Comments

not important • November 21, 2025 6:55 PM

The words you can’t say on the internet
https://www.bbc.com/future/article/20251118-the-words-you-cant-say-on-the-internet

=Algospeak, as it’s often called, is a whole coded language built around the idea that algorithms bury content that uses forbidden words or phrases, either to boost the political agendas of social media companies, or to sanitise our feeds for advertisers.

History is littered with examples of social media companies quietly manipulating what content rises and falls, sometimes in ways that contradict their claims about transparency and neutrality. Even if it doesn’t come down to individual words, experts say the tech giants do step in to subtly curb some material.

!!!!The problem is you never know why a post fails. Did you say something that upset
the algorithms, or did you just make a bad video? The ambiguity has encouraged a
widespread regime of self-censorship. On one end of the spectrum, the result is people talking about serious subjects with goofy language. But at the extremes, some users who just want to go viral avoid certain topics altogether.

it could mean there are ideas that some people never get to hear.

In practice, though, social media platforms have repeatedly meddled with which voices are amplified or buried, contradicting their rhetoric about openness and fair play, according to investigations by the BBC, advocacy groups, researchers and other news outlets.

The problem is the policies governing social media are heavy handed and largely
invisible, says Sarah T Roberts, a professor and director of the Center for Critical Internet Inquiry at the University of California, Los Angeles (UCLA).

People rarely know where the boundaries lie, Roberts says, or when the platforms quietly push some posts forward and others out of sight.

Social media companies make their money from advertising. Ultimately, that means their goal is to make apps that lots of people want to use, fill them with content that makes advertisers comfortable and do whatever is necessary to prevent government regulators from getting in the way, Roberts says. Every change to the algorithm and every content moderation decision comes down to that profit
motive.=

Clive Robinson • November 22, 2025 4:25 AM

@ Bruce, ALL,

My key, My Key, My Kingdom for my key!

Can there be too much security?

Yes, because systems can become more fragile than without security if not fully thought out.

Such an event has hit the International Association of Cryptologic Research (IACR) annual leadership election voting system last week.

https://arstechnica.com/security/2025/11/cryptography-group-cancels-election-results-after-official-loses-secret-key/

“Oops. Cryptographers cancel election results after losing decryption key

One of the world’s premier security organizations has canceled the results of its annual leadership election after an official lost an encryption key needed to unlock results stored in a verifiable and privacy-preserving voting system. “

With hindsight it’s easy to see why,

“Per the association’s bylaws, three members of the election committee act as independent trustees. To prevent two of them from colluding to cook the results, each trustee holds a third of the cryptographic key material needed to decrypt results.”

And thereby each trustee becomes,

“A single point of failure”

Any one who has worked with Intrinsically Safe or Safety Critical systems knows from almost their first day the hard and fast two fundamental foundational rules,

1, No single points of failure.
2, Fail to safety.

Whilst the second point can be problematic with Cryptographic based systems the first should always hold.

We even have cryptographic “M of N” Secrecy sharing protocols that can be used to ensure this.

At the very least the associations rules should have allowed for human frailty. Aircraft fall out of the sky, ships sink, and motor vehicles crash many on more than a daily basis, likewise sickness and unfortunate health events take many more, often quite unexpectedly.

Clive Robinson • November 22, 2025 5:33 AM

@ ALL

“Is Microsoft AI man an idiot or a liar?”

In,

https://www.pcgamer.com/software/ai/microsofts-head-of-ai-doesnt-understand-why-people-dont-like-ai-and-i-dont-understand-why-he-doesnt-understand-because-its-pretty-obvious/

PC-Gamer author Tyler Wilde asks the all important question,

“Microsoft’s head of AI doesn’t understand why people don’t like AI, and I don’t understand why he doesn’t understand because it’s pretty obvious

Microsoft AI CEO Mustafa Suleyman says it’s “mindblowing” that people aren’t more impressed with generative AI tools.

