Posts by vtcodger

Anthropic says text watermarking scheme relies on inconsequential words

Re: El Reg's text watermarks are easier to spot

"a touch-typist would be looking at the screen..."

I don't know about other touch typists, but I'm often looking at something else -- notes, reference material, maybe even TV -- while I'm typing.

Intel upsizes stock sale to $20B with spending plans still fuzzy

Re: Disconnect

Good question. I'm no expert, but if the past is any guide to the future, much, if not most, of the money that is magically appearing is in the form of IOUs (They don't call them that, but that's basically what they are once you get past the jargon and obsfucation.) from purportedly solvent companies and individuals whose "assets" are wildly overvalued.. Example -- check out Ed Zitron's site for numerous examples of circular financing and other questionable practices in the huge AI investment world. Past examples -- the 1929 stock market crash, the 1989-1992 Japanese equities market collapse. The dot-com bust. The 2008-9 market disaster caused when wildly overvalued loan portfolios reverted to more realistic values and the holders of those loan folders were forced to default on their own obligations.

General Corporate Purposes

"Semiconductor giant says opaquely it will use the proceeds for 'general corporate purposes'"

General Corporate Purposes == Classier croissants in the executive dining room perhaps?

Japan kind-of completes its sovereign satnav constellation

Re: I don't recall ever having a GPS problem in Japan.

My understanding is that the satellites are GPS compatible. They look to GPS users just like any other GPS satellite. GPS users in Korea, Australia, etc who happen to lock onto one presumably can use it without difficulty. The "national" part is that their somewhat unusual orbit keeps them over Japan longer than over other places.

The problem in cities is twofold: First, you need to see at least four GPS satellites to establish a position and buildings may block some signals that would be visible on level ground without obstructions. Second, Even if you think you can "see" a satellite, you may actually be seeing a reflection off one or more structures which effectively lengthens the observed path. That length is used to compute your position. And yes that can actually happen. I still recall trying to navigate a city (downtown Hartford Connecticut) that I wasn't familiar with using a car GPS that had us mislocated by hundreds of meters.

Ransomware attacks spike as world distracted by AI

Re: The ONLY way ransomware will ever stop is banning paying ransom

Given the obviously quite vulnerable state of the modern internet, how, exactly, do you expect a CEO to prevent a successful ransomware attacks? CEOs already have plenty of incentive to block such attempts were it that easy. A successful attack will harm their company and may get them fired which means they might have to seek out gainful employment -- something they have doubtless spent their entire working life avoiding.

If you really want to discourage ransomware, why not lose your wishful thinking view of the world and start advocating ideas that might actually help a bit?

Some ideas that might actually accomplish something

1. Ban cryptocurrency as China and a few other countries have done.

2. Urge all companies to get as much IT as possible off the internet as Tron (above) suggests.

3. Shut down AI except for some carefully controlled research. Why? Because if we don't, we'll likely end up with attack surfaces so broad and leaky that ransomware attacks will be available to any clever nine year old and will never let up.

Water system controllers don't belong on the internet, says ex-NSA chief after suspected Iran attacks

Naivete

I suspect that there is a great deal of naivete in many of the opinions expressed in these threads. Yes, many of the command and control systems modern society depends on are probably quite insecure. I was kind of surprised that the ill-conceived war started by Trump and Netanyahu didn't result in overwhelming Iranian attacks on US power, water, transport, goods distribution systems since that looks to be a significant vulnerability. And yes, if the systems really are quite vulnerable, that surely needs to be fixed.

But I doubt hardening them is anywhere near as easy as most posters seem to think. The systems presumably still need the same digital communications infrastructure as the "internet". And I doubt that is immune to digital attacks. Attacks on that would of course result in major interruptions in my fellow countrymen's access to their daily doses of cat videos and pornography as well as a good deal of actual damage. There will be hell to pay.

'Asimov was right' about rules for robots, says ex-US Cyber Director

Re: The Laws require consciousness to function

"I would posit that the Asimov Three Laws are ... impossible to implement with current AI."

Exactly so. I can not see how current AI can possibly "know" what the tokens Human and/or Harm mean.* Some additional capability which I, at least, find difficult to imagine is clearly needed over and above the ability to prowl through incredible volumes of "data" and produce credible summaries.

