Technical Thread

[c-a-r-f-r-e-w]

May 7, 2023 14:33:40 GMT alurqa said:

I downloaded www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationprojections/datasets/localauthoritiesinenglandtable2

which is an Excel spreadsheet. I exported the data as a CSV file, but using TABs to separate fields. I then wrote this awk program (see attachment) to sum up different age groups from this forcast data.
View Attachment

…

What is striking is over the next 20 years with the current assumptions the number of:

* under 25s will increase from 16.8m to 16.9m;

* 25 - 64s will increase from 29.0m to 30.1m;

* the 65 and overs will increase from 10.1m to 14.7m.

So there is an increase in all three age groups, BUT, in percentages of the total population:

* under 25s will decrease from 29.9% to 27.3% (a drop of 2.6%);

* 25 - 64s will decrease from 51.8% to 48.7% (a drop of 3.1%);

* the 65 and overs will increase from 18.1% to 23.8% (a rise of 5.7%).

So in 2043 3.1% fewer working people will have to pay 5.7% more pensioners. Hmm.

Useful analysis Al. Some estimates have immigration currently running at nearly a million, and most of those are liable to be working people. On the other hand, it looks like it’s -possible fewer people will be becoming parents because cost-of-living keeps going up etc.

[c-a-r-f-r-e-w]

May 12, 2023 21:29:22 GMT mercian said:

May 7, 2023 14:33:40 GMT alurqa said:

I downloaded www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationprojections/datasets/localauthoritiesinenglandtable2

which is an Excel spreadsheet. I exported the data as a CSV file, but using TABs to separate fields. I then wrote this awk program (see attachment) to sum up different age groups from this forcast data.
View Attachment

This is the result of my running my program against the OPCS data above:

…

What is striking is over the next 20 years with the current assumptions the number of:

* under 25s will increase from 16.8m to 16.9m;

* 25 - 64s will increase from 29.0m to 30.1m;

* the 65 and overs will increase from 10.1m to 14.7m.

So there is an increase in all three age groups, BUT, in percentages of the total population:

* under 25s will decrease from 29.9% to 27.3% (a drop of 2.6%);

* 25 - 64s will decrease from 51.8% to 48.7% (a drop of 3.1%);

* the 65 and overs will increase from 18.1% to 23.8% (a rise of 5.7%).

So in 2043 3.1% fewer working people will have to pay 5.7% more pensioners. Hmm.

Interesting results. There are lots of variables which could affect the forecasts though. For instance government estimates of immigration tend to be low, especially illegal immigration for obvious reasons, and immigrants tend to be young.  The government is already talking about increasing the state pension age again. Also perhaps a citizen's wage will come in as there could be mass unemployment as a result of AI. I was watching a couple of chaps doing a really good job of clearing a neighbour's gutters and cleaning flat roof etc today and thought that that sort of job would be one of the last to go. Office drones doing repetitive tasks will be phased out pretty quickly I think. Apart from in the Civil Service and NHS obviously.

It may well get rid of the clerical jobs down the ladder - the question for me is how far up the ladder does it go? How many of the essay-writers running things are going to get replaced?

[mercian]

May 12, 2023 23:53:32 GMT c-a-r-f-r-e-w said:

May 12, 2023 21:29:22 GMT mercian said:

Interesting results. There are lots of variables which could affect the forecasts though. For instance government estimates of immigration tend to be low, especially illegal immigration for obvious reasons, and immigrants tend to be young.  The government is already talking about increasing the state pension age again. Also perhaps a citizen's wage will come in as there could be mass unemployment as a result of AI. I was watching a couple of chaps doing a really good job of clearing a neighbour's gutters and cleaning flat roof etc today and thought that that sort of job would be one of the last to go. Office drones doing repetitive tasks will be phased out pretty quickly I think. Apart from in the Civil Service and NHS obviously.

It may well get rid of the clerical jobs down the ladder - the question for me is how far up the ladder does it go? How many of the essay-writers running things are going to get replaced?

