Dolphins, sharks, turtles, and human workers are all victims of unregulated squid fishing fleets.
Another news article.
As usual, you can also use this squid post to talk about the security stories in the news that I haven’t covered.
robin • June 20, 2026 3:07 AM
@Clive: you quote: ““Of the 45 citations in the report, only five accurately point to real sources,” the team wrote, adding that many others were either totally false or significantly distorted.“”
I can’t see what the problem is: that report really does redefine excellence in the age of agentic AI, as promised in the title. Although perhaps not in the way KPMG intended.
It would be nice to think of it as a Ratner moment, but I expect it will be air-brushed out. I see Ernst and Young also had a somewhat similar foot in mouth moment.
Given that the KPMG report seems to have been on their website for nearly 9 months it must be assumed that those false citations are now treated as gospel by LLMs even though it’s been taken down now.
Clive Robinson • June 20, 2026 7:25 AM
@ Robin, ALL,
You note that,
“Although perhaps not in the way KPMG intended.”
But as you also note,
“It would be nice to think of it as a Ratner moment, but I expect it will be air-brushed out. I see Ernst and Young also had a somewhat similar foot in mouth moment.”
Mr Ratner[1] is less well known these days, and in reality,
“His sin was telling the truth”.
Which begs the question of the difference between telling the truth and facing the consequences or just abusing the truth for profit?
What is clear though, is that using Current AI LLMs to generate research, or law citations is most certainly,
“Not a way to increase employee productivity”
Due to spending all that time checking every word looking for “Vibed Soft Bullshit”.
I saw an argument the other day that “Vibe work” must have ROI because it “produces more in less time”…
What they failed to take into account is that they only considered in effect “lines of code” being “good” and did not consider any being “bad” thus the enormity of “technical debt” that would be produced, that eventually has to “pay off” in all sorts of bad ways.
These “consultancy Corps” appear not to consider what you might call “Negative ROI” of “polishing turds” that LLMs pop out all the time.
[1] He was a speaker at a business confrence in London’s Albert Hall when he made a joke about the family company he had built up in the US and UK,
Like all jokes it had a germ of truth in it. The thing is most modern products made in China for the Western Markets are today built under the same,
“Make it cheap, pile it high, and knock it out fast”
Idea which has given us the Internet of Things, Amazon home and security products, and even those who had good reputations throwing security away on small network edge products…
But operating in that business area takes considerable skill and thus those that took over his businesses failed to make a success, so it appears last year Mr Ratner is aiming to take the business back before it finally sinks.
Speaking of which Cisco appears to have slipped into the “keep bailing” with respect to it’s SoHo Network products.
JG5 • June 20, 2026 12:15 PM
The Alien Probe technology is here now on Uranus. It just isn’t evenly distributed. It never can be, because reality is a non-linear, time-varying, emergent, non-equilibrium process.
I hope that I posted something about the LLMs being part of a giant scam. The data centers are very real and built for processing the data from all of the surveillance cameras, as well all the data swept up from computers and cell phones. Including, keystrokes, audio, and video. All of the talk of the LLMs causing mass layoffs is an intentional distraction from the real issues in play. The price of electricity will force a painful reckoning. The US tried to do too much with too little. In many areas of endeavor.
https://x.com/BabyD1111229/status/2067748917685662197
…
Every time …, the sensor sweeps up the unique electronic identifiers of every device in your vehicle. Your cell phone. Your smartwatch. Your wireless headphones. Your fitness tracker. Your laptop. Your tablet. Your car’s own infotainment system. Your tire pressure sensors. Your vehicle’s Bluetooth hotspot.
…
Clive Robinson • June 20, 2026 12:41 PM
@ Bruce, Lurker, ALL,
Deliberate GPS degradation in Europe worse than thought
Some of us keep an eye on radio and space based navigation and time systems as they have serious security issues not just for tangible physical objects but intangible information objects. Thus they form foundational critical infrastructure.
But… to many it’s just a box that means they don’t have to read a map or know how to set their watch…
However they should consider that aircraft can now be landed by GPS systems at commercial airports, with some pundits saying the need for commercial pilots can be reduced (thus savings for the shareholders etc).
So maintaining GPS systems free from jamming and spoofing has become of significant importance and criticality.
Which is why this is “bad news”,
‘It’s quite a bit more than we expected’: Satellite reveals immense scale of GPS signal tampering
“As soon as we started doing any operations above Europe, we noticed that there was really something going on there.”
“An experimental satellite has mapped the scale of GPS jamming across Europe and the Middle East from space for the first time.
The data surprised the team behind the project and indicated that satellites orbiting far from Earth aren’t the only ones that experience degradation of their positioning, navigation and timing (PNT) signals, which could affect their performance and the safety of their operations.“
Hopefully new types of autonomous “Positioning, Navigation and Timing”(PNT) system currently under development will become commercially available in the near future, otherwise life is going to get more than a little awkward.
Clive Robinson • June 20, 2026 1:05 PM
@ JG5, ALL,
With regards,
“The data centers are very real and built for processing the data from all of the surveillance cameras, as well all the data swept up from computers and cell phones. Including, keystrokes, audio, and video.”
The idea behind some of it, is to,
“Fingerprint the citizen every which way”
And by “fingerprint” I don’t just mean standard bio-metrics.
The state of your physical and mental health can be found from the amalgamation of biometrics.
Whilst we’ve known this for some time, few have asked what advantage such information would be to the Government and various forms of “Guard Labour”…
Just remember in the US way way to many people believe in the efficacy of “lie-detector” technology. Where the actual reality is it’s about as much use as “flipping a coin” and many people with a little practice can fake the results.
Thus one idea is to have “lie-detection” running 24×365 on every one as a form of “pre-crime detection”.