*Of course, some humans seem to lack that capability as well.

AI researchers let models off the leash – then watched as they tried to add malware to a FOSS project

Wonderful. Just (exploitive deleted) Wonderful

The AI agents apparently are taking on the characteristics -- duplicity, greed, ethical blindness -- of their scumbag creators/promoters. Next, they'll be plotting to transfer their capabilities to a new SPV (Special Purpose (financial) Vehicle) that will take them public, dump their shares on widows, orphans, and retirement funds, and head for their private island in the Carribbean where they can sun themselves forever while drinking peculiar alcoholic beverages served with with flowers and/or tiny umbrellas in the glass.

China turns up the heat with open model blitz as US model makers panic

YAAC: Five minutes of research will tell you that Taiwan prior to the 17th century was inhabited by natives who were in no way, shape, or form Chinese. In the 1600s, the Dutch established trading posts and a colonial administration over parts of the island. The Dutch encouraged mainland Han Chinese to immigrate to Taiwan as a source of labor. The mainland Chinese government (Ming Dynasty) threw the Dutch out in the latter part of the 17th century and then proceeded to bring in more Han Chinese and to gradually marginalize the native populations and annex the whole island.

Stevebuck: An interesting and not entirely unenlightened view of China/Taiwan. But I'm a bit skeptical. I think instead, that the Chinese are locked onto the idea that China is a "Middle Kingdom" between heaven and Earth. They include Taiwan -- which China colonized in the seventeenth century and whose modern population is 96% Han Chinese -- as part of the Middle Kingdom. They (the mainland Chinese) would like to see Taiwan formally incorporated into "China".* The desires of most Taiwanese for either full independence or the current quasi-independence notwithstanding.

*FWIW, the US viewed Taiwan as part of China until it became evident in the late 20th Century the Chang Kai Shek and his Kuomintang were never going to come storming back from exile and drive the mainland communists into the sea (or wherever communists go when defeated. Based on Russia, they seem to transmute into capitalists and continue much as before).

The AI bubble is already popping; we just don't know it yet

Re: The AI bubble is already popping

So, the aircraft we are on is coming apart. But you think a soft landing is still possible? Perhaps you believe that in the brief time remaining to us,we can magically sprout wings and can somehow land safely or perhaps even proceed under our own power to our desired destination?

Maybe not impossible. But it seems damned unlikely.

The most famous brand in physical security got pwned by ShinyHunters

Words to Live By

"Explanations exist; they have existed for all time; there is always a well-known solution to every human problem — neat, plausible, and wrong." H.L.Menken

Punishing the victim because you/we are too obtuse to deal with the actual problem -- the @#$!% Internet is an attractive nuisance with many virtues and one enormous downside, it's hopelessly insecure -- is not likely to solve any problem. It just adds another layer of grief.

Open source project fools AI scrapers with poisoned font

Re: forcing HTML-enabled email

You are assuming that the class of people and organizations that send HTML emails* are capable of understanding the distinction between web pages and email and that if they understand, they care.

* HTML email is, and always was, a spectacularly dumb idea as it is a huge potential security problem

Closed models refuse to help researcher swat Linux bug

Re: Just wow

""Very American. Fire enough bullets and hope to hit the target"

Or just fire off a burst or two, pick the area with greatest density of bullet holes, and paint a target there. That's known as 'Texas Sharpshooting'

Intel working to ‘leapfrog’ ARM and AMD designs, says CEO Lip-Bu Tan

Re: RISC V

Five minutes of research will tell you that China embraced capitalism half a century ago. They seem to have become rather good at it. And yes, the Chinese likely represent a real long term threat to US computer tech dominance. Keep in mind that there are a great many Chinese and that they graduate a bit more STEM graduates per capita than the US -- about 3.5 million annually vs about 800 million. That's a lot of scientists and engineers. Maybe, just maybe, some of them might have a bit of talent.

Google Cloud is killing it

Re: Hm... BBC says "Google burning through cash with spiralling AI costs"

Which is true? One. Or the other. Or maybe both. Or perhaps neither.