 It'll take longer, but many will go eventually.  For many years some senior execs refused to use computers because they were used to minions finding out information for them or typing their letters.  I doubt there are many of those left now, but those who have modernised are now doing quite a lot of what their predecessors considered to be menial work - e.g. typing their own emails instead of dictating a letter to a secretary.  Perhaps AI will deskill even high-level jobs eventually and senior execs will just get an AI to come up with some strategy and then basically rubber-stamp it.
 I think I said this recently somewhere else, but it does concern me a bit that a lot of traditional mental skills are disappearing - e.g. mental arithmetic, map-reading, memorising poetry and so on. Not necessarily a bad thing in itself, as we must keep moving on, but what are they being replaced with other than the ability to ask a machine a question?  

[leftieliberal]

A BBC Sounds programme that is well-worth listening to on the replication crisis in science: www.bbc.co.uk/sounds/play/m001lqvg

For years, science has had a dirty secret; research has been dogged by claims and instances of fraud, malpractice and outright incompetence. Suspicious-looking data sets, breakthrough results that can’t be replicated, eyebrow-raising statistical sleights of hand - science has been undergoing something of an existential crisis.

And at the forefront of keeping science honest has been a bunch of outsiders, some of them with no formal academic positions, no salaried posts, double-checking the published claims of researchers and academics. Their work is not without controversy, especially when they go public; nevertheless, they’ve achieved impressive results.

Presenter Michael Blastland meets some of these ‘Truth Police’, discovering their methodology and their motives, as well as asking how scientific institutions are reacting to the deep issues they have brought out into the light.

Presenter: Michael Blastland

[mercian]

I've been pondering the AI revolution. This is a trivial example but we all know that AI-generated pictures of people often have the wrong number of fingers. In the expected course of events one would expect a gradual improvement, but suppose some mischievous person decided to tell an AI to generate thousands or millions of such pictures and spread them around the internet. Then any future AI would find it difficult to decide what is the correct number of fingers. How could that ever be corrected?
That was a trivial example, but what if it was something more serious spread by a bad actor? Perhaps some political position such as the Holocaust never happened, or something more current (I'm trying to be uncontroversial). AI programs would just use whatever had been spread as input to any new work that it did? The Internet at the moment is already full of misleading rubbish but in time faulty AI-generated stuff could crowd everything else out including good stuff. How could it ever be corrected or controlled? Very worrying.

[c-a-r-f-r-e-w]

Ultrasound breakthrough could help humans venture into deep space
Scientists use pulses to send rats into hibernation and the technique could potentially be adapted for interstellar travel

“A future where humans are placed in hibernation to travel into space is a step closer after scientists proved they can put mammals into a sleep-state using ultrasound.

….

Now, a team at Washington University in St. Louis has shown they can send rodents into a state of reversible hibernation by firing ultrasound pulses at their heads, stimulating the hypothalamus preoptic area in the brain, and causing their heart rates to halve and their temperature to drop.

…

When the scientists dug into what was causing the effect, they found that the ultrasound activated an ion channel linked to metabolism and temperature.

Although mice can enter torpor naturally, the team showed the same process works on rats, an animal that does not hibernate.

The researchers said the rat experiments demonstrated that the brain regions regulating hibernation may also be present in non-hibernating mammals and gave hope that the technique could be used in humans.”

[c-a-r-f-r-e-w]

May 21, 2023 16:38:29 GMT mercian said:

I've been pondering the AI revolution. This is a trivial example but we all know that AI-generated pictures of people often have the wrong number of fingers. In the expected course of events one would expect a gradual improvement, but suppose some mischievous person decided to tell an AI to generate thousands or millions of such pictures and spread them around the internet. Then any future AI would find it difficult to decide what is the correct number of fingers. How could that ever be corrected?
That was a trivial example, but what if it was something more serious spread by a bad actor? Perhaps some political position such as the Holocaust never happened, or something more current (I'm trying to be uncontroversial). AI programs would just use whatever had been spread as input to any new work that it did? The Internet at the moment is already full of misleading rubbish but in time faulty AI-generated stuff could crowd everything else out including good stuff. How could it ever be corrected or controlled? Very worrying.

