* Posts by johnrobyclayton

84 publicly visible posts • joined 11 Jan 2018

Page:

Google dragged to UK watchdog over Chrome's upcoming IP address cloaking

johnrobyclayton

Now we need open first stage proxies

Google is providing the first hop proxy.

I see no reason why other organisations could not also provide the first hop proxy and browsers designed to implement this feature with a configurable first hop proxy.

Add in a configurable list if first hop proxies with a shuffle option.

Like having a list of DNS servers with a shuffle option to prevent any one server from having a complete list of dns requests that your computer is sending.

It may be a technology that Google is developing for its own benefit but at its core is something useful that could be extended into a powerful privacy enhancing tool.

EU lawmakers scolded for concealing identities of privacy-busting content-scanning 'experts'

johnrobyclayton

How to hunt predators

Ask any hunter and they will tell you.

You hunt predators by stalking the prey. Not by looking at predators.

Predators are hiding from the prey and hiding from each other.

Wait at the watering hole.

Wait on the game trail.

Look for big herds of prey.

Hang around them until the predator shows up.

If you want to protect children from predators,

sit next to them while they play on line,

hang out with them when they play in the playground,

converse with them when they are talking,

shop with them when they are at the mall,

be with them every day.

In other words, be a good parent/guardian for every child you want to protect.

Watermarking AI images to fight misinfo and deepfakes may be pretty pointless

johnrobyclayton

Only trust what you know, Know who signs

Deepfakes are an issue only for the stupid that believe what they see and those affected by the consequences of the stupid believing what they see.

Deep fakes are just visual candy. There is no need to go around believing it.

It is a choice only believe images that have been digitally signed with a modern robust encryption scheme.

It is choice to only make decisions based on data whose provenance can be proven to come from a known reliable source.

Digital signatures all the way.

If AI can defeat modern robust encryption then we have bigger problems than deep fakes.

US AGs: We need law to purge the web of AI-drawn child sex abuse material

johnrobyclayton

There is a question to ask here

Do we want to prevent any form of exploitation of children?

Or

Do we want to prevent any form of enjoyment of sexual pleasure from perceiving information that is perceived by anyone else as representing children?

The first is clear and easy to concretely describe.

Laws prohibiting the first relatively are easy to frame with little risk of unintended consequences.

Laws prohibiting the second are open to almost infinite feature creep and framing laws for this are almost certainly going to have unintended consequences.

Digitally aging images of a missing child so that they can be identified years after they disappeared has been used to digitally reverse aging of adult entertainers to produce what appears to be child pornography.

No children exploited here. This has been around for years.

Drawing cartoons of children to produce what appears to be child pornography. There are lots of artist that can draw a picture of what looks like a child without needing a child as a model.

No children exploited here. This has been around for millennia. I am sure there were a few statues in the Grecian Empire of individuals we would identify as underage without the carver needing to use a model.

Drawing a cartoon of a bandicoot or a fox with a flying tail of an indeterminate age, but looks like the character in popular children's entertainment might be identifiable as something that can be experienced as child pornography.

No children exploited here, though there might be some copyright infringement. There is a surprising amount of such content available.

Creating a customizable doll.

Lots of dolls and action figures with bendable and pose-able limbs.

Lots of dolls/action figures that can have their configurations changeable. Mr and Mrs Potato Head.

Lots of dolls/action figures that can simulate a number of biological behaviours.

Wetting themselves, sucking on a bottle, crawling, walking, speaking phrases, speaking context aware phrases, full on human language interface.

Writing a story about suck a customizable doll/action figure.

Drawing a cartoon or creating animation about such a doll/action figure. Astro Boy.

Creating Pornography about such a doll/action figure that that is brand new from complete cloth.

There is no child exploitation here, but there is an almost infinite variety of pornography that can be imagined and produced.

The doll/action figure/android/robot does not even have to look human for someone to perceive them as something that is recognizable as a child.

There is simply no limit on what can be objectionable just as there is no limit on what people are capable of enjoying.

There are an indefinite variety and indefinite count of individual that are capable of enjoying experiences that can be achieved through various forms of child exploitation. The only reason it is not infinite is that the population itself is not infinite.

This does not mean that these individuals necessarily need to exploit any children directly to enjoy these experiences.