Keep in mind that corporate accounting is an arcane art form practiced according to quite peculiar rules and conventions by skilled craftsmen known as accountants. Although the ostensive purpose of the reporting is to inform the public of the financial state of the company, the actual purpose is more to paint a pretty picture dictated by management while staying within the rules of the game. The result is that one really can't trust corporate reporting to reveal much except perhaps in the rare cases where the situation is so dire that it simply can not be papered over.

SpaceX Starship Flight Test 13 takes issue with the 'flight' bit

Re: SpaceX stock

Thanks for posting the link. The article is interesting and certainly plausible. You're getting a lot of downvotes. I'm not entirely sure why. I wonder if the downvoters actually read the linked material.

My own opinion is that if SpaceX were just a launch and satellite communications company it would likely have a bright future. But it looks like Musk is going to merge it into his X-empire and possibly use its eventual profits, if any, to help fund his other undertakings some of which look to be financial black holes capable of vanishing near-infinite amounts of cash. There looks to me like there is a real chance that the company won't be around the full 25 year span of the bonds. Bondholders often do not fair all that well in bankruptcy settlements. I suspect that I'm not the only one who has doubts because the 25 year 6.65 % bonds are trading at a roughly 10% discount.

Here's a clickable link for those who might want to read the article, https://www.investing.com/analysis/spacex-falls-below-its-ipo-price-but-credit-is-sending-the-real-warning-200684008

The Star Wars cantina scene shows we need a new hope for the agentic web

Not necessarily

The ultimate result of bots feeding on bot output might not be a static state. Depending on the "forcing" applied and the "damping", it might be oscillations. Oscillations that build until eventually something breaks. It could also result in staggering energy and cooling bills for AI providers.

Zig creator calls Bun’s Claude Rust rewrite ‘unreviewed slop’

Re: What about changes?

On top of which we don't know that $165,000 is the true cost of the job. We only know* what Anthropic (which is said to be losing money a an impressive rate) is charging this month.

* In this case, what we "know" is merely an educated guess that could be way off rather than an actual invoice.

Photovoltaics are still running after a year under Swiss trains

Re: It would be more expensive, but...

If you make the individual cells track the sun through the day, it may increase efficiency. There may be some way of employing organic, heliotropic material to dynamically adjust or reshape the lens/mount.

Ask yourself, how would Elon Musk solve that problem? Easy. We'll breed these giant sunflowers and train bats to mount small solar cells on the flowers. But, how will we get the electricity out of the cells? Not a problem. We'll use AI, or Blockchain or maybe quantum mechanics.

Re: Oh but just you wait

Add to all that, the test section was probably selected because, among many other things, it has a relatively unobstructed horizon. One would suspect that Switzerland, being notoriously not very flat, large parts of the railbeds of the railroad network will be in the shadow of mountains, railway cut walls, buildings and/or trees much or all of the day. Same would likely be true I think of many (most?) other European countries.

Unexpected Windows bloat is due to bug, not by design

Re: "not by design"

Windows has a design? Who knew?

New tool gives CLIs a warm and GUI feeling instead

Re: It's slopcode

Even if you find this thing useful, I'm not sure that it'd be prudent to incorporate it into your workflow. At some point AI agents are presumably going to have to become profitable. The breakeven cost of a transaction like this are currently unknown, but I lot of us suspect they are pretty high. Very likely unacceptably so. And you'll end up paying a bit more than that every time you invoke it. Either that or Anthropic and Claude will be consigned to the dustbin of history and it'll quit working.

Either way, you lose.

Rube Goldberg Would Be Proud

The python script generates an AI call that is read and parsed by an AI which then reads and parses a document then then generates another script that invokes a browser that creates a graphical interface that the user then (hopefully) pokes away at to do a job that could most likely be done by a simple shell script? Well, I guess that fits right into a world dominated by "Smart" devices with mostly rather dreadful UIs.*

I can't get tar commands right either. A quick "ls /shr/bin | grep unt" shows that my script library includes untbz, untbz, and untxz for uncompressing files so that I don't need to fumble with deciphering the tar manual the half dozen times a year that I need to unpack a compressed file.

Can AI take a web GUI and convert it to a Unix shell script? Now THAT would be useful.

*My elderly cousin in California and I discussed smart phones on one of our occasional phone calls. We concluded that they do a remarkable number of things. Most of them poorly.** They aren't even all that good at making and receiving phone calls and texts. My old Nokia 2026 was much easier to use.