Gettimg access to high quality sources of info. for the AI is currently a limiting factor. Presumably they would weight the training toward higher-quality sources? (Maybe we will end up with AIs dedicated to policing things and removing the pics with too many digits etc…)


meanwhile things move apace…

“In 1987, during a forecast for the BBC, Fish said that a woman had called in to say that a hurricane was on the way. “Well, if you’re watching, don’t worry, there isn’t!” he blithely told viewers. Had Fish had the benefit of AI, perhaps he would not have been so bold. During the past few months a quiet revolution has been brewing in meteorology — driven by the technology of the moment.

Peter Dueben, head of earth system modelling at the European Centre for Medium-Range Weather Forecasts (ECMWF), described Google DeepMind’s efforts in particular as “astonishing”. He said: “They have really surprised the weather and climate community how good those predictions can be in principle. There’s a lot happening right now.””

This month Atmo, a San Francisco-based start-up, released what it claimed was the first live global medium-range weather forecast, purely based on an AI model, that was as good as the gold standard. The difference? It is cheaper, faster and learns from its mistakes — and could herald street-by-street forecasts in the coming years, according to experts.”

…

“The AI models are more agile, Alexander Levy, the co-founder of Atmo, said. He added: “Since this model can run in a few seconds, instead of hours, you get the information about the forecast much earlier. I think that is the main breakthrough here.

“The computers necessary to produce a fast forecast would before cost billions of dollars. Whereas this can be done more cheaply or much more quickly, or both, or in greater detail.”

Cheaper and faster means greater accuracy because you can create a bigger “ensemble” forecast, which is where you re-run the model with different permutations to get a better prediction. Instead of running it 20 times, which would be the limit of traditional systems because of cost, you could run it 10,000 times or more.”


Times

[c-a-r-f-r-e-w]

“More directly, an AI bent on a goal to which the existence of humans had become an obstacle, or even an inconvenience, could set out to kill all by itself. It sounds a bit Hollywood, until you realise that we live in a world where you can email a DNA string consisting of a series of letters to a lab that will produce proteins on demand: it would surely not pose too steep a challenge for “an AI initially confined to the internet to build artificial life forms”, as the AI pioneer Eliezer Yudkowsky puts it. A leader in the field for two decades, Yudkowksy is perhaps the severest of the Cassandras: “If somebody builds a too-powerful AI, under present conditions, I expect that every single member of the human species and all biological life on Earth dies shortly thereafter.””

www.theguardian.com/commentisfree/2023/may/26/future-ai-chilling-humans-threat-civilisation

[leftieliberal]

May 21, 2023 16:38:29 GMT mercian said:

I've been pondering the AI revolution. This is a trivial example but we all know that AI-generated pictures of people often have the wrong number of fingers. In the expected course of events one would expect a gradual improvement, but suppose some mischievous person decided to tell an AI to generate thousands or millions of such pictures and spread them around the internet. Then any future AI would find it difficult to decide what is the correct number of fingers. How could that ever be corrected?
That was a trivial example, but what if it was something more serious spread by a bad actor? Perhaps some political position such as the Holocaust never happened, or something more current (I'm trying to be uncontroversial). AI programs would just use whatever had been spread as input to any new work that it did? The Internet at the moment is already full of misleading rubbish but in time faulty AI-generated stuff could crowd everything else out including good stuff. How could it ever be corrected or controlled? Very worrying.

An interesting article on why the threat from AI is not extermination, but something we are already seeing:
www.theguardian.com/commentisfree/2023/may/31/yes-you-should-be-worried-about-ai-but-matrix-analogies-hide-a-more-insidious-threat
The big AI/tech companies are waving the existential threat at us in the hope of distracting us from what they are already doing.

Rather than fretting over some far-flung fear of an “existential threat” to humanity, we should be concerned about the material consequences of far less sophisticated AI technologies that are affecting people’s lives right now. And what’s more, we should be deeply troubled by the way AI is being leveraged to further concentrate power in a handful of companies.