Distracting ourselves by trying to detect, track, legislate against, prosecute, incarcerate all of these various types of people and activities will dilute our focus on protecting children in our care (that's all of them)

Going after this infinite variety of people is low hanging fruit for legislators and law enforcement bodies. It allows them to crow from the rooftops that they are working hard to protect everyone from everyone else that can outrage and infinite variety of outrages. They can do this using easily applied technical tools to track and trace and spy on everyone.

Focus on the hard work of protecting children directly instead of sticky beaking on everyone in the hope of finding something that outrages someone.

AI-created faces now look so real, humans can't spot the difference

johnrobyclayton

Re: Detail and perception

I have never upvoted before.

This one made me laugh out loud.

QM, AI, LOTR, All in one funny package

University students recruit AI to write essays for them. Now what?

johnrobyclayton

First step seems to work.

In https://chat.openai.com/chat:

Question:

What is a good bibliography for a paper on the use of AI to generate academic papers?

Response:

There are many sources that you can use as a bibliography for a paper on the use of AI to generate academic papers. Some possible sources include:

"The Role of Artificial Intelligence in Scientific Communication" by Mark C. Wilson and Vincent Larivière, published in the journal PLOS ONE.

"Automatic Generation of Scientific Papers: A Review" by Rensheng Wang, published in the journal Frontiers in Artificial Intelligence.

"An Overview of Artificial Intelligence in Scientific Communication" by K. S. R. Anjaneyulu and M. S. S. Krishna, published in the journal Current Science.

"Toward the Automatic Generation of Scientific Papers" by Yves R. Kergosien and Marie-Christine Rousset, published in the journal Artificial Intelligence Review.

"Artificial Intelligence in Scientific Communication: Opportunities and Challenges" by Sabine Köszegi, published in the journal Science and Technology Libraries.

"Automatic Generation of Scientific Papers: A Survey" by M. A. Hossain and M. S. Rahman, published in the journal IEEE Access.

These sources provide a good overview of the current state of the field and discuss some of the key challenges and opportunities that have arisen as a result of the use of AI in scientific communication.

Anti-money laundering bill targeting cryptocurrency introduced in US Senate

johnrobyclayton

At last a little bit of competence

quote-->

In a summary of the bill, Senator Marshall's office said unhosted crypto wallets, meaning any that are controlled by an individual – such as cold storage on a flash drive, or in an offline wallet – are a serious regulatory gap that needs closing.

Such wallets "allow individuals to bypass AML and sanctions checks," the pair said in the summary, and the bill takes action to ensure such wallets aren't black holes for regulators.

-->

The only way to be safe when using cryptographic wallets is to start with personally managing your own wallet. That is just a start though.

This is the most significant part of the legislation. Everything else they have thought of trying is a bit silly.

Digital currency and blockchain do not in and of themselves provide anonymity.

There is a way to be anonymous though.

Use one time spending wallets.

If you have some digital currency and want to spend it.

Create a new wallet.

Transfer the value you want to spend into the new wallet.

Transfer the whole wallet to the entity that you are paying.

The payee then receives the value by creating their own new wallet and transferring the value from the spending wallet to the receiving wallet.

The payee can then store the receiving wallet until they need to spend something.

If the value in the receiving wallet is the same they want to spend then they just pass the receiving wallet to the new payee.

If the amount to spend is greater that what they have in one spending wallet they they can transfer a number of spending wallets to the new payee.

If its less then they can create new spending wallets and transfer the appropriate amount of value to the new spending wallets.

If everyone does this then there will be no clusters of linked transactions that can be linked together to identify the individuals that are related to the transactions.

Then there is just wallets being transferred from place to place and each being used only once.

Value simply gets recorded as going along a single path of wallet to wallet to wallet.

Anonymity is a bit more fiddly and a bit more work but it is not difficult to do.

A little bit of competence on behalf of legislators and regulators deserves a little bit of competence right back at them.

You get the internet you deserve

johnrobyclayton

Maybe there is hope

AI's generating content that only has a little bit of useful information.

Ad networks that advertise on content and pay for impressions and clicks based on how many searchers find it.

Search engines looking for useful stuff to offer searchers to make themselves relevant to searchers.

It is one big generative adversarial network.

