Re: "a little more significant than Doctor Who defeating the Daleks"
And don't forget the NSA/CIA. They've probably got complete Dr Who archives. Multiple copies
2372 publicly visible posts • joined 13 Sep 2017
My understanding is that bakers of the time caulked seams in their ovens with dough at the beginning of the working day. The thoroughly singed "cake" was scrapped out at the end of the baking day whereupon indigents would pick over it hoping to find some edible material. Could be a myth.
Even if true, I can't quite see an analogy to AI which seems to be somewhat usable on good days even if not remotely cost effective.
If there were a standard age file in Linux then Steam could use it too, and suddenly it all starts to make a bit more more sense
Inevitably, in many distributions, age will be installer age at installation time stored as an integer in /etc/age. But on others it will be user age at login time in seconds since epoch. And on red Hat and Debian it'll be in years in floating point. in /usr/local/age. There will be several other variations most of which will be against the law in India and Bolivia because those countries require age to be reported in hours since the birth of Ada Byron (aka Ada Lovelace) on December 10, 1815.*
*And, like the article says, underage users will most likely lie about their age anyway.
Sheesh!!! Is there a point to all this? From what I've seen, AI looks to be kinda cute but more or less useless except perhaps for the entertainment industry. Build a few research centers to explore quietly and objectively whether AI has any practical, cost effective uses, let Disney et.al. see if it makes movies/TV production cheaper/better and move on to trying to solve real world problems like getting a reasonable standard of living for everyone on the planet that wants it, and the practice of electing absolutely horrible human beings to political office.
Would it really make all that much difference in the long run to anyone except Sam Altman if humanity held off on AI for a few decades?
There are a bunch of potential problems -- pilot training, lack of obstruction free landing zones, weather, congestion, traffic control, insurance .... But the biggie is simply broken cars (mostly) stop, broken aircraft (mostly) drop. Aircraft falling out of the sky tend to be hard on both the payload and anything/anybody in the impact area. That said, some amount of airial taxi traffic will probably happen. I doubt it'll be particularly cheap. And the vehicles are going to have to be maintained like aircraft, not in the haphazard way cars are.
I once, in a former life at a very large semiconductor company, sat in a meeting to discuss a presentation that a Very Senior Executive was due to give, where an entire hour was spent discussing what color and size the heading on a powerpoint slide should be
One of the corollaries to Parkinson's Law (C. Northcote Parkinson, 1957 -- Work Expands to Fill the Time Available) is the Law of Inverse Proportionality The time spent debating an issue is inversely proportional to its objective importance.
Interesting thought, but No -- that's not realistic.
This is presumably a shipboard system. That means it has to compact, operate off ship power, not interfere with other shipboard systems, operate in the Arctic in winter as well as the Persian Gulf in Summer, tolerate high humidity, survive salt water spray, and a bunch of other things. Not only that, I was told back in the day, that, in practice, pretty much every ship in the US Navy is unique. No two ships have exactly the same equipment complement or layout. Every install is custom. I expect the same is true of the royal navy.
Then there's the issue of installation. Probably has to be done in a shipyard. Which means scheduling time in the yard. For every ship in the navy. And getting the ships -- some of which may be on the other side of the planet back to the yard in an orderly and timely fashion.
Ans there's the matter of training the crew to use the system. Probably requires simulators and course materials and instructors that have some idea how the thing works.
And I won't even go into the issues on the production side of manufacturing and quality testing a bunch of units in a hurry.
Overall, weeks sounds like wishful thinking.
Required reading: Franz Kafka -- The Castle. (Might be downloadable at https://dn790009.ca.archive.org/0/items/castle01kafk/castle01kafk.pdf)
Dealing with "Customer Service is already all too often a Kafkaesque experience. I can't imagine that AI will make it any better.
Moreover it will probably degenerate eventually into a "My AI will work out something with your AI" situation often yielding truly absurd results.
Error Correcting Code (ECC) memory requires a significant number of additional bits per word. The internet tells me 8 for 64 bit word. (Is that really how it's usually implemented?), Plus additional circuitry for checking and correcting? Is that likely to happen in the midst of a RAM shortage caused by our current race toward mediocrity?
But just restoring the parity bit that Microsoft talked RAM mongers into dropping three or four decades ago when expensive RAM was impeding deployment of GUIs would be a step in the right direction. Probably won't happen though if it looks like checking parity will reveal modern ram to be a bit flaky at times.
"You're screwed"
Maybe, maybe not. The case must wend its way through lower courts and various levels of appeal. Many lower court judges don't seem to be big fans of MAGA. And God only knows what the relevant laws actually say. It's probably pretty much a crap shoot as to whether Anthropic will prevail in the court system. Of course, the Supreme Court would likely uphold Trump. But they decide which cases they want to look at and actually only elect to review a select few cases.
Anthropic probably thinks their chances of winning in court in court are sufficiently good to justify the money spent.
Only the GDP of Ireland ($700B)? I thought we were talking real money here. Let me know when the bubble approaches the GDP of Canada or Italy. ($2500B).
One does have to wonder how they plan to recover these "Investments". They do have a plan, right?
Would it be rude to suggest that we unwashed masses might do well to develop realistic exit plans for products and services provided by this lot? It appears likely that computing services are going to get rather costly in the not too distant future.
Religious folk are very excited about the afterlife stuff, even though they haven't tried it yet.
Actually a good many of them seem to be a bit nervous about that afterlife thing. I can't think why. Presumably they have their reasons.
Yes, many people in the US drive compact cars. It sometimes seems like 10-15% of the cars in Vermont are gray Kia Souls. Even saw something that looked like a Japanese Kei car not too long ago. OTOH, a disturbingly large fraction of the vehicles are pickup trucks, some of which are gargantuan.