It's also worth reading what Scott Aaronson (a quantum-computing specialist, who also advises OpenAI) has written on the subject in his blog

scottaaronson.blog/?cat=8 (this category contains his posts tagged "The Fate of Humanity")

[c-a-r-f-r-e-w]

May 30, 2023 17:06:25 GMT leftieliberal said:

May 21, 2023 16:38:29 GMT mercian said:

I've been pondering the AI revolution. This is a trivial example but we all know that AI-generated pictures of people often have the wrong number of fingers. In the expected course of events one would expect a gradual improvement, but suppose some mischievous person decided to tell an AI to generate thousands or millions of such pictures and spread them around the internet. Then any future AI would find it difficult to decide what is the correct number of fingers. How could that ever be corrected?
That was a trivial example, but what if it was something more serious spread by a bad actor? Perhaps some political position such as the Holocaust never happened, or something more current (I'm trying to be uncontroversial). AI programs would just use whatever had been spread as input to any new work that it did? The Internet at the moment is already full of misleading rubbish but in time faulty AI-generated stuff could crowd everything else out including good stuff. How could it ever be corrected or controlled? Very worrying.

An interesting article on why the threat from AI is not extermination, but something we are already seeing:
www.theguardian.com/commentisfree/2023/may/31/yes-you-should-be-worried-about-ai-but-matrix-analogies-hide-a-more-insidious-threat
The big AI/tech companies are waving the existential threat at us in the hope of distracting us from what they are already doing.

Rather than fretting over some far-flung fear of an “existential threat” to humanity, we should be concerned about the material consequences of far less sophisticated AI technologies that are affecting people’s lives right now. And what’s more, we should be deeply troubled by the way AI is being leveraged to further concentrate power in a handful of companies.

It's also worth reading what Scott Aaronson (a quantum-computing specialist, who also advises OpenAI) has written on the subject in his blog

scottaaronson.blog/?cat=8 (this category contains his posts tagged "The Fate of Humanity")

Yes, good spot. Fi7nd this particularly salient: “The through-line here is that we’re not talking about the danger of some far-off sci-fi future, we’re talking about the amplification of systems and social problems that already exist.”

[leftieliberal]

Something to be aware of if you work for an organisation that uses these cameras: www.bbc.co.uk/news/technology-65975446

But experts say the UK needs to do more to protect itself from what Prof Sampson, the surveillance camera commissioner, describes as "digital asbestos".

"We have a previous generation that has installed this equipment, largely on the basis that it was cheap and got the job done," he says. "We've now realised that it has some serious and inherent risks - so what do we about it?"

Asked whether he trusts Hikvision and Dahua, he replies: "Not one bit."

[leftieliberal]

Sarah Silverman sues OpenAI and Meta for misuse of copyrighted material.

www.bbc.co.uk/news/technology-66164228

Apart from the exemption for "fair dealing and research" there are strong restrictions around the use of copyrighted material. It seems that these Large Language Models were trained using copyrighted material (one remembers Google getting into trouble over scanning all the books in the US Library of Congress - the USA copyright library) and it could well be that these big tech companies have broken the law again.

The lawyers assisting the group, Matthew Butterick and Joseph Saveri, are already involved in an earlier case against OpenAI brought by two authors.

They write that "since the release of OpenAI's Chat­GPT sys­tem in March 2023, we've been hear­ing from writ­ers, authors, and pub­lish­ers who are con­cerned about its uncanny abil­ity to gen­er­ate text sim­i­lar to that found in copy­righted tex­tual mate­ri­als, includ­ing thou­sands of books".

[johntel]

Can ChatGPT do a cryptic crossword?

www.telegraph.co.uk/lifestyle/2023/03/06/can-ai-chatbot-solve-crossword-better-human-put-test/

jameswillia.ms/posts/chatgpt-cryptics.html

[c-a-r-f-r-e-w]

Jul 12, 2023 13:43:47 GMT leftieliberal said:

Sarah Silverman sues OpenAI and Meta for misuse of copyrighted material.

www.bbc.co.uk/news/technology-66164228

Apart from the exemption for "fair dealing and research" there are strong restrictions around the use of copyrighted material. It seems that these Large Language Models were trained using copyrighted material (one remembers Google getting into trouble over scanning all the books in the US Library of Congress - the USA copyright library) and it could well be that these big tech companies have broken the law again.