They start out shite, sure,

But they do improve...

How GitHub Copilot could steer Microsoft into a copyright storm

johnrobyclayton

Explain how/where you got your code Vs Explainable AI

These are the same problem.

To solve this would require a lot of work in the initial setup of the training data.

You will not be able to solve this using training data that is not completely vetted and cleansed.

The training data needs to be correct and complete and fully attributed.

Unfortunately not something that is effectively done in today's mad scramble to create cool toys.

Code sample x made by author y does z using language w

Code Sample:

Code:

---

x

some piece of code

---

z

What this code does:

---

Some description of what this code does,

description of inputs,

description of outputs,

description of purpose

---

Language:

---

w

What language is this in?

---

Author:

---

y

Author of this code, Available licenses for this code.

---

For creating the code generator you need to generate suggested x given z and w

Given:

---

Description of what this code needs to do

---

z

What inputs are available?

What outputs are required?

How should it behave?

---

Language:

---

What Language are we generating in?

---

This produces generated code.

Questions about this provided code are:

What code samples from the training data are most likely to have contributed to the output.

What are the authors and what is the licensing terms of the code that most likely contributed to the generation of the output.

Generating code samples we seem to know a bit on how to do.

Identifying what code samples most likely contributed to the generation of the output is a second challenge that could be met by further research. This would likely require extensive supervised training using sample Training data-->generating sample output that human evaluators could then compare to the training data and identify the likely contributors Enough of this could train a model at identifying likely contributing code samples given generated code a a training corpus just like models can be developed for attributing purposes. Like identifying if some piece of prose might have been written by Shakespeare.

This can then be linked to the author and license information linked to the training code samples.

This is a vastly greater amount of work than just training a model on any tom dick and harry code set scraped from an essentially random source.

There is no easy shortcut way of doing this unless you are happy with the untoward consequences of not caring about any consequences.

Ever suspected bankers used WhatsApp comms at work? $1.8b says you're right

johnrobyclayton

The security and anonymity options in communications technology are only going to improve

It is a fiercely competitive market.

Everyone will have to adapt to the reality of not being able to monitor, track, trace or even be aware that communications between any pair of communicators is taking place.

It is even possible to have communications take place without the communicators being capable of knowing who each other is unless identity information is exchanged.

There are a lot of laws and regulations that are completely or partly dependent on there being some way to determine some details of a communication.

These need to adapt to an environment where this information is simply not going to be available.

Regulators and legislators can grump and moan all they like (it is always amusing to read about it on this site), but it is not going to stop completely anonymous, secure and private communications becoming more and more of a reality.

If you cannot make legislation or regulation that can handle a complete inability to monitor communications then you simply will not be able to enforce such legislation or regulation.

If you are in a position to need the protections of legislation or regulation that need to be able to access elements of communication that cannot be compelled to be available, change your position.

Tech world may face huge fines if it doesn't scrub CSAM from encrypted chats

johnrobyclayton

I am a bad actor - please help

Can someone send me all of these hash databases and deep learning models that are being developed to identify bad files or content?

For file hashes I can create innocuous files whose hashes collide with with bad file hashes, scatter them on social media, and tie up investigative resources.

For deep learning models that identifies bad content I can create an adversarial deep learning model that can generate content that the supplied deep learning model identifies. I can let the government provide the training tool for automated generation of bad content.

There are a lot of silly people that think that the range and options of information available can be constrained. It is disappointing really.

UK government opens consultation on medic-style register for Brit infosec pros

johnrobyclayton

Not to worry, its a long consultation

20th March 2345 is a long way away. Not much to worry about for a few centuries.

Hitting underground pipes and cables costs the UK £2.4bn a year. We need a data platform for that, says government

johnrobyclayton

This reminds me of a story

Several years ago I read a story of a Uni student who decided to try an map the various bits of cabling that existed in a particular location from public sources of information.

He observed some risky conglomerations of various bits of infrastructure on his map.

He observed that one particular bank had all of its communication infrastructure going through one point accessible by a manhole cover a little bit down the street and around a corner.

He went to the bank and asked then if they were aware of this vulnerability. The bank got very upset and would not let the student leave the building.

Some alphabet agency got involved and they promptly hired the student and got him to do the same project on a national scale.

