Greetings from the Isle of Wight, the land of my childhood, where I am visiting family this week!
It may be unnecessary to note that Blair's peddling of AI corresponds with venture capital's investment cycle aligning with AI. This will end, just as it did for VR, when it becomes apparent that the story being sold about the new technology is shown to be smoke and mirrors by the passage of time and the failure for the incredible new marvel prognosticated to appear.
As someone whose Masters degree was in AI, I can say with confidence that the new AI systems are much less impressive than the sales pitch, not reliable enough to be used in almost any context without hilarious side effects, and that the most plausible application for the large language model (which most new AI sales pitches are focussed upon) is far easier censorship operations online. Weighting online searches for censorship (Google's game) becomes far easier when you can just provide a prompt to an automated system to skew search results that have been converted into natural language without any human involvement. This, the robots can handle. More than this... not so much.
All of which is a reminder of why our dreams of escaping incompetent governance don't go very far. Those caught up in high-level political machinations are simultaneously in the game of presenting themselves as valuable to the electorate (which is your recurring theme, David) and ensuring that they are deemed as valuable to those involved in circulating capital (which is where they get paid when they cease to be politicians). Once again, I am reminded of the great potential for a 'vow of poverty' to be mandatorily taken by politicians as a wondrous - and hilariously unlikely - way of resolving the problem of career politicians. 😂
Stay wonderful,
Chris.
PS: There's a stub marked [LINK] where I suspect you intended to refer back to an earlier piece, unless this was a subtle joke, which seems the less likely interpretation.
When Blair 'peddles', the instinct is to cycle far away. He makes it seem like AI can 'refresh' the parts ( other) government's cannot reach, but AI is the product of its human trainers, ergo, prone to disseminating misinformation, at risk of 'corruption' by bugs in the system or external hijacking of its data bank. Still AI has more chance of not giving skewed and biased answers than the lamentable dissembler Blair.
One of the limitations of AI, I expect, is that it is trained on existing data. While it may discover hidden relationships I doubt that it can 'create' new political visions. Which is going to disappoint many people hoping for a glorious Utopia.
My view is that AI will limited by liability issues. Once upon a time we could hold people or organisations liable for harm cause by their actions or decisions. But with AI, where does the liability sit ? The authors of the software, the people who 'trained' it, the official who signed off and accepted the output of the AI system as the 'correct' way forward ? It seems almost as if it's another way of diffusing and therefore avoiding repsonsibility by bureaucrats who are beginning to realse that they can't cope. I spent over twenty years in IT contracted to government departments, and it seemed to me that the primary objective of the officials was to avoid being held accountable for anything in case it went badly. To my mind AI will just make that worse.
You have provided evidence for what I believe. AI will take over existing internet information search and give us the answer the politicians want us to see.
One of the biggest problems for politicians 'fixing things' is that there is an unspoken taboo that changes must not adversely affect people. Privileges or rights must be grandfathered in, there must be Government relief for people who who are 'caught out' by taxation or benefit changes.
And yet this is the dynamic of any change. There will be winners and losers - and as long as you cannot discuss or allow 'losers' the options for change are severely limited. But there comes a point (whether you accept the idea of the Laffer curve or not) where no further effective change is possible. And this, to return to the article, is where Governments tread water 'managing problems' and their only argument is that they tread water in a more stylish way than their opponents.
Pournelle's Iron Law of Bureaucracy and Parkinson's Law apply all the time unless there's a significant external shock like losing a war or national bankruptcy. It is possible Millei in Argentina is going to be able to slash Argentine bureaucracy but I'm not yet convinced he will succeed.
The lack of competence you observe is downstream of this. Too many bureaucrats meddling in education policy and insisting on DIE sorts of indoctrination instead of quality education. About the one thing the Tories k,ind of did right was to liberalize, a little (probably too little), education and to allow for different sorts of schools. Just possibly this will turn out to be enough assuming the graduates of these schools are not embraced and suffocated by the rest of the system
Your remarks about bureaucracies metastastising reminded me of a letter to Nature in 1980 , when it was a decent journal, by Robert Moss, titled 'Expanding on Parkinson's Law'. In it he presents a simple model based on routes of communication to explain how adminstrative 'work' never shrinks, but will expand unless tightly controlled. It can be found on the internet here:
I found this article in the mid 1980s when I had left the offshore oil industry and was discovering that not all organisations were as lean-and-mean as my previous employer.