“Jeez there so many cynics!” wrote Suleyman on X this week. “It cracks me up when I hear people call AI underwhelming. I grew up playing Snake on a Nokia phone! The fact that people are unimpressed that we can have a fluent conversation with a super smart AI that can generate any image/video is mindblowing to me.”“

Anyone get the feeling Mustafa has not progressed since first fingering his snake game?

How ever I’m more suspicious, Upton Sinclair noted a century ago that,

It is difficult to get a man to understand something, when his salary depends on his not understanding it.

In effect the man becomes deliberately dishonest for personal profit…

I’ve warned several times already that Microsoft has the BE Business Plan and how it would be used for surveillance and profit. I gave more details[1] as I became more certain that was what Microsoft intended.

Interestingly the PC-Gamer author concludes similar with,

“AI and machine learning may in fact transform the world, but there’s no reason to assume that the transformation will be a good thing unless we actively try to make it a good thing. So far, tech companies have given no indication that they care about anything besides the pursuit of profit. I don’t think that we’re the cynical ones here.”

But on another note if you want a little light relief and a laugh,

https://www.pcgamer.com/software/ai/grok-ai-temporarily-so-sycophantic-it-claims-elon-musk-is-the-best-at-drinking-pee-and-other-things-im-not-going-to-put-in-a-headline-you-cant-make-me/

Hopefully “Hell-on Rusk” had a chuckle rather than do the usual “Chuck the toys out of the pram”.

[1] I said in june this year at
https://www.schneier.com/blog/archives/2025/06/hearing-on-the-federal-government-and-ai.html/#comment-445807 on this blog

“But the AI business plan of Meta and Microsoft and I assume Alphabet as well is as a surveillance tool to gather as much “Personal and Private Information” from people as possible.

I’ve called it the “BE Plan” for,

“Bedazzle, Beguile, Bewitch, Befriend, and BETRAY”’

Because this just shows the steps which are already in full progress and as a result,

“All your privacy will be striped bare”

And endlessly be sold over and over, passing from hand to hand each making their percentage on you.”

JG5 • November 22, 2025 8:22 AM

No comment. Standard fare. “Stupid company stories are a dime a dozen.” – our gracious host

We Hacked Flock Safety Cameras in under 30 Seconds.
Benn Jordan 909K subscribers
777,218 views | Nov 16, 2025 | ATLANTA
Go to https://ground.news/benn for a better way to stay informed. Subscribe for 40% off unlimited access to world-wide coverage through my link. ⬇️ Epic amount of info below ⬇️
https://www.youtube.com/watch?v=uB0gr7Fh6lY

Bernie • November 22, 2025 3:45 PM

@ Clive,

Thank you for getting to it first; I was about to mention the lost key article on Ars Technica. Sometimes it feels weird to heard about certain news items from outside of Schneier on Security. Speaking of which, if Bruce ever changes the name of this blog, I’m going to have a heck of a hard time trying to leave a comment.

KC • November 22, 2025 8:17 PM

For those of you collecting books to read after ‘Rewiring Democracy’

Here’s a possible addition for your stack: ‘The Fourth Intelligence Revolution.’ It lists under AI.

An excerpt:

This revolution will expand so far that it will be marked by the democratization of intelligence and its incorporation into our personal lives to a point where we will all effectively become citizen spies.

So we can all become citizen spies, eh?

If you are curious about surveillance and don’t mind being disturbed, this book may be up your alley. A review: “it’s highly readable and up-to-date on current threats and capabilities.” Of course, I’m sure there’s nothing classified.

tags: ai, espionage, oof

mannix • November 22, 2025 10:27 PM

Watch out for online contact with Chinese spies, UK defence minister warns public

https://www.theguardian.com/world/2025/nov/19/watch-out-online-contact-chinese-spies-uk-defence-minister-warns-public

China’s power play: MI5 warns of relentless espionage attempts in Britain

https://www.theguardian.com/world/2025/nov/19/china-mi5-espionage-britain

China in ‘covert and calculated’ effort to recruit MPs and peers, minister says

https://www.theguardian.com/world/2025/nov/18/mi5-issues-alert-to-mps-and-peers-over-chinese-espionage

Clive Robinson • November 23, 2025 7:20 AM

@ Winter, ALL,

Is China salting the reactor?

