Comments
Winter • January 14, 2023 1:11 PM
@modem
Re: Is there a god?
I actually knew that story.
I always interpreted the Rise of the Machines as the old fear of the uprising of the slaves and underclasses.
The Robots of Karel Čapek, the Golems of folklore, the Morlocks of Wells, all descendants of the uprising of Spartacus.
Also, the alien invasions in SF are the sublimation of the fear that some other people will do to the Americans (and British), what the Americans and British did to the natives of the lands they conquered.
echo • January 14, 2023 2:19 PM
War and slavery existed before Europeans turned up. I am more minded to trust science and law before anyone regardless of their status. They are, in theory, a benchmark which can protect us from the worst of ourselves. Yes, both can and have been misused but generally we are at a stage where we should and do know better. Mostly.
As much as I dislike the unscrutinised happy clappy kind of narrative (often used by various politicians and on the make “entrepreneurs” I also dislike the, mostly American, tendency towards doomerism is sci-fi. I personally find the majority of US based sci-fi output unreadable and unwatchable for both reasons.
I’m really not that bothered by rogue AI. In fact I feel the current climate change and nuclear blackmail issues in the in-tray are useful learning opportunities. At some point there will come a tipping point: anyone mishandling AI (or similar hot potatoes) is simply going to earn themselves restrictions on practice, a long spell in a jail cell, or if they are really naughty and putting people in danger a unilaterally imposed swift execution.
There’s just so many ways to stop this whole thing before it gets out of hand that don’t even involve an direct law or action it’s not funny. The thing is all that’s rather dull and doesn’t make a good scary story or news bulletin.
If any moustache twirling billionaire evil villain with a private island tries it on they need to recruit without getting caught, source parts without getting caught, get past the environmental policy at the planning department of the local council, explain away their power usage, and so on and so on. There’s a million ways to get away with it on paper but in the real world not really.
Ted • January 14, 2023 3:20 PM
@echo
There’s a million ways to get away with it on paper but in the real world not really.
I think that’s an interesting point. So many actions, especially these days, leave a trail.
I’m going back to the decision Bruce linked to a few days ago, about the judge who ruled that a geofence warrant was unconstitutional.
What I’m amazed at is the level of research and detail that are included in her ruling. The process took more than two years and was the first to comprehensively examine the pros and cons of geofence warrants.
I haven’t read the whole decision, but did note that Google is working with various groups to figure out how to apply the laws. From the ruling:
To ensure privacy protections for Google users, … Google instituted a policy of objecting to any warrant that failed to include de[-]identification and narrowing measures… Seemingly developed as a result of Google’s collaboration with CCIPS, this de-identification and “narrowing protocol … typically includes a three-step process.”
https://s3.documentcloud.org/documents/21338805/2022-3-3-opinion.pdf
Ultimately, the judge thinks the future of geofence warrants should be taken up by lawmakers.
Clive Robinson • January 14, 2023 5:43 PM
@ echo, ALL,
Re : Size of cause v. Size of effect.
“If any moustache twirling billionaire evil villain with a private island tries it on they need to recruit without getting caught, source parts without getting caught, get past the environmental policy at the planning department of the local council, explain away their power usage, and so on and so on.”
That is based on the incorect assumption that,
“Big effects require a big cause”
Mostly they do not when you go beyond the purely mechanistic.
The ability to design a self replicating biological agent at home is almost here, it’s already been done in the lab for several years[1]. If nothing else then the last three years should have informed people of just how easily pathogens mutate in the home, then spread in the community and then around the globe.
But further consider the writing of computer malware, it requires little more than research and a computer, and has frequently been done by teenagers working at home on an Xmas / Birthday gift from their parents.
The reality that nobody wants to talk about, is that “Size of cause and Size of effect” are almost entirely independent of each other. What counts the most is the sensitivity of the environment the effect occurs in, along with how easy it is to create the environment.
The more sensitive the environment the less energy is required for each member of the environment to move from one state to another. If the process of a member moving from one state to another releases enough excess energy to cause further members to have changes of state, then it becomes self replicating and a chain reaction can occur.
If it does occur, it almost always follows an exponential rate of change unless subject to a form of moderation that modulates the replication rate.
Which leaves how easy it is to create a sensitive environment, generally it’s easy enough for any fool to do.
One of the reasons people do not wish these to be talked about, is that all human progress sits on the cusp of such changes of state being moderated to within limits.
That is the closer to the edge, the fasyer the progress and the greater the productivity thus potential for profit. The cost of moderation however significantly reduces profit and often productivity.