What an IDORable Giggle: AI-powered 'female only' app gets in Twitter kerfuffle over breach notification

johnrobyclayton

Being a minority

We have done a lot of work to enable minorities to be included in our rich Internet culture, and it has only made our Internet culture richer.

There has been a significant downside though.

The Internet was initially created and populated by a group of people that to a large degree were conditioned to believe that they were the apex predator in the room. Flame wars, Religious crusades, Trolls, they always existed, either in reality or in perception but because a large number of us knew that we were the apex predators in the room, we did not care.

The inclusion of every minority we could find or be discovered by, though, introduced a lot of people to the Internet who have been conditioned that they are not the apex predators in the room.

They are conditioned to take seriously every indication that they are under some form of attack, whether it is real or not.

The rough and tumble of the Internet is a big part of its richness. The determination by a lot of well meaning individuals and groups to cover all of the hard and pointy surfaces with polite, gentle and politically correct foam is reducing the cultural richness of the Internet.

We need to teach more people that the belief that you are the apex predator in the room is simply a state of mind. You do not need to be the apex predator to believe it, and the belief can give you the freedom to simply ignore the sharp, uncomfortable and pointy bits of the Internet you might not enjoy and enjoy all of the richness that remains.

Facial-recognition algos vary wildly, US Congress told, as politicians try to come up with new laws on advanced tech

johnrobyclayton

There are no shortcuts

When

They have 10000 different images of every person to be positively identified (scanned image is of this person)

And

They have 10000 different images of every person that they want to negatively identify (scanned image is not of this person)

And

They have 10000 different images of each person they are going to scan and attempt an identification on.

Then

They might have a chance of reliably identifying someone that they have scanned.

It still will not be perfect though.

Post Office faces potential criminal probe over Fujitsu IT system's accounting failures

johnrobyclayton

Re: Ooooh first post....

'Ere, I resemble that comment, except its Windows 2000 Advanced Server slipstreamed with SP3 and the registry hack that made the evaluation versions included in the MS Cert training packs full versions. Ahh the good old days. Devuan rules.

Put the crypt into cryptocoin: Amid grave concerns, lawyers to literally dig into exchange exec who died owing $190m

johnrobyclayton

Crohn's Disease + Indian Food = Easily Faked Death

If you have Crohn's disease and want to fake your own death, India is the place to do it.

The immovable object versus the unstoppable force: How the tech boys club remains exclusive

johnrobyclayton

Maybe it is time to move on.

I am quite fine with competing purely based on merit and effort.

I know that a lot of males in my industry have behaved in a way that allowed them to succeed in a way that does not correlate with merit and effort.

I understand the reaction to this that promotes the idea of enforcing a method for females to enable them to succeed in a way that does not correlate with merit and effort as a valid response.

I do not need to succeed beyond what I earn by merit and effort.

I also do not need to stick around in an environment that prevents me from being able to succeed in line with my merit and effort.

So maybe its time to create a new industry. There are enough mechanisms now to enforce anonymity while supporting the economic, informational and business transactions to support fully anonymised informational industries. Being fully anonymised, gender cannot become an issue.

Pair programming? That's so 2017. Try out this deep-learning AI bot that autocompletes lines of source code for you

johnrobyclayton

Can I get one trained on homework assignments?

There is a lot of training data available online apparently.

Would this be plagiarism?

DeepNude's makers tried to deep-six their pervy AI app. Web creeps have other ideas: Cracked copies shared online as code decompiled

johnrobyclayton

This is only the start

Some ideas that will definitely crop up if they have not been built already:

Chatbots that emulate children for the enjoyment of sexual predators. (These already exist in the hands of some law enforcement groups to identify and trap such predators.

Deepnude+deepfake videos = Porn films with the subjects of your choice.

Text story to Screen Play / Screen Direction conversion using AI

Screen direction / Screen Play to video /animation with 3D imaging using AI

Add force feedback

Add them all together and you will have the full Startrek Holodeck Experience.

It is inevitable that this will be used for any and all sorts of perversions along with entertainment, training and every other useful and culturally appropriate use.

There will be grumbling and crying and gnashing of teeth but that is not going to stop the torrent of new ideas and technologies that will explode to satisfy any and all perversions.

