad? There's going to be a Darwin Award winner there real soon now, I'm sure.

Doesn't count until we see some in the wild.

Actually, I think it was Fred Schwed Jr. 86 years ago, but still just as relevant. In case it's not obvious from the fee, this is totally not about making YOU financially better off.

I'm guessing the tank has enough positive pressure left internally that it can withstand the pressure exerted upon it by the upper stage's engine exhaust without collapsing until the upper stage is clear. If it deforms, then presumably it would not be able to be reused, but if it can withstand the pressure long enough just fine then that removes the need for some additional shielding, and the mass that entails.

You're assuming the companies with these fleets of (currently largely non-existant) robots are still going to solvent if the bubble pops. That seems highly unlikely in many cases given the business model for AI is apparently "borrow massive amounts of money to fund it using the promise future orders as collateral". Asset strippers have no interest in salvaging a business; their business model is to buy the physical assets cheap, dump the debt on to bagholders (the shareholders), and sell the assets off to whoever wants it, hopefully for more than the cents on the dollar paid they probably for it. I buy stuff from these auctions from time to time; it's a great way to get nearly new, and often still on the market, kit at a fraction of the retail price.

Also, Facebook might not be the best counter example there. Remember what happened to many of the hires, business units, servers, and services, Meta setup when Zuck went all-in on the Metaverse? What do you think he'll do if going all-in on AI doesn't pan out for him?

Yeah, but these are Humanities students. That, by its very definition, is an area where AI should have very limited use, where it is applied should be done really, really, carefully, and job losses are far less likely than in many other fields. Sure, there's analysis of datasets, especially of geographical and historical data, but that is one of the areas where a specifically trained model can really be of use, but an AI is never going to painstakingly brush away dirt from some ancient historical site, and I shudder to think what would happen if AI hallucinations get let loose on philosophy or religion. That said, it would probably be very amusing watching those who buy into the output; and doubly so if the model was trained on the Butlerian.Jihad from Dune, less so for actual crusades, jihads, and "holy" wars.

Still, if these presumably tech savvy Gen Z students are not fans of the tech, regardless of whether that's because the recognise how its being used by corporates or some other reasons, then I think the people that need to be more worried about this are those that have built the massive pyramid trillions of dollars of debt to build something that few seem to want or trust. Like the .com boom, the bubble must pop sooner or later and sort out who is a "pets.com" and who is a "google.com", and there are growing indication that, unlike .com, the demand that will be required to pay for it all just isn't there, and we're already way beyond the scale of any previous government bailouts. That kind of crash only has one outcome; a lot of shareholders (which includes pension funds) are going to lose their shirts.

According to TFS, the layoffs are due on 20th May. No one is going to voluntarily quit if they can just phone it in for another 8 working days and get at least some additional severence pay to tide them over while they look for a new job. If they don't get cut and are still hacked off enough on the 21st, that's probably when people are going to start to quit.

Of course, one thing Meta is very good at is profiling people. And another, as TFS points out, is being callous sociopaths. Chances are they've factored all that in and I wouldn't be at all surprised if their actual target is a 15% RIF and they've worked out that if they fire *this* 10% on the 20th, then *this* further 5% that have definitely had enough and were hoping to be laid off will be so fed up with the loss of their former colleagues and even more hostile workplace will quit of their own accord over the next few weeks. If Meta was aware you were looking for another job before they announced the 10% RIF, it's pretty good bet you're in the additional 5% they are hoping for.

It's not a lake under the city, it's an aquifer, so it takes quite a bit of time for the water to disperse, rather than flow, through it. Replenishing a little bit of the water in one area through a leak might stave off some of the sinking in that area, but the areas where water is being extracted from will continue to sink much faster, with the additional complication that the density of the aquifter likely varies as well. The net result is the same though; different parts of the city sink at different rates, with those near leaks or denser parts of the aquifer slower than those near extraction points or the more porous areas, hence all the tilting buildings.

I'll second this â" I did mushrooms once many years ago, and it ranks as one of the most intense euphoric experiences I have ever had. It was truly life-changing.

2 drivers

Sounds like a textbook example for why layoffs are a thing.