Biometrics?
I have a (purportedly) smart phone that requires a fingerprint to unlock but will erratically accept any of my fingers except the left index finger which never seems to work. I'm sure some other phones have less eccentric sensors/software. But I have to wonder how many don't.
Yes PayPal is clearly an online bank. But for two decades they have been operating in the US as an unregulated online bank. (I believe they have finally applied for a banking license in Utah). I've always found it hard to understand why anyone would deal with them without the consumer protections provided by banking regulation. I sure don't have anything to do with them.
I think we're still supposed to worry about the ozone (O3) layer because without it we'll have three eyed sheep in the Falkland Islands or some such. Ozone is quite reactive so I wouldn't be surprised that some of the material injected combines with the Ozone and depletes that layer. Honestly, I really don't know if ozone layer depletion is or ever was a serious concern. I'll leave it to those with a better grasp of atmospheric physics/chemistry to evaluate.
Dead right unfortunately. A short time limit will likely lead to automated takedown. That probably means "AI" which likely won't be cheap. Or particularly effective. And, for example, it's virtually certain given the number of crackpots in the world that takedown requests will turn up for online images of every nude work of art known to man from "Venus de Milo" to "September Morn."
There's probably some reasonable compromise. But it's going to take a lot of serious thinking to get the details right. And even then there will be problems. And abuses.
Hype: The tech bro's have sunk a lot of money into AI. And of course, the marketing scum really have nothing else on their cart other than AI to lie about. And AI isn't totally worthless. For example, it might have taken me two minutes to find a source for the definition of The "sunk cost fallacy". Goggle conjured one up in seconds.
The sunk cost fallacy is a cognitive bias causing individuals to continue an endeavor—such as a project, relationship, or investment—solely because of previously invested resources (time, money, or effort), even when current costs outweigh potential benefits. It is "throwing good money after bad" by ignoring that past investments are unrecoverable.
And the best part? ... Google is paying for the costs ... which will quite likely turn out in the long run to be unrecoverable.
"... how exactly does one create an AI Security company?"
Have you asked that question to your local AI?
I just asked Gemini (Google Search). The answer:
"Creating an AI security company requires assembling a team proficient in both AI engineering and cybersecurity, focusing on protecting, detecting, and responding to threats against AI models and data. Key steps involve developing, testing, and deploying specialized security tools, such as AI-driven vulnerability scanners or threat detection, while adhering to frameworks like NIST AI RMF, ISO 42001, and OWASP AI Security.
Of course your answer may be different
What popped into my mind about half way through the article was a line from The Incredibles "They keep inventing new ways to celebrate mediocrity."
What, exactly, is the point of all this effort to produce products that are pretty much Meh ... and apparently can not, by design, improve?
35 years ago I was banking online using a 300 baud modem. Then I moved to a new town a long way away. None of the local banks supported online banking. But the internet was slowly coming to life. So sometime around 2000, I took a serious look at moving the family finances online. Didn't look all that safe. Decided I'd wait until the numerous security problems were cleaned up on a non-nightmarish way.
Still waiting.
Indeed, the German AVR reactor failed. Early and often apparently. Which presumably isn't entirely unexpected for a research effort. But it did, as advertised, always fail safe.
That's likely an important consideration as it is clear that the problems associated with powering a modern industrial society with intermittent wind and solar are quite likely insurmountable. That seems to mean nuclear power. Lots of it. While it does seem to be possible to operate a conventional nuclear plant safely, I have my doubts that the assortment of bean counters, marketing-clowns, and assorted sociopaths humanity often puts in decision making rolls is going to be able to do so. At least not everywhere and always.
Good Summary of the history. But it misses what seems to many the whole point of pebble bed reactors. In concept at least, when they fail, they fail safe. They can't meltdown and spew long lifetime fission products over a good sized chunk of countryside because the fissionables can never reach critical mass no matter how ineptly they are handled. Instead you end up with a (very) hot pile of "rocks" which, if it bothers you, you can subdue just by pulling it apart.
Dependence on AI. At long last!!!! The great debate is over!!!. The proof has been revealed!! There is (probably) no intelligent life on Earth.*
* At least not amongst terrestrial vertebrates. We're still not sure about octopods so it's remotely possible that there are intelligent marine invertebrates on the third planet.
The Tao Ching tells us that a journey of 1000 miles starts with a single step. And modern AIs are certainly at least one step away from ELIZA (1966).* But one has to wonder if this giant step is even in the right direction to get to any reasonable destination.
*It has been 60 years after all.
"I wonder if the reason why you see fewer Starlink satellites in Jakarta"
It's just an artifact of the way that near circular satellite orbits work. It's not that easy to describe in words, but the result is that the density of satellites is lower near the equator than in higher latitudes. European, North American, and East Asian locations will see more Starlink satellites at any given time than will equatorial locations like Singapore, Jakarta and Quito.
The hi-tech answer for such problems is presumably to require 2FA to access the trash bin. That way your relative won't forget to take his phone outside with him.
(Tangentially, what was said relative's plan for access if/when his phone battery died, was bricked by an ill crafted OTA update, broken, or stolen?)
Typical 2FA implementations can be a monumental PITA if you don't have (reliable) cell phone reception. Maybe 2FA isn't a problem for you. But keep in mind that you are not everybody. I suspect that many 2FA implementations are actually unusable for some handicapped persons and other edge cases.
They don't even have to make the ads less gaudy. If they will just keep the @#$%^ things from displaying on top of content, then I can turn off my ad blocker and treat them like newspaper or magazine ads. It would help a lot if everyone would adopt one simple rule for a year or three: If an ad is displayed on top of site content, don't buy that product.