Bot war: Here's how you can theoretically use adversarial AI to evade YouTube's hard-line copyright-detecting AI

johnrobyclayton

Try this the other way around

Instead of breaking copyright in the research, get some original content that you own and send that through the neural net to make it trigger a copyright flag.

A lot more fun as you get to beat up the copyright enforcer instead of getting beat up.

What bugs me the most? World+dog just accepts crap software resilience

johnrobyclayton

Idiocracy

I remember a scene in the movie Idiocracy.

It was in the triage section of the emergency department of a hospital.

There was a girl in front of a full graphical patient classification interface.

It had pictures of red marks on a picture of a body and she would select which one that most matched the problem described.

It had all the complexity of the classic game operation without the requirements for good hand eye co-ordination. The icon pictures were rather large.

I have worked in support for many point of sale companies.

I have seen the evolution of this type of software on a daily basis. I have heard comments from the owners of more than one business that they wanted software that someone who was not very good at reading to be able to use.

Users get stupider and stupider because they are allowed to by companies that want to sell to the stupidest people they can find because these are the ones that will pay the most for software no matter how many issues it has.

The demands on software are increasing every day not just by the increasing complexity of our operating environment but by the requirements that stupider and stupider people need to be able to use it safely and profitably.

The problems of allocating blame for software that fails to meet requirements is a difficult one because it is the responsibility of everyone that contributes to its development, design, testing, marketing, use, selling, purchasing, training, legislating, securing, compliance, etc etc etc ad infinitum.

Click here to see the New Zealand livestream mass-murder vid! This is the internet Facebook, YouTube, Twitter built!

johnrobyclayton

You get what you...

You get what you pay for

Facebook, Youtube etc, are free or cheap enough, except for your attention and the information about yourself that you give up.

You get what you vote for

If you want something different, vote differently, if you do not like the choices, run for office yourself.

You get what you do nothing to stop

Evil happens when good people do nothing to stop it irrespective of what their opinions on what good and evil are.

If you do not like a company's behaviour:

Go to another company.

Create your own that out competes them that behaves as you see fit.

Buy them and change their behaviour.

Vote for legislators that create the regulations that enforce the behaviours you require.

Become a legislator and vote for the regulations that you require.

Do something that stops behaviours that you deem evil. Convince others to prevent evil from being done. Share your attempts with others.

You might want to create a manifesto in this case, just so people can understand your choices and the intent of your actions.

As has been said before, there are no easy options here.

There are plenty of easy to state wishes that are in a lot of cases, especially now extremely heartfelt.

But as has been observed many times before magical wish based thinking and requirements in the political/technological/social/environmental spaces serve no useful purpose.

It is only the choices that we make, the actions that we engage in that makes the world what we want. We live in a world full of other people that are choosing and acting as they see fit.

If these events prompt you to choose something or do something, then make your choice. Act.

Strewth! Apoplectic Aussies threaten to blast noisy Google delivery drones out of the sky

johnrobyclayton

Not all that usefull in urban areas, but in the sticks ...

Trialling this in an urban or suburban area is a mistake.

Too many neighbours to bother and it would only be used by people too lazy or time poor to go to the local chemist or corner store.

Providing deliveries to people in remote areas would avoid the noise complaints. Only the target of the delivery is likely to hear the drone and it would be serving a definite need for remote families.

Might be better to use a hybrid fixed wing - vtol drone configuration that can handle significant distance and weight while maintaining landing accuracy.

Did you know?! Ghidra, the NSA's open-sourced decompiler toolkit, is ancient Norse for 'No backdoors, we swear!'

johnrobyclayton

Perhaps they have moved on

Once you have pwned the compilers and other elements of the toolchains that build the toolchains that build the toolchains that ....

To get a clean from start you would have to wire up the processor from transistors, design and build your chip fabricators, code your compilers and bootstrap yourself into the modern age.

Not many people or organisations have the patience, ability or resources to go through such a process.

One iteration of Linux from Scratch was enough for me.

There is no way to fully trust that any technology today has not been pwned to some degree or another.

The only defense is to get as many different individuals and groups investigating and testing in as many different ways as possible.

It is probable that the NSA's largess is not something that is a significant threat to the NSA itself.