The problem with this is that the majority of people unaware of what constitutes a “sensitive environment” often go ahead and unwittingly create one (have a look at the cleaning products under the average kitchen sink, or hobby related materials in the garage etc).
Thus on average more people get hurt by the lack of knowledge of those around them than people being hurt by those who have the knowledge[2]. The result is that many get away with having the harms they cause being described as “accidents” not “negligent” or “criminal” behaviour that they should be.
If people know then two major effects come into play,
1, They tend to avoid doing or limit the things that cause harm.
2, They raise question when others around them do the things that cause harm.
Better educated and better informed individuals tend to consider their lives, and the lives of their loved ones, more highly because of their improved knowledge and education.
The more highly an individual values their life, the less likely they are to do harmful things that effect them or importantly others around them.
I could go on at length about the inherent security value of being better informed. But that brings up the consideration of,
“Once you start, where do you stop?”
[1] Often called “Gain of Function Research”(GoFR) the impression often given is the need for large expensive labs. Where as the first such experiment a little over a decade ago used cages of ferrets and a technician “manually infecting” them untill sufficient change occured for the pathogen to become transmissable by airborne droplets.
[2] If people do the research they find a nasty effect by corporations which is “Hazzard Exportation”. The cost of doing things in less sensitive environments or in more strictly controled environments is high. The effect of such regulation is not to stop things being done, or make them done more safely, but for the things to be moved to less regulated, often more corrupt places. This gets excused away under the notion of “unintended consequences” often a sign that the regulation had been “got-at” for political reasons.
echo • January 14, 2023 6:00 PM
@Ted
Buggerit. I just closed the page I was drafting a reply. I’m not typing all that again!
The judgment was interesting and a couple of comparisons can be made between US law and European law under the European Convention. There’s an issue of the invasiveness of corporations versus safeguarding the public, remedy, and proportionality and legitimacy, exceptional circumstances, and qualified rights and so on. I think the judges conclusions that the question is for legislatures is correct. I don’t agree with the absolutism but the original warrant did seem overbroad without sufficient justification. That said denying suppression of evidence was pretty fair too.
With regard to Bruce’s linking to a claim the search is unconstitutional it’s more accurate to say the search in this specific case was unlawful. Law and guidelines and procedures seem quite under-developed, and those arguments can get quite windy as the judge indicates. “Blanket just because we can” tracking may be unconstitutional but there’s also an awful lot of space between threat versus proportionality it probably doesn’t quite mean what journalists think it means. That’s what happens when journalists don’t read the judgement thoroughly. The judgement is more of a “Get your sloppy assed shit together” instruction than anything else.
echo • January 14, 2023 7:39 PM
@Clive
That is based on the incorect assumption that,
“Big effects require a big cause”
I never said that. You assumed I said that. Different thing. I’m well aware of bathtub WMD and the like.
The point is the public policy questions and administrative questions are really going to limit anyone doing anything too stupid with AI. It may even be regulated with a scale like drugs. Computers and networks of computers can provide enough metrics to track unusual activity, or who is friends with who or has been in the same room with who. The list goes on.
So your soldering iron slips and you accidentally create the world’s first super AI with an evil streak. Then what? It’s not going to be able to do a lot or will trigger a lot of tripwires if it does.
There may even be regulated containment rules at some point. Good luck escaping a lead lined room with a large axe next to the power supply lead.
Climate change and nuclear blackmail provide useful starting points to begin discussing the ins and out of crafting a policy position. WMD too and Crispr and all that jazz.
It’s none glamorous and dull work. One assumes people with half a clue will be involved not random journos or experts on everything who can’t catch the sky falling on them.
AI may evolve but like a lot of people forget everything else evolves too. The evil AI of tomorrow will be facing the policies and mitigations and infrastructure of tomorrow not the half baked nonsense of today. I’m not saying there is zero risk. I just think people tend to seize on the narratives written by journos and amateur sci-fi authors and their brains go faster than their smarts. They drive past a bazillion realities and throw a million hypotheticals around but their position is untested fantasy with a million holes baked in. There isn’t much a super AI could do today even if switched on. There’s just too many breaks in the way but I’m leaving that for the dull and boring desk jockeys to work on.
It’s all pretty much implied in what I said.
Clive Robinson • January 14, 2023 9:53 PM
@ echo, JonKnowsNothing, Gunter Königsmann, SpaceLifeForm,
“I never said that. You assumed I said that.”
What you said and I quoted it was,
“If any moustache twirling billionaire evil villain with a private island tries it on they need to recruit without getting caught, source parts without getting caught, get past the environmental policy at the planning department of the local council, explain away their power usage, and so on and so on.”
All of which are, that for your presumed “moustache twirling billionaire evil villain” to do his big event, they will have to muster sufficient resources to be so noticable they can not be ignored thus they will get stopped.
Hence you were to paraphrase, saying,
“Big effects require a big cause”
And that those
“Big causes would be the moustache twirlers undoing.
I pointed out that we actually know from real world experience this not to be true.
We know and have done for years you can bring the Internet down with a worm or similar it’s happened several times but the Morris Worm in 1988 and WannaCry in 2017 are two that are memorable due to the US NSA being the source of the vulnarabilities used.
In all worm cases and most modern malware it just requires,
1, Knowledge from research.
2, The ability to write code.
Neither of which are in anyway “red flag” events.
Also we know that the “launch point” could be any of several million public access Internet points. Many of which such as those in small cafés have no in building CCTV or Internet Security, and can also be reached from quite far outside the cafés buildings anyway. So the chance of “identifying the perp” is negligable at best, and we know of many ways the odds could be made better such as a variation on “The Moscow Rock”[0].
Likewise many millions have been killed by “Varients of Concern”(VoC) of C19 all of which came about in peoples homes, with no fanfare, or red flags like fancy lab equipment, etc or other indicators than the change in “community spread” which was pretty much to late (for political and criminal reasons).
I also noted that just over a decade ago, the first of the so called “Gain of Function Research”(GoFR) experiments was carried out with just cages of ferrets.
In case it missed your attention at the time, @JonKnowsNothing, @SpaceLifeForm and myself had long discussions about the same thing happening with “Mink Farms”. I had serious concerns that SARS-2 would get from “farm stock” to “wild life” which then became “disease reservoirs”. Note the times of our concerns and that it duely went on to happen in various ways. I was also very concerned about domestic rodents and what would happen with them, likewise before the entirely new strain based on domestic mice came out of Africa and dominated the world in fairly short order (thankfully as it’s increased infectiousness came with reduced pathogenicity).
All of that could have been done in open sight and nobody would have noticed or cared[1].
But the point I was making was the exponential or chain reaction growth of such events.
I’ve already pointed out earlier today[2] in answer to @Gunter Königsmann just how the use of Search-AI and Chat-AI could start changing the internet in a way that could produce rapid exponebtial growth.
So yes I do believe with enough reason and logic that,
1, It can be done.
2, It won’t be seen untill to late.
3, It won’t get stopped easily if at all.
I guess time will tell which of us is right
[0] The Moscow Rock from more than a decade and a half ago would have been funny but for the fact we have reason to believe people lost their lives and liberty over it,
https://www.theguardian.com/world/2012/jan/19/fake-rock-plot-spy-russians
[1] Fun fact, the James Bond Film OHMSS loosly based on the sixty year old Fleming book, had the hero Bond face arch villain Ernst Stavro Blofeld (played by Telly Savalas of later Kojak fame). Blofeld is planning to hold the world to ransom by a threat to render all food plants and livestock infertile through pathogens he has created. In the film he describes their creation and also how they would be distributed by a group brainwashed hypnotized young women in his “clinic” who would take the pathogens around the world to live stock shows and the like…
[2] My reply to @Gunter Königsmann,
Ted • January 14, 2023 11:55 PM
@echo
The judgement is more of a “Get your sloppy assed shit together” instruction than anything else.
Yes, precisely. I just finished reading the decision. It’s a detailed illustration of technology and investigative techniques jauntily outpacing a legal framework.
Suffice it to say, Google collects hundreds of location data points on users everyday. And most people are faintly aware this is a “feature” they have opted into.
On top of this, the courts aren’t providing certain guidance on how the geofence warrants should be executed, or what objective criteria would allow for progressively invasive searches.
The judge makes one particularly wild observation that gets a paragraph on p27-28 but no further attention.
There is something called a ‘search warrant return’ that an officer files to notify the court what info they have received back from a warrant. “Inexplicably” the detective filed this return a day before he even sent the geofence warrant to Google. It contained a “sizable amount” of precise location info for at least 19 user devices.
This is same number of de-identified devices that Google actually sent to the detective under Step 1 of their process.
JonKnowsNothing • January 15, 2023 2:18 AM
@Ted, All
When doing a deep dive into a USA court opinion, you might want to study up how Marcy Wheeler does her work over on EmptyWheel (.net).
Marcy has carved out a specific detailed method of analyzing court proceedings. From the day to day transcripts, the filing of warrants, submission of evidence. Details on how to read BATES numbers, which are the old-fashioned hand pressed number stamps, used to mark each item as evidence and which department or group submitted the item(s). It’s an interesting classification system.
IANAL
Courts do not create evidence (normally). They do not do primary or secondary research. That is the job of the prosecution (unlimited funding) and sometimes the job of the defense (very limited funding).
Some courts and judges have clerks that help them collate and clarify all the aspects of the law(s) and sometimes provides a topic list as it relates to a particular case. The SCOTUS clerks get a the most publicity.
Judges rule on what is presented and compare that with the law(s). Depending on the type of case, appeals may happen but only based on the transcript of the case. If you want to really understand what took place, The Transcript is what you want to read. Not just the opinion.
Application of the 4th Amendment has been eroded for some decades now. The key is a comma space in the statement regarding the need for a warrant.
- Warrants are only needed for UNUSUAL searches
- Any searches are considered NORMAL do not need a warrant
- There are no more UNUSUAL searches; all searches are NORMAL; therefore no warrant is needed
SCOTUS current rulings have led Congress to create the secret FISA Courts. From FISC comes the authorization to use boomerang warrants targeting an international search that boomerangs back to target people inside the USA who would not be covered by a such a warrant.
There won’t be reason to celebrate until or unless SCOTUS changes their rulings on this topic and Congress disbands the FISA Courts and amends the laws that created it.
Don’t hold your breath, that has as much chance of happening as the public release of the full Senate Report on CIA Torture (GITMO).
echo • January 15, 2023 8:04 AM
What you said and I quoted it was,
An evil AI is bringing nothing new to the table, and I’m not repeating myself.
I’m not surprised your ex is your EX…
On top of this, the courts aren’t providing certain guidance on how the geofence warrants should be executed, or what objective criteria would allow for progressively invasive searches.
I’m probably the only person here who read the transcript thoroughly and chased up the case law. There is barely anything in the transcript I didn’t know already. The UK has its equivalents and they’re all fairly rudimentary.
For matters of strategic policy development and assessment of legal changes the Law Commission is a thing.
Statutory guidance is a thing.
When doing a deep dive into a USA court opinion, you might want to study up how Marcy Wheeler does her work over on EmptyWheel (.net).
I personally have zero interest in the structures or rabbit warrens of the US legal system. That’s a critique all of its own. I’m personally interested more in legal principles and their underpinnings which are universally portable. The US could learn something from other jurisdictions instead of sticking its exceptionalist head in the sand.
Actually, I think the problem with the US is too much of the wrong kind of democracy. There’s too many votes for too many people right down to appointing the local dog catcher. Then you have a judiciary where judges dealing with cases which raise serious points of law need nothing more than a degree to be appointed and they can lack any formal understanding or experience of law. As for politicians and their arm waving about the “deep state” that’s usually right wing blowhards who find that their wild ideas and attitudes crash with reality then begin abusing their public office and relationship with the media to absolutely mess everything up. And for people who hate the “deep state” so much they sure like to stuff it with their friends and weaponise it.
I don’t follow the Tr*mp saga much at all lately. Why? The US legal system basically allows any Johnny come lately to mouth off in the media plus all the other political horsetrading and nonsense. Most European systems are much more sane.
Ted • January 15, 2023 9:29 AM
@JonKnowsNothing
Courts do not create evidence (normally). They do not do primary or secondary research. That is the job of the prosecution (unlimited funding) and sometimes the job of the defense (very limited funding).
Ah, okay, thank you. Upon looking again, I see that Chatrie (of US v. Chatrie) had a fair amount of wherewithal in his corner.
His lawyers included a public defender PLUS Michael Price, the lead litigator of the National Association of Criminal Defense Lawyers’ Fourth Amendment Center.
I will keep a lookout for Marcy Wheeler’s work. Thank you for the heads up 🙂
modem phonemes • January 15, 2023 7:01 PM
The website for Capricon is titled Capricon 43 and the theme is
perhaps, just perhaps, might The End actually be… just another
Beginning?
Is the 43 then what comes after 42 , the answer in Life, the Universe, and Everything ?
Subscribe to comments on this entry
Leave a comment
Sidebar photo of Bruce Schneier by Joe MacInnis.
modem phonemes • January 14, 2023 12:22 PM
The tl;dr
https://cyborgink.com/2022/10/30/answer-fredric-brown/