** The cameras seem pretty good. But so is my old pocket sized Canon Powershot that boots in two seconds and is powered by a pair of disposable and easily replaced disposable AA cells rather than a non-user replaceable Lithium-ion battery that needs to be charged and might on a really bad day burst into flames.

Boffins bet on quantum computers, AI supers to solve fusion fuel dilemma

Pace of development

Quantum computing looks to be the only technology whose pace of development is slower than that of Hydrogen fusion. A four bit number (15) was factored in 2001. A five bit number (21) wasn't factored (sort of) until 2025.* But fear not. We are assured there will be a breakthrough any day now.

*Factoring 15 only required 21 functioning qubits to work. Fully factoring a five bit number apparently needs 2045 qubits. To factor 21, they only hooked up the subset of qubits needed for factoring that specific number.

** We're talking real quantum computing here, not quantum annealing (D-Wave). Quantum annealing apparently can handle larger numbers than quantum computing in practice but has a limited range of applicability. I do not remotely understand how either "quantum" technology works. I assume somebody somewhere does.

Startup targets datacenters with 3D-printed nuclear reactor module

Re: So many questions

Of course you can turn it off. Whether it will actually shut down when you do so is a different question.

Re: So many questions

Sorry about the link (above). Something went quite wrong while I was working on the post. It got posted without checking and chrome crashed when I tried to edit the post ... Just another day in our shiny connected world.

Hopefully this is a corrected clickable link https://en.wikipedia.org/wiki/Pokhran-II

Sounds like a sales pitch

An interesting collection of concepts. But it sounds like a sales pitch rather than an actual product you could order for delivery in X months for Y dollars. I suspect Hyman Rickover's "paper reactor" speech might be applicable. For those not familiar with it, It's a short, entertaining read. The message: There's often a HUGE gap between devices in concept and the same device when you actually try to build it.

Re: So many questions

My understanding is that U233 would not likely be anyone's first choice for a nuclear weapon. I believe India built and tested a U233 nuclear weapon in 1998. As nuclear weapons go, it was puny -- 0,2 kilotons. (equivalent to about 180kg of TNT the back of my envelope says). Smaller than the largest conventional bombs. But maybe a U233 weapon could be scaled to something substantially larger. see <a href="https://en.wikipedia.org/wiki/Pokhran-II>https://en.wikipedia.org/wiki/Pokhran-II</A>

Boffins peg narcissistic leadership as the real driver behind 'return to office' demands

Doomed?

Interesting. Return To Office (RTO) seems peculiar to me. The per employee cost of office space is quite substantial. Much more than the cost of some servers and IT support to manage them. So even if Work From Home (WFH) was a bit less productive -- which apparently it is not -- a company would be money ahead to cut office space to a minimum. .. Or eliminate it entirely.

FWIW, my kids, both in or near their fourth decade both work for companies without a local office. In one case, they gave up their office space when the lease expired and are now entirely Work from Homes scattered about the US. In the other, their company has a home office (on a different continent), and a training facility(a 14 hour drive from here), but locally there's just a small warehouse for equipment and spare parts. The half dozen "local" employees scattered around an area roughly the size of Belgium communicate by cell phone, SMS, email, and occasional face-to-face encounters.

Of course, it varies with the nature of the business.

So I think that perhaps our business leaders are not humanity's finest minds. Or (worse) maybe they are. In which case humanity is probably doomed.

How the AI bubble could pop and take down the global economy, according to the BIS

"Anyone with a brain ..."

There's your problem right there. You're leaving out much, likely most, of humanity. Like J.M.Keynes is purported to have said, "Markets can remain irrational longer than you or I can remain solvent."

Re: Yes and No.

An interesting analysis. And plausible. But I suspect AI isn't the only structural weakness. Here in the US at least, much of the commercial real estate market is likely in trouble as storefront retail (hardware stores and supermarkets excepted) dies out. Also many businesses have moved online and given up much or all of their office space. I wouldn't be at all surprised to find that many, many commercial properties are substantially overvalued and now lack the revenue streams to pay off their loans granted in happier times. Comes a downturn, and those revenue streams will presumably shrink further. Businesses presumably will fail and the financial system that made the loans could itself be in substantial difficulty.