It will be though, something that can increase the exposure of their counterpart's efforts along with those of non-state threat sources.

When the bits hit the FAN: US military accused of knackering Russian trolls, news org's IT gear amid midterm elections

johnrobyclayton

The IRA in news is dead, Long live the IRA

Irish Republican Army

Internet Research Agency

Isn't it nice when the current affairs torch gets passed on.

Fool ML once, shame on you. Fool ML twice, shame on... the AI dev? If you can hoodwink one model, you may be able to trick many more

johnrobyclayton

AI Buster Buster Buster Buster

Train an AI to recognise something.

Train the next AI to fool the first AI.

Train a third AI to recognise the attempts to fool the first AI

Train the fourth AI to fool the third AI

Train the fifth AI to recognise the attempts to fool the third AI

Wash, rinse, repeat

Tech sector meekly waves arms in another bid to get Oz to amend its crypto-busting laws

johnrobyclayton

If anyone is interested in my thoughts on secure communications

Some thoughts that I have had on anonymity and security.

https://github.com/johnrobyclayton/SecureCommunicationsInfrastructure/wiki

Oz government rushes its anti-crypto legislation into parliament

johnrobyclayton

I have been thinking

Elected governments can only promote legislation that can be understood by, and desired by the voters.

There are plenty of technical people that know and understand the futility of effectively controlling decent encryption technologies.

But there a lot more people that do not understand the impossibility.

Therefore elected governments have to say silly things like the law of the land overrules the laws of mathematics even though that is cringe worthy.

What is needed is a description and demonstration of a secure communication infrastructure that is as impossible for any governments to effectively control as possible. This needs to be as simple in the individual operational elements as possible. There may be a lot of operational elements but if each piece is simple enough then a lot more people will be able to understand it.

I have been writing up something to do this as a hobby for the last couple of years.

Its a bit of a read and is a work in progress but I think that there is enough for sharing.

I have been using a github wiki for this:

https://github.com/johnrobyclayton/SecureCommunicationsInfrastructure/wiki

Brain brainiacs figure out what turns folks into El Reg journos, readers

johnrobyclayton

C, C++, C#, C99 shaped nubs?

Explains a lot.

NSA boss: Trump won't pull trigger for Russia election hack retaliation

johnrobyclayton

Instead of attacking the predators, reduce the vulnerability of the prey.

There are always going to be enemy's trying to attack.

In this case they are attacking through a vulnerability in how our civilisation works.

We can counter attack or we can reduce the vulnerability.

Improving the critical thinking capabilities of our populace and improving their political awareness and comprehension would be an effective method of reducing our vulnerability.

Allowing yourself to be effectively lied to is a choice. Never trust, always question, always verify, always fact check and never let the desire to fit in override your good sense.

Hospital injects $60,000 into crims' coffers to cure malware infection

johnrobyclayton

Predators always exist in the presence of prey

If you choose to be prey then you encourage the existence of predators to the detriment of everyone.

Choosing to be prey is not taking preparatory steps to avoid being taken advantage of as well as the immediate surrendering of value to predators.

A Hospital holding up their patients to defend themselves from the righteous criticism of their behaviour is deplorable.

What do we want? Consensual fun times. How do we get it? Via an app with blockchain...

johnrobyclayton

Objective Measurement

Receiving consent is required before engaging is sexual activity by another.

This is enforced by law and evaluated in the courts.

At this point in time, as far as I am aware, there is no generally agreed upon standard for objectively measuring or determining consent prior to engaging in sexual activity with another.

Therefore, as far as I can see, any sexual activity between two or more people must be defined as suspect until a legally binding statement has been recorded by all participants that they did indeed consent to the sexual activity.

This is insufficient.

Consent has a very complex definition and it is getting more complex all the time.

It includes any one or any combination of:

State of mind of any of the participants, before during and after

State of mind of others, before during and after

Age of any participants

Information available to any of the participants

Information provided by any of the participants

Differing local statutes

The list goes on ...

I for one do not like an environment where the appropriateness of my sexual activities with others simply cannot be defined before or during the event and can sometimes only be guessed at after.

You might dislike the solution presented in this article.

If so, present a better solution to this problem.

Romance, love and tenderness are all fine and dandy but they are a pile of crap in a court of law.

Page: