Gangster AI

Palantir is here to disrupt and make the institutions we partner with the very best in the world and, when it’s necessary, to scare our enemies and, on occasion, kill them,” Mr. Karp told shareholders earlier this month. “We hope you’re in favor of that, we hope you’re enjoying being a partner.”

https://www.washingtontimes.com/news/2025/feb/18/palantir-goes-script-reveal-ambitious-agenda-real-reason-tiktok-ban/

“Spawn of the Devil”

Alex Karp is a twisted human being. This quote says pretty much all you need to estimate what kind of human being he is – no morality worth speaking of (and when you see pictures of him he can barely bring himself to have a shower – he’s not only twisted, he’s grubby!).

Speaking at a shareholder’s meeting he is quoted as saying:

Palantir is here to disrupt and make the institutions we partner with the very best in the world and, when it’s necessary, to scare our enemies and, on occasion, kill them,” Mr. Karp told shareholders earlier this month. “We hope you’re in favor of that, we hope you’re enjoying being a partner.”

What a creep! (He has the blood of many Palestinian children on his hands …)

His mate Peter Thiel is no better, just lives in the shadows.

https://www.washingtontimes.com/news/2025/feb/18/palantir-goes-script-reveal-ambitious-agenda-real-reason-tiktok-ban/

Too much information

Until fairly recently, say 19th Century, the survival of ‘content’ from the past has been as much as matter of chance as of design. The invention of printing, of course, arriving in the 1450s, made a huge difference to the survival of documents of all kinds including from the previous ages of scribal production. But, as is evident from Roland Allen’s research in “The Notebook” the development and spread of paper as a medium not only made printing feasible but also the general use of paper as a means of recording all kinds of information. Yet by its nature paper is also a vulnerable medium and even if printing ensured the survival of a great deal much has been lost. For the everyday use of paper as a personal as well as an official medium of record, survivals have been rare when considered against the huge scale of creation over the centuries since the 1400s.

Today, digital methods of production and storage of text, images and sound, have the result that very much less is lost.* In fact, we may be entering an age when we experience a new kind of cultural problem, the overabundance of ‘content’. 

Is this a new kind of problem?**

============

* I use the word ‘less’ deliberately for much of the world’s population remains unaffected by the presence of digital media. and probably cares even less! Yet, it has to be true that over the next few decades the digital archive provides a deeper pool of content that represents in greater variety and detail the ‘output’ of human beings.

===========

** In fact, reading Yeo, Notebooks, English Virtuosi, and Early Modern Science, and Allen (see below) this is not really a new problem. During the period leading up to the of the creation of the Royal Society and then beyond into the 17-18 Centuries we find much anxiety about how to manage the flood of content and information that printing made available

===========

Posted to Bluesky

(1)

Takeaways: paper enabled print but also the spread of paper notebooks. Most now lost (tens of thousands?). Many books too but notebook loss is greater.

Contrast with the scale (~100%?) and longevity (200+ yrs?) of digital storage. This near-perfect archive may be our cultural nemesis (e.g. LLMs!).

(2)

For those interested in the history/evolution of pre-digital media see ‘The Notebook: A History of Thinking on Paper‘. Some useful insights into our media precursors.

===============

Black Mirror, Series 2, Episode 3.

Broadcast i Black Mirror, Series 2, Episode 3. Available on Amazon Prime.

Broadcast in 2013 this film has a prescient quality about it. An example of how an LLM, which in 2024 are now commonplace. It also posits the idea that well, they could be used to mimic the personality of an individual. This episode of Black Mirror illustrates the limitations of that idea. Something like an LLM cannot replicate a human life; and creating an automata built to to mimic the physical person in every physical detail (adjustable in real-time!) may just make their limitations worse.

A young couple, Martha and Ash, off to a new home, somewhere rural. He is always engrossed in his online chatter. She, a graphic designer, is good humoured about it, if at times a little impatient with his distraction.

After they unload their stuff he goes off the return th hire van. He does not make it. He dies in a crash. She is, of course, traumatised; grieving.

At the wake. A friend, also bereaved, tells her she can sign up to ‘something’ (an online system, just software, that uses all the information it has about someone, that you can have conversations with). She says it will help her grief, as it did for her. Martha is upset, denounces the whole idea, screaming – “It’s sick …”.

Later, while also discovering that she she is pregnant, she gets a message from the same friend ‘I’ve signed you up’ and in the same list of new messages there is also one from Ash!

She finds the conversations both weird but also comforting and she spends a lot of time talking ot online-Ash. It sounds like Ash, and Ash seems to know he is ‘not there’, and online-Ash does seem to have ‘gaps’ i his knowledge about himself, and about Martha – but these are quickly repaired after he performs some sort of lookup!