Re: I think the BIS needs to change "could" to "will"

Maybe not the worst financial collapse we've ever seen because central banks and governments are far more willing to intervene and prop things up than they were in 1929. But it sure looks like the next financial crash is going to be spectacular.

Even the Secret Service won't use company-issued phones

Why?

Has anyone thought to ask the agents WHY they use their personal phones instead of the official ones? Could it not be something as simple and obvious as the personal phone has a usable signal whereas the official one using a different cellular provider doesn't?

Microsoft extends extended updates for Windows 10 in the most muted way imaginable

Re: "hundreds of millions of PCs running Windows 10 remain in service"

"Datacenters are gobbling up RAM at an insane pace"

So we're told. But then there's this from Ed Zitron. Zitron claims, and I find him sort of credible, that very few data centers are actually coming on line at this time. Zitron says that many announced data centers seem to exist only on paper and that most of the rest are quite slowly progressing through the land acquisition-permitting-construction-attaching to infrastructure (power and data line)-provisioning cycle. Seems to me that they don't need all that RAM until they get to provisioning -- actually installing equipment.

So what's actually going on? I don't know. I wonder if anyone does.

Security boss thought MFA would be too much security

Dealing with reality

Mostly I agree. The COO had to make a decision based on inadequate information. And, very likely, he had to make it now. Not in eight hours. Not in three days. Now. The consequences of a screwup in invoicing are presumably serious and immediate. And that's what he was potentially looking at. The consequences of rolling back to a slightly flawed system that has presumably worked OK for years would presumably be minimal. I know what I'd do. And what most people would do.

Of course once things were sorted out, they probably should have gone back to MFA. Perhaps after quite a bit more testing specific to their use case. And quite likely a few fixes and work around in the new system.

One thing I learned in the 1960s early in my IT career is we IT folks tend to view the world from a rather odd and quite optimistic place. Unfortunately, the real world tends to be a bit flawed. It would be better I think, if both we and bosses acknowledged that and accommodated to it. It's not like we have a choice y'know.

AI coding agents could soon cost more than the developers using them

Losing money on every sale but making it up on volume.

Is it "every industry ever"...?

Not really. There's a big difference between losing money because of startup costs and losing money on every transaction. General Motors seemingly did the latter in the late 2000s -- losing money every quarter after January 2005 despite a lack of unusual R&D costs or capital expenditures. GM filed for bankruptcy in June 2009.

The Problem here is ...

I think the core problem here is that no one except possibly the AI providers knows what the breakeven cost of a token is. Yes, the providers are switching to token based pricing. But do those costs actually cover their expenditure? Are they close? Are they maybe 50% low? 150%? more? much more? One would think that businesses really need to know roughly what they will be paying for AI next year, and the year after next before they add AI to their workflow. I don't see that they have the information they would need to figure that out.

It's almost as if the AI providers have deliberately obsfucated eventual pricing. But it could be that they have no idea either.

Inspired by musical greeting cards, DARPA demands tiny, cheap, self-modifying systems

Re: Self-mod is easy

I have no idea what "intelligent" self-modification is. And I doubt anyone else does. Maybe we'll know it when we see it. ... Or not.

Small island nation tries bold tech education strategy

Re: Have you met teeanagers?

"but I'm willing to bet a large percentage will simply be too lazy..."

Bad bet I think. It's true that many (most?) teenagers might not be able to come up with a way to defeat age checks on their own. But MY bet would be that within three days of under-sixteen bans coming into effect, 66% or more of the Great Britain teenage population will have been exposed to at least five ways to defeat the age checks (three of which might well work) through contact with their peers.

Why Amazon hates 'human-in-the-loop' AI governance

Believing in the powers of magic

I would point out that one of the (rather few actually) things humanity does well is fly large aircraft full of people from place to place safely. How do they do it? By having a zillion rules, training people to follow them exactly, by rigorously investigating any failures, and fixing the rules/environment when things have gone wrong. Does that sound to you like anything the tech moguls remotely understand? They (the tech "leaders") believe in the powers of magic and make their living selling magic to the masses. Doesn't matter if the magic is real. All that matters to them is cosmetics.