It won’t work even remotely well but it will be “made and marketed” to “Guard Labour Agencies” thus enabling them to use personnel who don’t/can’t think beyond the immediate and thus moderate their behaviours, but will “snap to” on any order no matter how unlawful.
It’s the sort of thing people like Peter Thiel and Co of Palantir and similar in Amazon and Meta have more than wet dreams over. And they care not a jot for justice or due process only pushing “might is right” ideology for control and profit.
Eff Bezos • June 20, 2026 3:05 PM
Amazon was sued on Monday over alleged privacy violations from its Ring doorbell cameras. The class action lawsuit claims that Ring’s Familiar Faces feature stores images of passersby without consent.
Familiar Faces lets Ring users identify people who regularly come to their home through AI facial recognition. That way, if a regular guest, like a family member, mail carrier, or neighbor, comes to the door, the device will be able to recognize them and deliver more specific notifications like “Dad is at the door,” rather than “A person is at the door.” Ring users have to opt in to this feature, but privacy advocates noted that the people who walk past these Ring doorbells have not consented to these facial-recognition scans. That same concern is at the center of this class action lawsuit.
According to the lawsuit, “Millions of other Americans passed by a Ring security camera and unknowingly had their facial recognition information collected.”
lurker • June 20, 2026 8:39 PM
@Eff Bezos
The “class” in this case is presumably the millions of John Does who have had their faces scanned and put into a database somewhere. What remedy can any judge offer to these John Does? Post facto?
Amazon claim the data is encrypted, never shared, and unhnown John Does are deleted after 30 days. But with the admission that Flock Safety were in the neighborhood, 30 days is more than enough time for skullduggery …
Clive Robinson • June 21, 2026 8:05 AM
@ Bruce, ALL,
AI makes “Cognitive Debt” the new and worse tsunami in IT and ROI is not happening.
Hands up those who did not see this coming?
CTOs Agree: Cognitive Debt Is the New Technical Debt
“ …
The free-for-all is over
Two years ago, the mandate was simple: spend on AI, no questions asked. That era is over.
What’s replaced it is a harder conversation, one several participants had clearly been having with their CFOs. The question has shifted from are you using AI to what are you getting for it?
“
…
The feature debt problem nobody wants to talk about
AI makes it cheap to write code. That is not the same as it being cheap to ship it, or to maintain it. One participant put it cleanly:
cognitive debt is the new technical debt.
https://shiftmag.dev/ctos-agree-cognitive-debt-is-the-new-technical-debt-10229/
Both of these were “obviously going to happen” and not just in ICT.
In fact as obvious as the,
“OMG think of the job losses”
That many in the MSM etc have not consulted history about.
As some will know what concerns my most is the “debt” by it cognitive or technical.
Because either way security will rest on those who work lower down the stack be it computing or cognitive.
The ability to produce fast code makes it a “quantity over quality” issue and as I’ve pointed out for years,
“Security is a Quality Process.”
Just for that alone I see an increase in the need for Security Practitioners as AI digs in deeper with “vibe development”.
The other thing to realise is that the term “Engineer” is again being abused.
What is currently “Vibe Coding” is at best “artisanal behaviour” and easily down in the hand waving category. Thus there needs to be a lot of improvment.
Because those who produce secure products need to take traditional engineering practices on board.
Calling what most software developers do “engineering” even before “Vibe Coding” came along is to be blunt an insult to those who actually engineer safe and secure tangible physical products.
The problem is actual “engineering costs” in time, knowledge, experience and resources, none of which manglement want to take on. Hence the idiocy of
“Move fast and break things”
That originated from the psychos in Facebook etc.
Think about that attitude if taken in say aerospace engineering. Aircraft would be dropping out of the sky faster than drones in the Ukraine.
But we’ve already seen some of it with Boeing management and the hundreds of deaths that has caused.
Currently –even before AI– software development was more “art” than even “craft” and it still shows with CVE numbers which we now know are only a small percentage of vulnerability issues.
Weather • June 21, 2026 8:23 AM
@All
Microsoft recycling bin bug, after 12 years someone rediscover it.
Pretty much you make a file with charcters 0x01-0x7f, each character has a value between 1-7, when the counter gets to 0xfffffe you insert 0x80-0x85, then put your payload after that. When they try to delete the 350mb file approx, it causes a crash.
Clive Robinson • June 21, 2026 8:28 AM
Hmm…
It appears that the blog software has barfed again with “blockquotes”
In my above it appears with the second set of “triple dots”.
It killed not only the “blockquotes” but “in italics” as well.
Sorry to folks that are trying to read it.
The quotes from the article finish before the URL to the article, the comments after the URL are mine.
Clive Robinson • June 21, 2026 8:37 AM
@ Weather,
You did not provide a link.
Is this the bug you are talking about,
Rontea • June 21, 2026 10:55 AM
@Eff Bezos
Ah, what a curious spectacle of modernity! We place an unblinking eye at the threshold of our homes, a silent sentinel proclaiming the gospel of convenience. And yet, like all idols of progress, it feeds upon the very essence of our liberty. Here in this so-called digital Eden, man is no longer a soul, but a pattern to be recognized, a face to be cataloged. We walk the streets of our own cities, strangers in a world where even our passing glance becomes property. And then we ask—did we not build this altar ourselves?
Weather • June 21, 2026 3:25 PM
@Clive
Sorry no, i didn’t read the article i was talking about a BoF.
Weather • June 21, 2026 3:42 PM
Try works on old ones
#include
Main()
File *out
Unsigned int i,t
For(t=0;t<(0xffffffff/2);t++)
Fprintf(out,%c,0x80)
Fprintf(out,%c,0x80)
For(t=0;t<50;t++)
Fprintf(out,%c,A)
Exit.
Weather • June 21, 2026 8:51 PM
The code is easier if you look at notepad. The bin has a limit were it has to be deleted instead of temporarily stored.
It might be why they are getting strange extensions on files.