Words from politicians like "deliver" and "roll-out" tend to bring me out in a cold sweat and induce vomiting. Not sure how AI is going to "prevent" this health issue...
This was such a great read, it's so heartening to know there's others out there as interested in this as I am.
Your post reminded me of a concept I came across somewhere (maybe in Thinking In Systems book) called Bureaucratic Drag.
It refers to the inefficiencies and delays caused by excessive bureaucracy within organisations, particularly in governmental and large corporate environments.
I like to imagine this becomes a problem - as they appear to only be able to add to the system and never take anything away.
Thanks very much. Yes, bureaucratic drag is definitely 'a Thing'. Working at a university and seeing how much the tail wags the dog in terms of the relationship between the bureaucracy and academic staff has made me fascinated by this issue.
Brilliant. Incidentally, the problem you describe here -- learned helplessness in the population and metastasizing state bureaucracies -- is probably worse here in Canada than anywhere else.
Naturally, the members of the new government and their MPs are flushed with pride at their “great” victory and thrilled and full of enthusiasm to act swiftly to implement their wonderful vision for a future socialist Britain. Perhaps as they go about their great schemes they hum to themselves little ditties such as:
The “State” is the true embodiment
Of everything that's excellent.
It has no kind of fault or flaw,
And we, my dears, embody it all.
(apologies to W.S Gilbert)
I am an old man. I have witnessed many politicians come into office, driven by ideology, full of pride and ambition and after a few years slink away in failure and shame.
I forecast that the projects proposed by this new government, if it seeks to lord it over us, will have six phases:
1. Enthusiasm
2. Disillusionment
3. Panic
4. Search for the guilty
5. Punishment of the innocent
6. Praise and honours for the non-participants
As for AI, I’m reading, or rather listening to, “Technology is Not the Problem” by Timandra Harkness. She tells us that the way we use technology is the real problem because governments, business and advertisers, streaming services etc are all using technology to manipulate and influence our actions and choices and thought processes. So we should be aware of this.
In the early years of computing, when a computer filled a whole room and punchcards and punchtape were the medium for feeding data into it, the dictum we often used was "rubbish in-rubbish out". I guess the same applies to AI. Although I am amazed that my laptop and Tablet are more powerful than the huge computers I worked with 50+ years ago.
My own worldview of how society should function is guided by the moral theology of Catholic Social Teaching.
Not at odds at all. It's fun making lists! I claim no originality for the list I suggested. It was David's mention of AI that caused me to dredge it up from my memory. As I recall, the list was created as a joke about, or mockery of, the design process of computer systems that were heralded, with great fanfare, as likely to be the best ever for increasing productivity enormously, taking customer service to never before seen levels of perfection and cutting costs to the bare minimum; only find after a few years that the opposite was true and the system was a complete disaster.
It seems to me that all of the UK governments in recent years have followed such a model or the model you suggest.
And the irony of the whole thing is ... Tony Blair's report on the future of AI -- WAS WRITTEN BY AI. What data did he use? Whatever would get him the answer he had already decided on. How self-referencing can you get?
Greetings from the Isle of Wight, the land of my childhood, where I am visiting family this week!
It may be unnecessary to note that Blair's peddling of AI corresponds with venture capital's investment cycle aligning with AI. This will end, just as it did for VR, when it becomes apparent that the story being sold about the new technology is shown to be smoke and mirrors by the passage of time and the failure for the incredible new marvel prognosticated to appear.
As someone whose Masters degree was in AI, I can say with confidence that the new AI systems are much less impressive than the sales pitch, not reliable enough to be used in almost any context without hilarious side effects, and that the most plausible application for the large language model (which most new AI sales pitches are focussed upon) is far easier censorship operations online. Weighting online searches for censorship (Google's game) becomes far easier when you can just provide a prompt to an automated system to skew search results that have been converted into natural language without any human involvement. This, the robots can handle. More than this... not so much.