The lawyers assisting the group, Matthew Butterick and Joseph Saveri, are already involved in an earlier case against OpenAI brought by two authors.

They write that "since the release of OpenAI's Chat­GPT sys­tem in March 2023, we've been hear­ing from writ­ers, authors, and pub­lish­ers who are con­cerned about its uncanny abil­ity to gen­er­ate text sim­i­lar to that found in copy­righted tex­tual mate­ri­als, includ­ing thou­sands of books".

similar issues with music. But as more AI-generated content gets put out there, AI may increasingly be “cannibalising” itself as they put it. Copying other AI stuff.

AI lawyer bots prosecuting a case against an AI for copying summat by another AI. Which in turn lifted summat from a human. (Who used an AI to generate some of it…)

🤯🤯🤯🤯

[leftieliberal]

Jul 19, 2023 18:25:35 GMT c-a-r-f-r-e-w said:

Jul 12, 2023 13:43:47 GMT leftieliberal said:

Sarah Silverman sues OpenAI and Meta for misuse of copyrighted material.

www.bbc.co.uk/news/technology-66164228

Apart from the exemption for "fair dealing and research" there are strong restrictions around the use of copyrighted material. It seems that these Large Language Models were trained using copyrighted material (one remembers Google getting into trouble over scanning all the books in the US Library of Congress - the USA copyright library) and it could well be that these big tech companies have broken the law again.

The lawyers assisting the group, Matthew Butterick and Joseph Saveri, are already involved in an earlier case against OpenAI brought by two authors.

They write that "since the release of OpenAI's Chat­GPT sys­tem in March 2023, we've been hear­ing from writ­ers, authors, and pub­lish­ers who are con­cerned about its uncanny abil­ity to gen­er­ate text sim­i­lar to that found in copy­righted tex­tual mate­ri­als, includ­ing thou­sands of books".

similar issues with music. But as more AI-generated content gets put out there, AI may increasingly be “cannibalising” itself as they put it. Copying other AI stuff.

AI lawyer bots prosecuting a case against an AI for copying summat by another AI. Which in turn lifted summat from a human. (Who used an AI to generate some of it…)

🤯🤯🤯🤯

www.bbc.co.uk/news/uk-wales-66099850 AI: Digital artist's work copied more times than Picasso

[c-a-r-f-r-e-w]

Another use for coffee:

Full of beans: scientists use processed coffee grounds to make stronger concrete
Australian engineers say they can make concrete nearly 30% stronger by incorporating processed coffee grounds into the material

www.theguardian.com/science/2023/aug/23/full-of-beans-scientists-use-processed-coffee-grounds-to-make-stronger-concrete

and…

‘The office is for socializing’: how work from home has revolutionized work
Even as some firms are mandating workers come into the office, work from home has permanently changed the work landscape

www.theguardian.com/business/2023/aug/22/work-from-home-revolution

[c-a-r-f-r-e-w]

Aug 24, 2023 20:01:10 GMT jen said:

Aug 24, 2023 19:19:17 GMT c-a-r-f-r-e-w said:

Wow, thanks Jen! You’ve got a Freqout! I’ve been keen to try one of those. I found the EQ7 can be fun trying to recreate other pickups like a P90, but these days with modelling, maybe superceded for that. Like the idea of the Caline 57 deluxe sim - Have you been tempted by any of the more all-encompassing modellers, Kemper etc.?