Later, her friend writes to her to tell her about a new beta project by the same company, This turns out to be an automaton that looks like Ash, sounds like Ash, but is also ‘gappy’ about things he should know especially physical behaviour and real-time interactions.

The sex, however, is terrific! But he does not sleep, which Martha finds disturbing and is oddly dispassionate about everything. He cannot leave the house, for he is programmed to stay within a fixed distance of his point of origin (the bathroom in this case where he was activated after delivery!) …

Eventually she becomes so disturbed, angered, by the entire situation that she walks him to the nearby cliff, overlooking the sea shore where they live, and tells him to jump. He says “OK” but, again, Martha is so disturbed by his dispassionate compliance she screams at him that she cannot take it any more. He then enacts, mimics, how a person might passionately and fearfully to such a demand, to kill himself …

Final scene some years later: Martha has a daughter and it’s her birthday. They come home and the daughter asks to take a piece of birthday take to Ash, who now ‘lives’ in the loft. We see him standing there immobile, with some basic furniture – a bed!, a lamp! – none of which he needs. He speaks kindly and softly to his daughter, “Happy Birthday”; she replies, handing him the cake, “I know you can’t really eat it but let’s pretend …”.

Last shot: Martha climbs the loft ladder to join them.

===============

Posted to BlueSky:

A ‘case study’ of the limitations of LLM-type models of mind mimicry: Black Mirror, Series 2, Episode 3 (on Amazon Prime).

Produced 10yr ago) in 2013 it is an insightful depiction of the absurdity of believing that mind-mimicry built from data archives is a ‘real’ mind.

It is also a warning.n 2013 this film has a prescient quality about it. An example of how an LLM, which in 2024 are now commonplace. It also posits the idea that well, they could be used to mimic the personality of an individual. This episode of Black Mirror illustrates the limitations of that idea. Something like an LLM cannot replicate a human life; and creating an automata built to to mimic the physical person in every physical detail (adjustable in real-time!) may just make their limitations worse.

A young couple, Martha and Ash, off to a new home, somewhere rural. He is always engrossed in his online chatter. She, a graphic designer, is good humoured about it, if at times a little impatient with his distraction.

After they unload their stuff he goes off the return th hire van. He does not make it. He dies in a crash. She is, of course, traumatised; grieving.

At the wake. A friend, also bereaved, tells her she can sign up to ‘something’ (an online system, just software, that uses all the information it has about someone, that you can have conversations with). She says it will help her grief, as it did for her. Martha is upset, denounces the whole idea, screaming – “It’s sick …”.

Later, while also discovering that she she is pregnant, she gets a message from the same friend ‘I’ve signed you up’ and in the same list of new messages there is also one from Ash!

She finds the conversations both weird but also comforting and she spends a lot of time talking ot online-Ash. It sounds like Ash, and Ash seems to know he is ‘not there’, and online-Ash does seem to have ‘gaps’ i his knowledge about himself, and about Martha – but these are quickly repaired after he performs some sort of lookup!

Later, her friend writes to her to tell her about a new beta project by the same company, This turns out to be an automaton that looks like Ash, sounds like Ash, but is also ‘gappy’ about things he should know especially physical behaviour and real-time interactions.

The sex, however, is terrific! But he does not sleep, which Martha finds disturbing and is oddly dispassionate about everything. He cannot leave the house, for he is programmed to stay within a fixed distance of his point of origin (the bathroom in this case where he was activated after delivery!) …

Eventually she becomes so disturbed, angered, by the entire situation that she walks him to the nearby cliff, overlooking the sea shore where they live, and tells him to jump. He says “OK” but, again, Martha is so disturbed by his dispassionate compliance she screams at him that she cannot take it any more. He then enacts, mimics, how a person might passionately and fearfully to such a demand, to kill himself …

Final scene some years later: Martha has a daughter and it’s her birthday. They come home and the daughter asks to take a piece of birthday take to Ash, who now ‘lives’ in the loft. We see him standing there immobile, with some basic furniture – a bed!, a lamp! – none of which he needs. He speaks kindly and softly to his daughter, “Happy Birthday”; she replies, handing him the cake, “I know you can’t really eat it but let’s pretend …”.

Last shot: Martha climbs the loft ladder to join them.

===============

Posted to BlueSky:

A ‘case study’ of the limitations of LLM-type models of mind mimicry: Black Mirror, Series 2, Episode 3 (on Amazon Prime).

Produced 10yr ago in 2013 it is an insightful depiction, and a warning about the absurdity of believing that mind-mimicry built from data archives is a ‘real’ mind.

It is also a warning.

Creative Destruction

Our beginning and our end …

Photo of the first image of a black hole with a dark centre surrounded by a glowing, uneven ring of orange, yellow and red.