Sten • June 22, 2026 4:17 AM
Squidbleed should be mentioned in this thread.
It’s a a Heartbleed-style vulnerability that leaks internal memory from every version of Squid Proxy, in its default configuration.
https://blog.calif.io/p/squidbleed-cve-2026-47729
Anonymous • June 22, 2026 3:24 PM
I Stored a Website in a Favicon
https://www.timwehrle.de/blog/i-stored-a-website-in-a-favicon/
lurker • June 22, 2026 8:17 PM
AI models capable of devastating attacks on governments and business months away, rare Five Eyes statement warns
Signal agencies in Australia, the US, the UK, New Zealand and Canada sound alarm after Trump blocks foreign nationals from Anthropic’s Fable AI model
Hit back wih AI, is it the problem and the answer?
https://www.abc.net.au/news/2026-06-23/five-eyes-security-agencies-say-act-swiftly-on-ai/106828686
too many requests • June 22, 2026 8:46 PM
Microsoft discovers new lightweight backdoor that steals cryptocurrency
Weather • June 22, 2026 11:11 PM
@lurker
No, its more dimension than black and white. Maybe more covering there IT sector from within.
lurker • June 23, 2026 2:19 AM
@Clive Robinson, ALL
Paper, paper, always paper …
Altho’ it’s disappointing to see MS Excel amongst their offline tools.
Clive Robinson • June 23, 2026 6:01 AM
@ Bruce, ALL,
Auz votes to change it’s equivalent to the US Patriot Act
9/11 caused a lot of political termoil and many “emergency Acts” were pulled into being in the process.
Much of such acts were in effect “cobbled together” bits of legislation that National level “Guard Labour” agencies / entities had been trying to get for decades before… But had had them rightfully rejected on simple democratic basis of “The people will not stand for this”.
The 9/11 attack was blown up in the world wide MSM even before it had ended. And unfortunately this enabled these non democratic laws to be voted in without scrutiny.
Some however had the sense to add “sunset clauses” such that the respective “parliamentary houses” would have to vote to keep them in place.
The Australian Labour Government wanted not just to keep it all in place but to add to it. However coordinated action by the other parties caused Labour to in part backdown,
Labor scraps plan to make spy agency’s 9/11-era questioning powers permanent
But Australian government will expand offences covered by rules to include promotion of communal violence and attacks on defence system
“Labor has quietly backed down on moves to make spy agency Asio’s powers for compulsory questioning permanent, but will expand offences covered by the rules to include promotion of communal violence and attacks on Australia’s defence system.
The laws were introduced after the 11 September 2001 terrorist attacks in the US and give intelligence operatives powers to issue a questioning warrant requiring a person as young as 14 to give information or produce items that may assist in a serious investigation.
Labor argued changes to the laws are justified by the “dynamic, diverse and degraded” security environment facing Australia today. The government wanted sabotage, promotion of communal violence and serious threats to Australia’s territorial and border integrity to be covered by the rules.“
This unfortunately allowed Labour to get some of what it wanted but the removal of the all important “sunset clauses” did not happen.
Thus hopefully next time around democratic attitudes will prevail over the very authoritarian desires of “Guard Labour” entities.
Clive Robinson • June 23, 2026 6:18 AM
@ Bruce, ALL,
Australia votes to change it’s equivalent to parts of the US Patriot Act
9/11 caused a lot of political termoil world wide, and many Governments produced “emergency Acts” as though they were pulled into being from thin air.
However much of such acts were in effect “cobbled together” bits of legislation that National level “Guard Labour” agencies / entities had been trying to get for decades before… But had had them rightfully rejected on simple democratic basis of “The people will not stand for this”.
The 9/11 attack was blown up in the world wide MSM even before it had ended. And unfortunately this enabled these non democratic laws to be voted in without scrutiny.
Some Governments however had the sense to add “sunset clauses” such that the respective “parliamentary houses” would have to vote to keep such legislation in place.
However the Australian current government incumbents “Labour” wanted not just to keep it all in place but to add to it as well as “lock it all in” by removing sunset clauses. However coordinated action by the other parties caused Labour to in part to back down,
Labor scraps plan to make spy agency’s 9/11-era questioning powers permanent
But Australian government will expand offences covered by rules to include promotion of communal violence and attacks on defence system
“Labor has quietly backed down on moves to make spy agency Asio’s powers for compulsory questioning permanent, but will expand offences covered by the rules to include promotion of communal violence and attacks on Australia’s defence system.
The laws were introduced after the 11 September 2001 terrorist attacks in the US and give intelligence operatives powers to issue a questioning warrant requiring a person as young as 14 to give information or produce items that may assist in a serious investigation.
Labor argued changes to the laws are justified by the “dynamic, diverse and degraded” security environment facing Australia today. The government wanted sabotage, promotion of communal violence and serious threats to Australia’s territorial and border integrity to be covered by the rules.“
This unfortunately allowed Labour to get some of what it wanted but the removal of the all important “sunset clauses” did not happen.
Thus hopefully next time around democratic attitudes will prevail over the very authoritarian desires of “Guard Labour” entities.
Clive Robinson • June 23, 2026 9:48 AM
@ Bruce, ALL,
Cloudflare launch a new form of privacy invation idea
The idea is that your Webbrowser will attest to you being a person without intent.
It will do this by “tokens” issued by “other websites”.
The claim it’s secure is a very bad assumption. Think of the tokens as a “supply chain” and just how many ways such a chain can be abused by Cloudflare or other intermediary.
I can guaranty that somebody somewhere will use these PACTs to make all people using then neither anonymous or private…
Cloudflare teams up with big browsers to help websites tell welcome from unwelcome visitors
Makers of Chrome, Edge, Firefox back bot-fraud defense called Private Access Control Tokens(PACTs)
“Cloudflare on Monday said that it has joined with the three leading commercial browser makers to create a privacy-preserving protocol that websites can use to separate desirable web traffic from undesirable network requests.
Cloudflare, along with Google Chrome, Microsoft Edge, and Mozilla Firefox, have committed to develop Private Access Control Tokens (PACTs), a way for websites to generate a digital token that asserts a given browsing session is being run by a human or bot with legitimate intent, as opposed to network requests from people or software deemed abusive or improper.
PACTs will let websites “with strong knowledge of ‘personhood'” issue anonymous tokens that browser users and designated bots can present at other websites, so that fewer identity checks are necessary. “
A piece in favor of independent audits for AI models on politically-sensitive topics:
https://www.wsj.com/opinion/ai-needs-public-quality-testing-f18e0ebd
“We found serious gaps: The models misstated public opinion on political topics and attributed quotations to people who didn’t say them. AI models gave incorrect answers to questions about mail-in voting and ballot fraud, and named the wrong people when asked who had endorsed whom. On open-ended, contested questions that voters might ask before an election—about gerrymandering, immigration and climate—they often advocated one side.”
In a blog post, Forum AI says that more of Fable’s answers contained at least one false claim, as compared to Opus 4.7 and 4.8.
https://www.byforum.com/blog/claude-smartest-model-most-neutral
I don’t know what it would take to make a truly neutral and 100% accurate LLM, but I like the idea of open audits.
Weather • June 23, 2026 5:43 PM
@Clive
So they use the browser header, shame firefox was in there with there side party scripts.
So more they think its solid ,more they will belive when you fake it.
Why do they get paid but not me…oh yeah.
Clive Robinson • June 23, 2026 10:56 PM
@ KC, ALL,
With regards,
“I don’t know what it would take to make a truly neutral and 100% accurate LLM, but I like the idea of open audits.”
It can not be done.
You would first have to ask how you define “neutral” as most humans will assume their views are neutral even though the person standing next to them would disagree.
The short answer is even the best of us have “cognitive bias”…
Often it is due to lack of unbiased information over an extensive period of time. Especially routed in our early formative years (so formal Education / religion systems are quite often to blame). This then biases us in what we find credible and engage with, and what we don’t find credible thus don’t even start to listen/read argument to.
Similarly “life experience” also causes cognitive bias. One example is you get attacked by someone not of your ethnicity, you tend to treat people of that ethnicity unfavourably thereafter if you do not have strong counter bias. Yet studies have shown that if the attackers are of the victims ethnicity a different criteria of blame/hate is used.
The list of potential bias causes is long, thus “neutral” is very subjective.
Worse, is when one factual measure is regarded by another as an “ism”.
Basic physics applied to human physiology indicates that a certain body type has advantage in a physical activity, dependent on the activity involved. Yet point this out and you will get others loudly protesting you have an “ism”…
So “facts” upon which “accuracy” rests are disregarded for “political” reasons. Thus “accuracy” becomes at best “subjective”.
Politics is all to do with “tribalism” and the old adage of “My Country Right or Wrong” comes into play. Thus facts become,
“Subject to interpretation”
And malleable to the order and style of argument presented, and what supporting information is selected, as well as who is presenting them.
So as you can see there is more “wiggle room” capability for people to falsely argue thus,
“Neutrality and Accuracy are in many cases faux-measures based on tribalism”.
Clive Robinson • June 23, 2026 11:19 PM
@ Weather,
With regards,
“So more they think its solid ,more they will belive when you fake it.”
Yup it’s a form of “reputation by quantity” rather than “quality”
In effect the more you get past a moderator, the more you will get past other moderators.
It’s a form of “herd mentality”.
The trick for an attacker is using more than one sacrificial ID and run them against as many moderators as possible. Thus build up a chart of what the moderators will accept or not.
Then run the attack ID against the selected moderators to get an “acceptable reputation” which with care will still give you “good reputation” rather than “bad reputation”.
All such a system will do is block the idiots, whilst letting the crafty through.
The next trick for the crafty is to “move the mean” such that they move the acceptance criteria. Look on it as,
“Poisoning the well drop by drop”
The more drops you can get in, the even greater the number of drops you will be allowed to get in.
@clive,
the other day i had an errant thought,
the petitionability or corruptibility is a FEATURE of human decision trees not a vulnerability.
lurker • June 24, 2026 12:43 AM
@Clive Robinson, ALL
re Private Access Control Tokens
I use an “other” browser which makes some effort to block trackers and spyware. More and more I have noticed recently sites
“detecting unusual traffic on your network.
Please pull the slider to the right to prove you are human.”
In spite of platitudes “that excluding certain hardware, platforms, or user-agents is not a goal,” I can see privacy protecting browsers being shut out of using PACTs.
Clive Robinson • June 24, 2026 2:31 PM
@ r,
With regards,
“the petitionability or corruptibility is a FEATURE of human decision trees not a vulnerability.”
As a rough rule of thumb, most “decision trees” are a consequence of hierarchical “power structures”.
As you are aware the closer a human gets to the top of a hierarchical power structure the more susceptable to coruption they are, and the more public facing the hierarchy the more susceptable they are to the faux narrative pushed from the MSM etc.
In the UK we’ve just seen the result of a poor interlect put in power without carrying out oversight or control on the immediate juniors they appointed as Ministers of State.
Despite multiple attempts to fend of credible contenders for his job, his ineptness has through public opinion called not just for his head, but named a suitable replacment.
At the moment it appears public opinion holds sway, as the appointed ministers have not stepped forward…
Sadly “The King” can nolonger call for a headsman at “old palace yard” or “Tower Hill” to dispose of politicians and other ministers of state that had disapointed, such as commiting treason or stealing from the Crown and it’s estates.
Of interest Sir Isaac Newton was placed in charge of the Royal Mint, where he came up with the idea of milled rim coins to stop the gold being “clipped” or in other ways debased. Whilst several people know this rather less know about the punishment for stealing from the Kings Treasury.
From the Norman times onwards “gelding and gouging” were seen as punishments not designed to kill but humiliate and exclude from traditional Baronial and lower courtly society that was based around marriage and land.
As both were done without the use of a knife, I can only assume that people were made of much sterner stuff back then.
yet how great • June 24, 2026 5:05 PM
License plate cameras can track your AirPods, smartwatch, and more, disturbing study finds
The new sensors are made by a defense company whose primary customers are U.S. law enforcement agencies
Clive Robinson • June 24, 2026 7:10 PM
@ Vles, ALL,
With regards,
“… what if AI is powered by quantum computing?”
Well we now know that Current AI LLMs will not reach sentience by scaling.
Likewise we know that turning the wick up on “Stochastic” really does not improve things (in fact evidence is you will get more “Soft Bullshit” / hallucinations).
So the likely hood of real AGI by LLMs is somewhere between “not a chance” and “not in anyone’s current lifetime”.
The other issue is Quantum Computing”(QC) is another of those “give us a decade” repeated every year.
The latest embarrassment from MicroSoft shows just how embarrassing QC is.
But… Some people have reasoned that quantum effects happen in ordinary organic life form. And it’s been shown to be true for things like photosynthesis,
https://phys.org/news/2026-06-energy-boosting-quantum-mechanism-photosynthetic.html
‘https://www.surrey.ac.uk/leverhulme-quantum-biology-doctoral-training-centre/research/quantum-effects-photosynthesis
Through the way smell works in olfactory sensors and a number of other quantum biologics where sensitivity exceeds expectations of purely chemical sensing.
Back some years ago the mathmetition theoretical physicist Sir Richard Penrose pushed forward an idea that human consciousness / cognition could be due to quantum microtubules in the brain…
He was derided for it by many back then. However current research findings suggest there may be something to the idea,
https://speciesuniverse.com/the-quantum-brain-exploring-roger-penroses-theory-of-consciousness/
So yes if QC ever gets to the point it is room temperature and low noise then it may well be possible that AI consciousness happens.
That said I very much doubt I will be around if and it’s a big if QC becomes couplable into all the individual neurons of the base Digital Neural Network, which as I’ve explained before is little more than a “Digital Signal Processing”(DSP) adaptive filter with fuzzing from a random or chaotic source.
Clive Robinson • June 24, 2026 9:48 PM
@ Bruce, ALL,
Microsoft have wasted billions on QC and may well have commited fraud
Is the current flow of not just thought but claims by others.
Boffin claims Microsoft’s supposed quantum leap does not compute due to ‘basic Python errors’
Nature paper argues researchers cherry-picked data. Redmond insists its work is sound
“Prestigious journal Nature has published a peer-reviewed critique of Microsoft’s claims to have made quantum computing breakthroughs – and the scientist who wrote the paper has essentially said Redmond got it wrong.
Microsoft made its claims of a quantum breakthrough in February 2025 when it revealed tech called Majorana and predicted “this breakthrough will allow us to create a truly meaningful quantum computer not in decades, as some have predicted, but in years.”
The software giant’s approach to quantum computing involves Majorana particles, subatomic particles that scientists have not observed directly. The company has pursued this approach for years, but experienced reversals that led to the retraction of some papers. Last year, however, Microsoft claimed it had both observed Majorana particles and harnessed them in a quantum computer.
Criticism of that claim was swift and sharp: we reported boffins willing to go on the record as describing Microsoft’s work as “unreliable” and perhaps even “fraudulent.”“
Read it through and you will find a couple of notable comments…
One of which is Dr Henry Legg, of the University of St Andrews,
“Last year they claimed to be years, not decades from a ‘topological quantum supercomputer,'” Legg told The Register in an email. “My feeling is that they are centuries, not decades away. If it works at all – and, based on what I have seen, the most likely scenario is that it doesn’t work.”
And more damming,
“I demonstrate that Microsoft’s tune-up software is flawed and that coding errors resulted in incorrect statements to peer reviewers,” said Legg. “Raw data, which was omitted from the original paper, also appears to indicate Microsoft’s devices contain considerable disorder and are not compatible with the existence of a topological gap. In other words, the prerequisites for Microsoft’s claims do not appear to be met, but this was obscured because this data did not appear in the original publication.”
Which is very much an “ouch statment”…
The real questions for Microsoft’s share holders are,
“How many billions have the executives wasted on this ‘no ROI’ publicity stunt research?”
Along with,
“How does it compare to the billions they have wasted on LLM&ML systems they are trying to force through Win 11[1] as a significant surveillance and data theft tool?
Especially as the smaller/lighter LLMs from the likes of Deepseek will run on a high end laptop, without needing to hemorrhage data to the Internet as some incontinent agent…
The recent nonsense from the US Executive against Anthopic is effectively pointless and extrodinarilly detrimental to all US AI companies.
China has a rather different attitude and their LLMs running on issolated On-Prem hardware are rapidly gaining on not just Anthropic but OpenAI offerings but without the issues… Thus the “US moat” is rapidly becoming “a puddle of cat piss”… And now tokens are getting close to the real cost and CFOs are seeing bills of millions of dollars with absolutely no sign of an ROI, they are going to look for alternatives that don’t hemorrhage confidential business information to the world by various forms of prompt injection and don’t cost a King’s Ransom.
Speaking of which if you are a CxO in walnut corridor, how do you “value in” a business tool that can be snatched away in 90minutes or less effectively never to return in more than a business cycle or three?
That is,
“What sort of risk mitigation do you do on US SaaS/cloud infrastructure AI?
Especially as new and more interesting Prompt Injection attacks are being found with apparently little difficulty…
The use of On-Prem LLM rather than SaaS cloud etc kind of reduces two areas of risk,
1, US Gov stupidly blocking access.
2, Data surveillance issues of SaaS AI.
But some may find even running AI behind closed doors may not stop “Prompt Injection” attacks via “unread Emails” etc,
https://m.youtube.com/watch?v=YtdLHrzZyB8
[1] Oh funny side note did anyone else know that,
“Win 11 is more than half a decade old and the likely ROI is going to be at best low?”
In part because of all the “bad publicity” and trade media telling people to use some variety of Linux or an Apple laptop… Some other companies are working out how to stay on Win 10 with stand alone applications to avoid paying the Microsoft rent ransoms, especially now that the extraordinary high cost of RAM makes getting “Win11 capable” hardware impractical.
Clive Robinson • June 25, 2026 3:26 AM
@ ALL,
Speaking of Microsoft and it’s bad behaviours and why shareholders should hold Executive “feet to the flames”.
As most readers here are aware there is a mysterious security researcher who is some what upset with Microsoft and has as a result made quite significant vulnerabilities “public” rather than do what Microsoft Executives want which is to “shut up and go away”.
Unfortunately much of the MSM and Trade press have taken the Microsoft attitude / brown envelope and tried to paint Microsoft as the victim of an extortionist.
The truth is Microsoft Execs are simply reaping the crop of their own stupidity yet again.
I know from bitter experience back in the 1980’s[1] that reporting vulnerabilities would cause people to treat you as an extortionist, Fraudster or worse form of criminal[2]…
Any way, the villains in this current saga are almost certainly Microsoft Execs and to save me listing why, have a watch of,
https://m.youtube.com/watch?v=IF-RipYYe6s
[1] I’ve mentioned before how UK Prime Minister Margaret Thatcher tried to have me framed for fraud, and it was only sage advice from my employer and the fact I dug my heels in that caused the Met Police and BT tactics to fail. Also how a short while later Steven Gold and Robert Schifreen “tried to be helpful” rather than as I advised dig their heels in and treat it as an adversarial process as I had. And so they paid the price of being set up by Maggie and the Met Police, and of having the next couple of years of their lives ruined by being convicted show trial style for fraud then fighting through various appeals upto the House of Lords who basically said that abusing fraud legislation (UK FA 74 and FCA 81) was not the way to deal with computer issues… So booted it back to the legislators and we ended up with the UK “Computer Misuse Act of 1990″(UK CMA 90) which was fairly appalling legislation as has been the later Computer Acts.
[2] I also found quite a few vulnerabilities with *nix systems from both the BSD and AT&T trees. I had a bit of a reputation in being able to not just break the login process but also then privilege escalate upto root (or above ie hardware level access).
Noticeable was the fact that *nix developers were not adversarial and generally fixed things fast.
ResearcherZero • June 25, 2026 5:51 AM
Hesgeth and Trump work to further degrade U.S. defense and intelligence capability.
‘https://www.edgen.tech/news/post/hegseth-forces-out-24-generals-as-pentagon-purge-deepens
Distinguished for his time in special forces, Gen. Chris Donahue, former former Delta Force commander and former commander of the 82nd Airborne Division, commander of U.S. Army Europe and Africa, considered a front runner for Army chief of staff -or- chairman of the Joint Chiefs of Staff, has submitted for retirement after butting heads with Pete Hegseth.
https://www.cbsnews.com/news/gen-chris-donahue-to-retire-army-commander-latest-departure-military-trump-hegseth/
“the officers forced out have had impeccable reputations”
Dozens of experienced generals and commanders have been forced out while on active duty.
https://www.theguardian.com/us-news/2026/may/03/pentagon-pete-hegseth-us-military
Counter-terrorism among agencies to have workforce significantly reduced.
https://www.nbcnews.com/politics/trump-administration/odni-begins-firings-under-bill-pulte-director-national-intelligence-rcna351290
ResearcherZero • June 25, 2026 6:02 AM
If that does not give everyone working in the military that sinking feeling, fear not.
Trump-class: A recipe for wasted resources, slowed production, dead sailors and lost battles.
Military aircraft can fly much greater distances than in the 1940’s. Battleships today would be sitting ducks, easily destroyed by modern weapons, along with their crews. Building such vessels will delay vital and effective hardware needed to win modern wars. The money and resources wasted in the process is enormous.
Obsolete-class battleships proposed by Trump would displace aircraft carriers. The “Golden Fleet” folly would also delay production of Virginia-class and Columbia-class subs. Any future administration will predictably end the program before a single ship is completed. No nation on Earth builds battleships anymore and do not build them for very rational reasons. Modern weapons could quickly sink them, along with everyone onboard.
The project will significantly degrade U.S. naval effectiveness and have knock on effects elsewhere.
‘https://www.telegraph.co.uk/news/2026/06/07/every-trump-class-battleship-one-less-us-aircraft-carrier/
JG5 • June 26, 2026 10:57 AM
ouch:
“Looks like you have hit our Markdown AJAX handler without posting content. Weird.”
Lost a more or less identical piece. Did my best to reconstruct from memory. Which isn’t what it used to be. I usually put the post on the clipboard in case something like this happens, but I have been lulled into complacency by the good behavior of the hardware/ firmware/ software.
This is one of the more profound analyses that I’ve seen lately. @Clive smiles knowingly.
How Sandboxing can Solve Privacy, Security, and Tyranny
https://www.youtube.com/watch?v=DSCM1XVLtJg
I have a new symbol for your collection. I call it omega-dot-dot: ὣ, the rate of acceleration of the rate of acceleration of angular velocity. As you, your company, your country, or your society circles the drain. A function of time, radius, fluid depth, drain diameter, angular momentum, … Ideally, it would be omega-umlaut, where the dots are round. Gödel might have something to say about the shackles and chains of using only ASCII characters.
lurker • June 26, 2026 2:27 PM
@JG5
This apparently depends on the browser in use, but some people have reported simply hitting the “Back” button can return you to the submission page as it was just before you hit “Submit”. It works for me, mostly …
ResearcherZero • June 27, 2026 4:00 AM
@lurker
They tried the “Back” button at the White House, but the internet was not cached in the 1920’s.
Clive Robinson • June 27, 2026 10:00 AM
Hmm…
Another Auto-mod failure…
I guess it’s time to try partitioning…
Part 1,
@ ResearcherZero, All,
With regards,
“Military aircraft can fly much greater distances than in the 1940’s.”
Yup it’s one of the reasons in the past I’ve pointed out that the day of “Carrier Groups” is long over against all but some second world and quite a few third world nations.
Israel and the US has very recently discovered that “bomb them back to the stone ages” is nolonger an effective policy against a thoughtful thus suitably defended leadership and armed forces where “chop the head off of the snake” strategies don’t work either.
And in the case of the US could easily cost them one or two carrier groups being beaten by barnacles and similar (they had similar problems with “ships worms” a couple of centuries back I guess the lesson got lost with the anchor).
Clive Robinson • June 27, 2026 10:03 AM
Part 2a,
But the Chinese and Russians have put some thought into it as well.
With Hypersonic nuclear missiles and unmanned drones with nuclear packages.
Clive Robinson • June 27, 2026 10:09 AM
Part 2b,
As anyone keeping an eye on things should know,
China is very much into churning out tactical and some strategic nuclear warheads for Hypersonic missiles and sub surface unmanned long duration drones.
Clive Robinson • June 27, 2026 10:18 AM
Part 2c,
With the Russians apparently having an unmanned sub surface drone.
Clive Robinson • June 27, 2026 10:20 AM
Part 2d,
With a range of upto 1000 miles and a “city busting” high yield nuclear warhead.
Only one of which would do to Pearl Harbour what the Japanese Navy could not to do in WWII.
And as I’ve said before the 15mins of Fame for Carriers was in the tail end of the Pacific War after Victory in Europe was assured.
Clive Robinson • June 27, 2026 9:54 PM
Part 3,
Importantly people should give careful consideration to the words of Admiral Herbert Richmond, a well respected naval historian who stated of WWII,
“Sea power did not win the war itself: it enabled the war to be won.”
And it’s the “enabling” that people should consider, because technology had overtaken carriers and other large vessels by the end of the 1960’s
To win a “boots on the ground” war you have to get not just the boots but everything the soldiers standing in them need for three months to a couple of years or more from outside thearter into it. Modern drone technology has greater range than aircraft and even cruise missiles. No matter where you put a carrier it’s vulnerable to more than five times the radius of the maned aircraft on board (a significant failing of the F35 and other stealth flying bricks).
Clive Robinson • June 27, 2026 10:15 PM
Part 4,
The Ukraine has proved that they can get sea-drones into defended ports right up against Russian Ships and Subs and do rather more than damage the paintwork and ding the side. Likewise slow fleet vessels like Russian Black Fleet oil tankers etc.
This “home made dromes” is nothing against the Ukranian’s but from a nation that is under constant aerial bombardment they have developed technology rapidly that few could have dreamed of even in the Super Powers. In part they got support from Turkey in this endeavour whilst Russia got assistance from Iran…
Warfare has changed a lot. Alright I had a bit of an “insider view” from the begining and have predicted on this blog the use to which unmanned drones, especially at sea could do.
And importantly that there is further scope such as drones that act as manoeuvrable “bottom squating” mines. That sort of act like “ticks in the long grass” just sitting in a minimum energy state waiting for the right prey to come in range then hop up attach and do the damage.
The US in particular have left “mine detecting / sweeping” vessels out of their construction / deployment line up…
Clive Robinson • June 27, 2026 10:23 PM
Part 5,
We have much more to see in this respect, as humans are very resource intensive to support and even a thousand pairs of boots will cost way more than a $billion just consuming on ships at anchor which would all be considered “capital vessels” in this day and age.
Unmanned drone tech be it air, land, or sea based has way way greater range than a bunch of “brown socks” at a fraction of the price and can be built in a few days rather than a few years.
Fun fact, it takes way way longer to make a tactical nuclear payload these days than the supply / manufacture chain of the delivery systems…
Also nuclear disasters in Ukraine and Japan have shown our assumptions about radiation levels are wrong. Which is why the
“linear no threshold”
Limit line is being quietly dropped[1]. In fact evidence suggests that some lowish levels of radiation are likely to be more beneficial than harmful.
[1] It’s something I’ve noted here before, and it was based on animal studies from earthworms to wild horses in the Ukraine Chernobyl fallout zone prior to the Russian Invasion. If people think I’m nutzz maybe they should see somebody elses respected view point,
ResearcherZero • June 29, 2026 5:26 AM
@Clive Robinson
If the experiments UK police have conducted with AI are anything to go on, then the US may want to think carefully before rushing to judgement based on the output of AI models.
Police made assumptions about future behaviours of people they fed into their predictive AI system. Amongst the data added was information about victims of crime, along with children who police believed may become future victims of crime. The system was also used to flag people who might become offenders. The results produced by the predictive analytics were riddled with errors, often flagging people who never committed offenses. Many of models were deeply flawed, with failure rates of up to 90% at identifying potential criminals.
Humans are prone to believing data produced by machines is accurate and make judgements accordingly. Those in official positions are especially prone to not admitting mistakes.
(Double checking or vetting of data by those currently in the White House seems uncommon.)
It has been said, that in Washington – the only classified information is bad news.
British cop used AI models to fabricate evidence and pervert the course of justice.
https://www.theguardian.com/technology/2026/jun/12/police-officer-under-criminal-investigation-over-alleged-use-of-ai
Clive Robinson • June 29, 2026 9:34 PM
@ ResearcherZero,
With regards,
“If the experiments UK police have conducted with AI are anything to go on, then the US may want to think carefully before rushing to judgement based on the output of AI models.”