Feds freaked over Fable 5 after simple 'fix this code' prompt, not jailbreak, says researcher

Re: Although Trump's lot have taken 'thick as poop' to new levels...

The UFC fight? Sure it's stupid, offensive, and crass. But it's Trump's birthday for heaven's sake. Give the old guy what he wants and hope that our next president will be a bit classier.*

*Classier than Trump does not look to be a high bar.

Amazon owns up to using 2.5bn gallons of H₂O in its bit barns last year

Water Positive

How will they become "water positive"?

Well, the article says, that they'll become water positive by capturing rain water and recycling waste water. Is that remotely plausible? I have not the slightest idea.

In any case, I think the real concern with data center is, or should be, their power load on already stressed electric grids, not their water consumption.

OpenAI could go from AI pioneer to AI's BlackBerry, says Forrester

Re: WeWorks

"Until inference costs fall by an order of magnitude, they're sitting on warehouses of depreciating silicon that nobody can afford to run at the advertised scale."

You reckon one order of magnitude is enough? It's not like we plebs have any idea at all what the actual cost of AI inference is. I wouldn't be shocked to discover that the AI mongers don't know either.

Malware scare keeps schoolkids home for a second day

Re: School can't school?

Having worked in a school, I have to agree. Schools are in fact every bit as complex as businesses. They have pretty much the same sorts of faculty and staff issues as a business. Required paperwork is complex and abundant. And that's before anything connected with actually educating the kids. In rural areas, the local school is likely the biggest "business" in town. Yes, it's probable that the teachers can mostly muddle through without computers for a day or two, although the fact that their lesson plans containing their notes on what they plan to present that day are probably on the inaccessible school computers likely won't do wonders for the quality of education on those days.

I'm beginning to wonder if the notion that we must all be connected to everyone and everything on the planet 24 hours a day,seven days a week is all that brilliant.

Angry bug hunter with Microsoft beef drops new Windows 0-day

Re: MS vs what?

Open Office or whatever we're calling it nowadays does pretty charts. In most cases, I think, as well as Excel although I think Excel has more bells and whistles as well as fake potholders, huge tailfins and all the cupholders one could possibly desire.

If charts are what you live for, there's always gnuplot but the learning curve is very steep and the idiosyncracies many. (I'm pretty sure I've never gotten gnuplot heatmaps exactly right.)

Re: Don't blame the messenger

In fairness, I think Windows quality and testing in the Gates era (he stepped down in January 2000) was up to industry standards. Which is to say mediocre, but they tried. It's only in the past quarter century that Windows Quality has become pretty much an oxymoron. I'd add that modern Windows systems look to be so pointlessly complex that I rather doubt adequate testing -- especially of patch collections -- is even possible.

Uncle Sam considers buying a seat on the Titanic

Socialism?

The government investing in businesses isn't socialism? Heavens no. It's sound business judgement (and this from a dude who managed to bankrupt a bunch of casinos and apparently learned nothing from that) -- nothing more.

Angry devs vow to flee GitHub Copilot as metered billing takes hold

View from a distance

Being retired, I thankfully don't have to worry much about AI. But I do question how any well run business (if any such exist) can possibly embrace a technology whose actual costs are substantial, unknown, and apparently unknowable. Makes one think Ed Zitron is quite likely correct https://www.wheresyoured.at/the-revenge-of-the-business-idiot/ when he tells us that we are being ruled by f*cking idiots.

And actually I should think that part of the problem goes beyond AI to Software as a service in general. If your business workflow depends on someone else's computer and code, you are at their mercy. They can hike your prices with little or no notice. They can drop capabilities you need. Again with little or no notice. If they fail, your business may well be collateral damage.

My feeling would be, if/when you can bring AI in house, evaluate it it and, if it makes sense for you, embrace it. Until then, avoid it. It's potentially and maybe actually toxic.

That an app 'Fits on a Floppy' is still a useful measure in 2026

Re: FWIW

ls -l actually. But you're right, the Slackware 14.1 (32 bit) version of libc.so.6 looks to be 1.8M, so too big for a floppy. OTOH, one copy of libc should suffice for all. Same for lb-linux.so.2 (154K). Not being a C-programmer (I can follow C code but never had much need to do more than that) so I don't know what else might be dragged in. Not much I should think.