The new image of the black hole in the Messier 87 galaxy. L. Medeiros (Institute for Advanced Study), D. Psaltis (Georgia Tech), T. Lauer (NSF’s NOIRLab), and F. Ozel (Georgia Tech).
Source

What the image shows is the ‘photon ring’ that surrounds the gravitational mass of the  black hole. The photon ring is a boundary layer, the last gasp of light as it is dragged into the black hole (yes, that’s the void in the middle of the ring that we cannot actually see).

Interesting detail: the photon ring is brighter towards the bottom of the image because the ring is spinning mind-bogglingly fast and where the ring is brighter is the result of the Doppler effect – the brighter part of the ring is spinning towards us (…well, that’s the  hypothesis anyway!)

Even more beautiful is the black hole at the centre of our own galaxy: Sagittarius A*

Futures (from Kurt Vonnegut)

 1988: A vision of the future from Kurt Vonnegut:

Interesting to note that way back in 1909 this was pretty much the same vision as laid out by E.M.Forster in ‘The Machine Stops’. Here are Vonnegut’s headline bullet points:

1. Reduce and stabilize your population.

2. Stop poisoning the air, the water, and the topsoil.

3. Stop preparing for war and start dealing with your real problems.

4. Teach your kids, and yourselves, too, while you’re at it, how to inhabit a small planet without helping to kill it.

5. Stop thinking science can fix anything if you give it a trillion dollars.

6. Stop thinking your grandchildren will be OK no matter how wasteful or destructive you may be, since they can go to a nice new planet on a spaceship. That is really mean, and stupid.

7. And so on. Or else.

See also: Benedict Cumberbatch reads Kurt Vonnegut

============================

Futurism (Marinetti) was nuts, in a way we know so well!

Big Tech’s Partnership with Authoritarianism

Ereader reviews never tell you what it’s like to actually read with them …

I have always been perplexed, and annoyed, by reviews of ereaders that focus almost entirely on hardware issues – how fast, how sharp, how bright, how much memory, is it made of metal or plastic etc. Important details but, really, the issue is does it make reading a pleasant and satisfying experience? The technical stuff takes one or two sentences.

Reviewers hardly ever tell you what it is like to actually read books with them. They might say it is comfortable to hold in the hand but what is the highlighting system like, does it work easily when actually reading, can you retrieve and download your highlights (how do you do that, and in what format), is searching for words and phrases in the text managed well, does it provide easy to dictionaries and other internet resources, are page layouts and formats easily adjusted, does the ereader tie you in to a particular cloud service (and if so how easy is that to use?) etc. etc. 

No it;s always minor details about processors ad memory – important enough bu not the main reason for buying an ereader.

The situation has not improved now that we moved in to the era of ereader/notebook hybrids. In fact with notebooks the issue is even more important. Reviewers forget that there are at least two kinds of notebook users – journallers and workbooks. The former are well suited to eink notebooks for, by and large, they are ore likely to keep a notebook as a running, routine commentary on a topic. The man feature of journalling is that it is a continuous record, developed over time. A prose diary is the most obvious case.

For me at least, a workbook is much less coherent and usually involves a ore fragmentary approach. Entries are not necessarily sequential though they might be but never spread across an entire notebook. To coin a word, workbooks are ‘heterotopic’.   

So when I read reviews of eink notebooks I want to know what purposes they suit best. Like many people I have on three occasions decided, or been persuaded, to try them out because they seem like the next best thing (I’ve own a Boox, a reMarkable, and lately a Scribe). I do not currently have one because they do not suit my ‘workbook’ style – disparate notes – yes I’ll call them ‘jottings’  – that, when written into the ‘white box’ of an eink screen are not easily ‘retrieved’. 

But reviewers never tell you about these issues – how well does an eink notebook enable search and how easily do they all the writer to extract and organise retrievals into useful lists – much like ‘notes’ from a Kindle – No reviewer tells you about how tools for find, retrieve and organise scribbles from disparate sections of disparate workbooks. And I have yet to to find an eink notebook that does this.

This example is bit extreme, but it sums up how irrelevant so-called reiews can be. This one aims to be general – to cover 5 reasons why eink notebooks are great – but it never talks about note taking! 

https://www.xda-developers.com/e-ink-tablets-are-the-best-notetaking-devices

5: No more inky fingers (Ink! OK maybe but now excuse and can’t even use a pen correctly!)

4: Instant backups (doh! standard these days across almost all software’/hardware and especially handhelds

3: No more wasting paper (dubious – wasting? not if you are maintaining your paper notes; besides this is small beer compared with eh production footprint of devices)

2: I can do other tasks too (well, this isn’t the point is it?)

1: Handwriting recognition (agan, fair enough, but come on – if you can’t read your own handwriting what’s the point?)