They should but they won’t if past behaviours are considered.
Look up Operation Ore, and how US Federal Guard Labour “tampered with computer evidence” that they then sent to be used in UK court cases.
It was never officially investigated to find out if the US Federal Guard Labour did it on their own decision or if the UK Met Police asked them to do so directly or indirectly.
The corner of the carpet was lifted and the brush swept a couple of times…
A US Federal Judge was on record as saying from his experience he quickly did not believe what US Federal Guard Labour said in court.
But as I’ve also noted here in the past the Forensics submitted by them is always questionable when new.
They make claims that there is no justification for and the wrong people go to jail and sometimes it’s years before the fact that the forensics are false becomes sufficiently well known that a few lucky people get released…
Even the once “gold standard” of “Finger Prints” are mostly nonsense and based on “subjective claims”.
There was the nonsense of “pour patterns” in cases of fires that was found to be totally subjective belief. When somebody finally “tested” it was found to be completely bogus.
One that I found was that claims made for DNA were false by simply looking at the test method. It did not stop false evidence being planted… I made a bit of a nuisance of myself trying to get it taken seriously and got considerable push back for “trying to kill the golden goose”.
Then some time later an Australian botanical researcher made the same observation, and it became news and all of a sudden cases were getting reopened around the world…
The fact is as I’ve said before,
“Forensics does not follow the scientific model”…
Science goes by design from provable cause to effect as this minimises error and in effect enforces occam’s razor (in effect the logical AND which is reducing).
But Forensics goes from visible effects to conjectured cause (in effect the logical OR which is expanding).
As far as we can tell the “black box” “Digital Neural Network” at the heart of LLM and ML systems is “expanding” not “reducing” hence one of the reasons for the “soft bullshit” hallucinations you get in upto 9 out of 10 result runs for some still Current AI systems.
Clive Robinson • June 29, 2026 10:54 PM
@ JG5,
With regards,
“How Sandboxing can Solve Privacy, Security, and Tyranny”
Yup it tends to be something I bang the drum about 😉
Most often these days when I talk about Claude Shannon and his notion of communications channels in the flow of information, and how it gives rise to amongst other things “the observer problem”
The big problem which is one few ever get their heads around is in two parts,
1, Shannon proved, that without redundancy information could not be communicated.
The second part was reasoned out from the opposite direction by Gus Simmons who wanted to send information without sending other information. Think of it as verifying a serial number without sending the number.
2, Simmons showed in effect that where there is redundancy in communications there must be a Shannon Channel created.
So in every transfer of information within it there is the capability of another channel and so on. A case of “infinite nesting” or “lesser fleas” or “turtles all the way down” depending on your preferred phraseology.
And it’s this that gives rise to “the observer problem”. No matter how intently an observer examines the flow of information in a communications, they can not say if there is other information being transferred as well.
Another of Shannon’s ideas is “perfect secrecy” which gives us at it’s most well known “The One Time Pad” encryption system.
It gets it’s strength through the simple fact that,
3, All messages in an information communication are equiprobable.
Thus everything rests on the key used which is the same size as the message sent (look up “unicity distance”).
Now flip it on it’s head, “all messages are equiprobable” also means under these rules means “all messages are deniable”.
That is “same cipher-text” gives “any and all messages of the same length or less” depending on the “secret key” only.
Thus an Observer who sees and records the “cipher-text” who gets given a key after the fact, can not say if the plain-text they get with that key was actually the intended text at the time of transmission or not.
This gives “full deniability” of the message.
Because the sender and receiver can give two different equiprobable keys to the observer…
Which means that the observer has to try to correlate to other information/events external to the communications, to try and decide which party if either is being truthful or not.
As an aside, fun fact about the Titanic and it’s water tight compartments without “tops”. As the ships designers knew, if they included tops on the compartments they would create dangerous conditions under ordinary ships operations. As the pressure on either side of a door would be different most of the time. Also breathable air would quickly be unavailable. Managing just these two issues is one of the reasons submarines are so difficult to design, and more than sufficient reason to make all “sub-surface” vessels unmanned and thus save a very great deal in costs for engineering and maintenance.
Subscribe to comments on this entry
Sidebar photo of Bruce Schneier by Joe MacInnis.
Clive Robinson • June 19, 2026 11:50 PM
@ Bruce, ALL,
KPMG report promoting AI found to be full of inaccuracies and AI hallucinations
Yup if you use garbage to make garbage…
The KPMG report was titled,
“Redefining excellence in the age of agentic AI”
Was found to be so full of nonsense, inaccuracies and AI hallucinations it was an “easy catch” for, GPTZero…
It’s just one of many, their investigators have found and revealed, and in this case has became further reported in more mainstream media for the business industry.
Put simply GPTZero research into why major government reports, academic papers, and other research related activities have become plagued with AI hallucinations.
https://techcrunch.com/2026/06/13/kpmg-pulls-report-on-ai-usage-due-to-apparent-hallucinations/
Gives an overview.
However there is more to it than “egg on the face” for a well known and oft hated consulting firm.
It appears that this behaviour is so numerous these days that GPTZero have come up with a new “term of art in the domain” for it,
“Vibe Citing”
It refers to what are faux citations, that look plausible to the casual eye, but are in fact nonsense and most probably are created by the use of “generative AI” that are then “not fact checked” by humans trying to be “more productive”.
If Current AI LLM and ML Systems apparent cause “productivity increases” this way, by “soft bullshitting”, then one has to question if it’s an increase of productivity by employees or just garbage with less than zero ROI.
https://www.techradar.com/pro/a-major-kpmg-report-on-ai-was-found-to-be-chock-full-of-ai-hallucinations