All of which is a reminder of why our dreams of escaping incompetent governance don't go very far. Those caught up in high-level political machinations are simultaneously in the game of presenting themselves as valuable to the electorate (which is your recurring theme, David) and ensuring that they are deemed as valuable to those involved in circulating capital (which is where they get paid when they cease to be politicians). Once again, I am reminded of the great potential for a 'vow of poverty' to be mandatorily taken by politicians as a wondrous - and hilariously unlikely - way of resolving the problem of career politicians. 😂
Stay wonderful,
Chris.
PS: There's a stub marked [LINK] where I suspect you intended to refer back to an earlier piece, unless this was a subtle joke, which seems the less likely interpretation.
Thanks for the heads up - oops! Anyway, yes, on the AI point I am very dubious. It sounds very much like clutching at straws to me.
It's more believing how own bs and/or not grasping the hype of silicon valley is the same hype he was guilty of in other circumstances
When Blair 'peddles', the instinct is to cycle far away. He makes it seem like AI can 'refresh' the parts ( other) government's cannot reach, but AI is the product of its human trainers, ergo, prone to disseminating misinformation, at risk of 'corruption' by bugs in the system or external hijacking of its data bank. Still AI has more chance of not giving skewed and biased answers than the lamentable dissembler Blair.
One of the limitations of AI, I expect, is that it is trained on existing data. While it may discover hidden relationships I doubt that it can 'create' new political visions. Which is going to disappoint many people hoping for a glorious Utopia.
Yes - it just creates pastiche. And pastiche of good government doesn’t cut it.
Governments **
My view is that AI will limited by liability issues. Once upon a time we could hold people or organisations liable for harm cause by their actions or decisions. But with AI, where does the liability sit ? The authors of the software, the people who 'trained' it, the official who signed off and accepted the output of the AI system as the 'correct' way forward ? It seems almost as if it's another way of diffusing and therefore avoiding repsonsibility by bureaucrats who are beginning to realse that they can't cope. I spent over twenty years in IT contracted to government departments, and it seemed to me that the primary objective of the officials was to avoid being held accountable for anything in case it went badly. To my mind AI will just make that worse.
You have provided evidence for what I believe. AI will take over existing internet information search and give us the answer the politicians want us to see.
One of the biggest problems for politicians 'fixing things' is that there is an unspoken taboo that changes must not adversely affect people. Privileges or rights must be grandfathered in, there must be Government relief for people who who are 'caught out' by taxation or benefit changes.
And yet this is the dynamic of any change. There will be winners and losers - and as long as you cannot discuss or allow 'losers' the options for change are severely limited. But there comes a point (whether you accept the idea of the Laffer curve or not) where no further effective change is possible. And this, to return to the article, is where Governments tread water 'managing problems' and their only argument is that they tread water in a more stylish way than their opponents.
We know the state always gets bigger.
Pournelle's Iron Law of Bureaucracy and Parkinson's Law apply all the time unless there's a significant external shock like losing a war or national bankruptcy. It is possible Millei in Argentina is going to be able to slash Argentine bureaucracy but I'm not yet convinced he will succeed.
The lack of competence you observe is downstream of this. Too many bureaucrats meddling in education policy and insisting on DIE sorts of indoctrination instead of quality education. About the one thing the Tories k,ind of did right was to liberalize, a little (probably too little), education and to allow for different sorts of schools. Just possibly this will turn out to be enough assuming the graduates of these schools are not embraced and suffocated by the rest of the system
I basically agree - I just hope it doesn’t take too much of an external shock…..
That reminds me of a saying of one of my tutors in the 70s:
"This country needs a revolution, it's just that I don't want to have to live through one"
Your remarks about bureaucracies metastastising reminded me of a letter to Nature in 1980 , when it was a decent journal, by Robert Moss, titled 'Expanding on Parkinson's Law'. In it he presents a simple model based on routes of communication to explain how adminstrative 'work' never shrinks, but will expand unless tightly controlled. It can be found on the internet here:
https://www.nature.com/articles/285009a0
I found this article in the mid 1980s when I had left the offshore oil industry and was discovering that not all organisations were as lean-and-mean as my previous employer.
That's so interesting - thanks for that.
Words from politicians like "deliver" and "roll-out" tend to bring me out in a cold sweat and induce vomiting. Not sure how AI is going to "prevent" this health issue...
Spot on Dave.
This was such a great read, it's so heartening to know there's others out there as interested in this as I am.
Your post reminded me of a concept I came across somewhere (maybe in Thinking In Systems book) called Bureaucratic Drag.
It refers to the inefficiencies and delays caused by excessive bureaucracy within organisations, particularly in governmental and large corporate environments.
I like to imagine this becomes a problem - as they appear to only be able to add to the system and never take anything away.
Thanks very much. Yes, bureaucratic drag is definitely 'a Thing'. Working at a university and seeing how much the tail wags the dog in terms of the relationship between the bureaucracy and academic staff has made me fascinated by this issue.
Brilliant. Incidentally, the problem you describe here -- learned helplessness in the population and metastasizing state bureaucracies -- is probably worse here in Canada than anywhere else.
We could have a really interesting debate about which country has it worse!
An excellent analysis, David, thank you.
Naturally, the members of the new government and their MPs are flushed with pride at their “great” victory and thrilled and full of enthusiasm to act swiftly to implement their wonderful vision for a future socialist Britain. Perhaps as they go about their great schemes they hum to themselves little ditties such as:
The “State” is the true embodiment
Of everything that's excellent.
It has no kind of fault or flaw,
And we, my dears, embody it all.
(apologies to W.S Gilbert)
I am an old man. I have witnessed many politicians come into office, driven by ideology, full of pride and ambition and after a few years slink away in failure and shame.
I forecast that the projects proposed by this new government, if it seeks to lord it over us, will have six phases:
1. Enthusiasm
2. Disillusionment
3. Panic
4. Search for the guilty
5. Punishment of the innocent
6. Praise and honours for the non-participants
As for AI, I’m reading, or rather listening to, “Technology is Not the Problem” by Timandra Harkness. She tells us that the way we use technology is the real problem because governments, business and advertisers, streaming services etc are all using technology to manipulate and influence our actions and choices and thought processes. So we should be aware of this.
In the early years of computing, when a computer filled a whole room and punchcards and punchtape were the medium for feeding data into it, the dictum we often used was "rubbish in-rubbish out". I guess the same applies to AI. Although I am amazed that my laptop and Tablet are more powerful than the huge computers I worked with 50+ years ago.
My own worldview of how society should function is guided by the moral theology of Catholic Social Teaching.
My alternative set of phases, not especially at odds with yours...
1. Enthusiasm
2. Realisation
3. Panic
4. Deflection from the guilty
5. Lessons that must be learned (and won't be)
6. Praise and honours for the great and good (especially for those coming to the 'correct conclusions')
Not at odds at all. It's fun making lists! I claim no originality for the list I suggested. It was David's mention of AI that caused me to dredge it up from my memory. As I recall, the list was created as a joke about, or mockery of, the design process of computer systems that were heralded, with great fanfare, as likely to be the best ever for increasing productivity enormously, taking customer service to never before seen levels of perfection and cutting costs to the bare minimum; only find after a few years that the opposite was true and the system was a complete disaster.
It seems to me that all of the UK governments in recent years have followed such a model or the model you suggest.
I’ll be meeting Harkness next week. I’ve not read the book but have heard her being interviewed about it.
Her narration of the book on Audible is delightful.
And the irony of the whole thing is ... Tony Blair's report on the future of AI -- WAS WRITTEN BY AI. What data did he use? Whatever would get him the answer he had already decided on. How self-referencing can you get?
Thank you for the link. I am interested in AI and religion.
Franciscan Ilia Delio, a scientist-theologian, explores the connection between AI and
Faith. A review of one of her recent books may be found here
https://aiandfaith.org/ilia-delios-re-enchanting-the-earth-why-ai-needs-religion-review/
I discovered the aiandfaith website only very recently. I haven’t explored it much so far
but it looks most interesting.
Dr (Sr) Delio is the founder of the Center for Christogenesis. Here is a link to a video
where she talks about why AI needs religion.
https://christogenesis.org/why-ai-needs-religion/