I’ve been hearing stories about China breeding uranium from thorium in a “Molten Salt Reactor”(MSR)[1]. Of which it’s difficult to find articles that are not MSM re-hashes of Chinese government anouncments.

So after hunting around,

https://www.world-nuclear-news.org/articles/chinese-msr-achieves-conversion-of-thorium-uranium-fuel

[1] If true such a reactor would not need to be refuelled for about 175,000 hours (~20years) and in theory could be made fully automatic for that period as well… Apparently It’s one of the things Rolls-Royce are allegedly looking into for “Small Modular Reactor”(SMR) designs.

Clive Robinson • November 23, 2025 7:48 AM

@ passive rubber chicken, ALL,

It’s not the Chinese spys that go after politicians and the military that worry me.

It’s what are known as,

“Economic agents of a hostile Foreign Power.”

Of which I’ve had the misfortune to come across from China, France, Israel and Russia as just a “design engineer”. I’ve mentioned the reasons why before and more importantly the complete lack of interest the UK Gov Security Services had (and still do).

Whilst “High Secrets” might be grand they are mostly “pointless” in effect all to often they are not really secret but “position papers” and what can be discovered via Google Maps and a drone these days.

What are considered “Low Secrets” are business and industrial trade secrets aquired via Industrial Espionage. What many do not realise is these have very real, immediate value, and “Economic Warfare”, is a very real thing and extraordinarily harmful to the economy.

The only country that appears to take “Industrial Espionage” as a serious threat and act on it untill recently is South Korea…

lurker • November 23, 2025 1:01 PM

@Clive Robnson, ALL
re Chinese MSR, “… such a reactor would not need to be refuelled for about 175,000 hours (~20years)”

Will either of us be around in 20 years time to see the refuelling? The “interesting” part about thorium MSRs is the first stage of the breeding produces the highly active U233, which is still present in the salt mix when all the thorium has been consumed.

Clive Robinson • November 23, 2025 1:26 PM

@ ALL,

How much to get a DOGiE mutt out of the pound?

Apparently since Hellon-Rusk and the Doh-gnarled-Trumper had their falling out the muts supposedly employed by DOGE have realised they have no protection from prosecution or people in general giving them the treatment they “Right Royally disserve” (apparently heads on spikes above “traitors gates” is an English King Thing).

https://www.rawstory.com/doge-employees/

Notes,

“Current and former Department of Government Efficiency (DOGE) staffers are growing increasingly concerned that the work they did slashing government programs and eliminating jobs will come back to haunt them with the possibility of criminal prosecutions.

Worse still is their growing belief that the billionaire Elon Musk, who recruited them, won’t step up to save them by appealing to Donald Trump on their behalf should things take a turn for the worse.”

A longer version with more history of events can be found at,

https://www.politico.com/news/magazine/2025/11/21/doge-elon-musk-succession-00641110

Rontea • November 23, 2025 2:31 PM

@KC
“For those of you collecting books to read after ‘Rewiring Democracy’

On Apple Books, Rewiring Democracy, is part of the Strong Ideas series. It includes:

-Data Feminism
-The Smart Enough City
-Sharenthood
-The Digital Closet
-Hacking Life
-On The Brink of Utopia
-Artificial Communication
-Fantasies of Virtual Reality

not important • November 23, 2025 6:13 PM

https://www.timesofisrael.com/chatgpt-for-genetics-nvidia-sheba-join-forces-on-ai-engine-for-personalized-medicine/

=Scientists say the AI research engine will be able to decode the majority of the human genome that remains obscure, opening the door to new drug discovery and personalized medicine.

Now Israeli scientists at Sheba Medical Center have teamed up with US chip giant
Nvidia and New York City’s Mount Sinai Hospital to create large language models
(LLMs), using the same type of technology that powers applications like ChatGPT, but are trained on the biological language of our bodies, to better understand and treat the diseases that afflict us.

“We want to create a type of ChatGPT of genomics that will allow users to put in
a whole genome sequencing of a person and will be capable of answering questions
about health risks, or what’s the best drug or treatment for a disease based on the unique genetic makeup of a person,” Avner Halperin, CEO at Sheba Impact at the ARC Center of Digital Innovation, told The Times of Israel.

The three partners are embarking on an ambitious three-year project at an investment of tens of millions of dollars to create a research engine that harnesses generative AI to decode the majority of the human genome – the genetic blueprint for human life – that remains poorly understood.

genomic research engine will be capable of identifying patterns and mechanisms that link a person’s genetic makeup to disease risk and therapeutic response.

The human genome – the set of instructions to build and sustain a human – is made up of 3.2 billion DNA characters or letters. Over the past two decades, science has made advances in the sequencing of a complete human genome, but only 2% of the human DNA consists of protein-coding genes, while the function of the remainder 98% has been difficult to interpret using traditional approaches, according to Halperin.

The Sheba-led collaboration aims to start decoding the mysteries of the remaining
98% through LLMs and machine learning technologies. Nvidia will provide the
computational power and AI infrastructure, and Sheba and Mount Sinai will lend the scientific and clinical expertise to pool and synthesize vast quantities of
genomic datasets.=

ResearcherZero • November 24, 2025 12:53 AM

@Clive Robinson

Most of the people sent to investigate theft of economic, business and industrial trade secrets do not even know what they are looking for. They are not from any of those sectors and have no training in any of those fields. The warnings to members of cabinet and any of the other officials and politicians are not taken seriously and state and federal police are also completely unequipped (unenthusiastic) to tackle the problem. Many of individuals involved in the theft of such secrets have been in place for decades with little to deter them from continuing in their activities or to prevent them from recruiting further assets.

UK politician jailed for taking bribes from Russia to further the Kremlin’s interests.

‘https://apnews.com/article/britain-politician-reform-bribery-sentence-372e14bfb629aa413814154d6321736f

Clive Robinson • November 24, 2025 2:30 AM

@ lurker,

With regards,

“Will either of us be around in 20 years time to see the refuelling?”

The thought that occured to me is,

“Will the reactor actually survive that long?”

As you probably know Russia has dumped a few hot reactors around the globe as the ships and subs they were in have become to brittle or to corroded…

Which also brings us onto the more interesting part,

“The “interesting” part about thorium MSRs is the first stage of the breeding produces the highly active U233, which is still present in the salt mix when all the thorium has been consumed.”

Compared to U233 thorium is considered by many as fairly benign… because U233 can be turned into nuclear weapons, the first attempt being back in 1955 by the US if memory serves. Subsequently both Russia and India have played with U233…

Thus there will be questions about recovery and purification that some are going to ask.

Clive Robinson • November 24, 2025 5:23 AM

@ ALL,

The Cuckoo in the Nest Squawk of IoT

Have you ever thought what happens to IoT devices that get abandoned by the supplier and their “ET Phone Home” behaviours?

Well, lets take a look at Google and the Nest thermostat,

https://www.theverge.com/news/820600/google-nest-learning-thermostat-downgraded-data-collection

Google is collecting troves of data from downgraded Nest thermostats

Google officially turned off remote control functionality for early Nest Learning Thermostats last month, but it hasn’t stopped collecting a stream of data from these downgraded devices. After digging into the backend, security researcher Cody Kociemba found that the first- and second-generation Nest Learning Thermostats are still sending Google information about manual temperature changes, whether a person is present in the room, if sunlight is hitting the device, and more.

So “Still phoning home” and “squawk ingvlike a cuckoo in the nest”.

But why is this alien bird still communicating? After all Google severed the users contracts… This is the “official reply”

In a statement to The Verge, Google spokesperson Laura Breen says users “who prefer to stop providing these logs can simply disconnect their device from Wi-Fi” by accessing the option in the on-device settings menu. “Diagnostic logs, which are not tied to a specific user account, will continue to be sent to Google for service and issue tracking,” Breen says.

What the article does note is that others have simulated the Google API so that the users can have some functionality back.

I’ve warned for some time that the “ET Phone Home” devices that have nearly all their features via a suppliers servers are not just a security risk but very vulnarable to service termination without warning or massive hikes in service fees.

My argument is,

“They are a total failure waiting to inflict harm on the users in all sorts of ways.”

In all seriousness they are not worth the risk or cost, just have nothing to do with them for your own safety and sanity.

Clive Robinson • November 24, 2025 3:34 PM

@ robotorch, ALL,

With regards CrowdStrike comments in the link.

Two things to note,

1, What a company claims and what is true may not be the same thing.
2, Insider threats are actually the real problem in ICTsec systems.

Combine the two and you can see why a company might “claim insider” when it’s actually not.

Hence I advise an open mind on what has been said but no verifiable evidence given.

Clive Robonson • November 25, 2025 2:19 AM

@ Bruce, ALL,

Memory “ain’t wot it was!”

Something to consider in your computing life is the impermanence of long term data storage. Or,

“Bit Rot is every where you look”

But it’s worse than you might think…

The unpowered SSDs in your drawer are slowly losing your data

SSDs have all but replaced hard drives when it comes to primary storage. They’re orders of magnitude faster, more convenient, and consume less power than mechanical hard drives. That said, if you’re also using SSDs for cold storage, expecting the drives lying in your drawer to work perfectly after years, you might want to rethink your strategy. Your reliable SSD could suffer from corrupted or lost data if left unpowered for extended periods. This is why many users don’t consider SSDs a reliable long-term storage medium, and prefer using hard drives, magnetic tape, or M-Disc instead.

https://www.xda-developers.com/your-unpowered-ssd-is-slowly-losing-your-data/

So much for SSDs as backup media 1-3 years is not what many want as a “data retention period”. They want “Grandparents photo album” permanence of atleast a hundred times that.

The reality is data even on magnetic media and write once optical disks is going to die in 50-100 years.

It’s why some are even looking at using the equivalent of a variation on QR/iQR Codes on what many would consider “old school cine-camera” 16mm film stock that gets longterm stored in a freezer.

Which then brings in the problem of “reading it back” in even 10years let alone a century or three. Technology moves on to the point that it is more than obsolete. So just does not get supported in any way any longer. Seen a 5&1/4 floppy drive recently? How about a Zip Drive or even CD/DVD drive?

I have a Sony DAT audio recorder I paid quite a bit of money for when they first came out (think more than a good condition second hand car). I can nolonger use it because Sony nolonger make the “rubber band” used in the drive mechanism…

Our history and knowledge is,

“Fast disappearing like smoke in an autumn breeze.”

Clive Robinson • November 25, 2025 4:12 AM

@ Bruce, Winter, ALL,

More fuel on the AI v human Brain fire

Besides all the nonsense over LLMs there are other aspects to AI that have almost “flame war” type intensity in some, but it’s hard to see through the smoke of the dumpster fire Current AI LLM and ML Systems have become.

One of these “other” areas is about how do you “boot up” any intelligence entity/system –be it biological or mechanical– from “freshly minted” nothing, to fully functioning as a reasoning and intelligent entity.

As with all such boot ups there is a flip or tipping point where development moves from self to within environment.

As with computers there has to be a certain minimum of development for the system to function. After all if a computer can not get and put data into a communications channel then it is effectively useless. Similar applies to the biological brain.

In some creatures it’s fairly apparent they are highly functional after birth and can “run with the herd” others such as primates are almost entirely helpless and can take months or years to develop independence in an environment.

So what of the human brain and it’s boot up?

This might be of interest and answer a few questions,

https://news.ucsc.edu/2025/11/sharf-preconfigured-brain/

However it will also cause some feathers to be ruffled in the AI-v-Human crowds.

Press pack article,

https://news.ucsc.edu/2025/11/sharf-preconfigured-brain/

Paper,

Linked in article to Nature’s “paywalled” site. Fine if you have a Uni/work access, but otherwise well over priced.

Clive Robinson • November 25, 2025 8:12 AM

@ ALL,

More on non hybrid PQ only pushed from the NSA

For some time now I and others have been commenting that non-hybrid PQ only Crypto would be a quite bad idea.

Yet for some reason this bad idea has been receiving the equivalent of “back door funding” from the likes of the NSA…

Well it’s been documented to a certain extent,

1, ‘https://blog.cr.yp.to/20251004-weakened.html

2, ‘https://blog.cr.yp.to/20251123-corruption.html

3, ‘https://blog.cr.yp.to/20251123-dodging.html

4, ‘https://blog.cr.yp.to/20251123-scope.html

And it does not read very well.

notch in the side, now it's a double sided floppy! • November 25, 2025 8:17 PM

@Clive,

“Seen a 5&1/4 floppy drive recently?”

Believe it or not, while some of the old hardware kinda sucks, I have 5 1/4 floppies from the 80’s which read better than some of the modern media today!

ResearcherZero • November 26, 2025 2:56 AM

GrapheneOS moves infrastructure after French prosecutors threaten arrests and backdoors.

‘https://www.privacyguides.org/news/2025/11/22/grapheneos-migrates-server-infrastructure-from-france-amid-police-intimidation-claims/

Legislation regarding Chat Control may be held behind closed doors without discussion.
https://demstate.com/article/eu-plans-to-pass-controversial-chat-control-legislation-without-discussion

The threats add weight to concerns that the bill will allow for forced mandatory scanning.
https://techreport.com/news/new-eu-chat-control-proposal-privacy-experts-see-dangerous-backdoor/

ResearcherZero • November 26, 2025 3:02 AM

SitusAMC, a finance vendor that handles compliance, accounting records and legal agreements was breached. The FBI is investigating as the vendor’s systems handle large volumes of nonpublic information for banking and real-estate (much of it private and confidential).

‘https://www.nytimes.com/2025/11/22/business/bank-data-hack.html

Clive Robinson • November 26, 2025 4:08 AM

@ notch in…,

“Believe it or not, while some of the old hardware kinda sucks, I have 5 1/4 floppies from the 80’s which read better than some of the modern media today!”

I’ve mentioned before I’ve an Apple ][ from the 1970’s and the original language disks still read OK.

But they were “low capacity” by modern standards, and single sided. But as you indirectly note, if you cut a notch in the other side of the floppy you could do an ELO Bluesky with a “Now please turn me over” at the end of the A Side track. It was an interesting decade, that gave us 8bit CPU chips and the birth of personal computing…

The things the younger generations have missed out on, like good music, bad taste in clothes and the combination in Noddy Holder,

https://www.youtube.com/watch?v=lhSGZiMQIqc

Who now gets remembered mainly for what he called his “pension fund”, and the famous line of,

“Look to the future it’s only just begun!”

And as it’s that time of year almost,

https://m.youtube.com/watch?v=WM7M7zSMJcw

ResearcherZero • November 26, 2025 4:31 AM

Software projects often fail because of the delusional aspirations of decision-makers.

‘https://spectrum.ieee.org/it-management-software-failures

Now that the “cost cutting” experiment of DOGE is over, the public is not being informed of expenditure for the new AI plans, or the effect on existing and established supply chains.
https://venturebeat.com/ai/what-enterprises-should-know-about-the-white-houses-new-ai-manhattan-project

Following the Trump administration’s further cuts to tech contracts and consultancy firms, it has ordered changes to software development. The demands include mandatory inclusion of AI and favor large, established companies with their own data centers over new startups.

https://codewave.com/insights/trump-ai-policy-software-developer-opportunities/

Leave a comment

Blog moderation policy

Login

Allowed HTML <a href="URL"> • <em> <cite> <i> • <strong> <b> • <sub> <sup> • <ul> <ol> <li> • <blockquote> <pre> Markdown Extra syntax via https://michelf.ca/projects/php-markdown/extra/

Sidebar photo of Bruce Schneier by Joe MacInnis.