(I’ve got some older pedals, like a Mesa Boogie V-Twin, a Cream Extreme etc., and a VG99 using a GK Pickup, but am rather behind the current curve)

Haha! Now you've caught my interest... Yes I've tried digital modellers like the Kemper, they don't sound that bad these days, but it's the latency that is a turn off for me. I know it's only around 3ms, and if you're jamming at home you are unlikely to notice it, but these day all medium to large stages use a digital mixer, so your monitors are already going to have twice that latency (1 x in + 1 x out), say around 5ms, and I can just about live with that. Adding the 3ms from a modeller would bring us up to approx 8ms and that, quite frankly, drives me mental. That's also why I avoid digital pedals apart from the Freqout (which merely adds a digital signal to the analogue input) and the Ricochet (which I only use on long sustained notes where exact timing is of no consequence). The Caline Americana is completely analogue and has zero latency (it's a Chinese knock-off of the Tech21 Blonde, but I prefer the Caline's EQ curves).

I have a number of other analogue amp sims and quite enjoy them all. If you would like to discuss them I suggest starting another thread.

I would never use a modeller to try to simulate other pickups, for various reasons it can never really work. And anyway, I have lots of guitars with lots of different pickup configurations, I would have no need, even if the sim worked.

Yes, I take your point about latency. I have wondered about the Blonde, I’m quite keen on the Fender Twin sound though and some say pedals don’t tend to have the headroom to match the Twin?

What other analogue amp sims do you use Jen?

(I know what you mean about modelling pickups; in situations where you just have the one guitar, it can be handy, as can switching to different tunings without switching guitars if you use something like the GK pickup).

[jen]

Aug 24, 2023 20:43:28 GMT c-a-r-f-r-e-w said:

Aug 24, 2023 20:01:10 GMT jen said:

Haha! Now you've caught my interest... Yes I've tried digital modellers like the Kemper, they don't sound that bad these days, but it's the latency that is a turn off for me. I know it's only around 3ms, and if you're jamming at home you are unlikely to notice it, but these day all medium to large stages use a digital mixer, so your monitors are already going to have twice that latency (1 x in + 1 x out), say around 5ms, and I can just about live with that. Adding the 3ms from a modeller would bring us up to approx 8ms and that, quite frankly, drives me mental. That's also why I avoid digital pedals apart from the Freqout (which merely adds a digital signal to the analogue input) and the Ricochet (which I only use on long sustained notes where exact timing is of no consequence). The Caline Americana is completely analogue and has zero latency (it's a Chinese knock-off of the Tech21 Blonde, but I prefer the Caline's EQ curves).

I have a number of other analogue amp sims and quite enjoy them all. If you would like to discuss them I suggest starting another thread.

I would never use a modeller to try to simulate other pickups, for various reasons it can never really work. And anyway, I have lots of guitars with lots of different pickup configurations, I would have no need, even if the sim worked.

Yes, I take your point about latency. I have wandered about the Blonde, I’m quite keen on the Fender Twin sound though and some say pedals don’t tend to have the headroom to match the Twin?

What other analogue amp sims do you use Jen?

(I know what you mean about modelling pickups; in situations where you just have the one guitar, it can be handy, as can switching to different tunings without switching guitars if you use something like the GK pickup).

The AMT F1, it does a decent Fender Twin sim, and you can run it at 18v to get more headroom.

I also have the DSM & Humboldt Simplifier, an excellent Fender Deluxe (and switchable to AC30 or Marshall) sim, though it breaks up too early for me to use it on a live board, just like a real Deluxe.

Marshall JMP-1 - 19" tube preamp with optional speaker sim output. Does all your standard Marshall tones. Used it for years (live and on many records) when I was in my Gibson + Marshall phase.

Joyo AC-tone - decent enough, but when I want the Vox tone I'll generally take the trouble to mic up one of the two amps I have that do that.

I also had the Joyo version of the Tech21 Blonde/Caline Americana. Decent enough, it's like the Caline but has less headroom. I've given it to the other guitarist in my main band and he uses it all the time.

Also got the AMT Chameleon Cab - a very tweakable speaker sim. Good fun, but don't need it on a live board.

FYI, I always play live through an amp. I have several 90s USA Fender solid state combos, but if the gig has backline provided I can work with any solid state or tube amp (just need to tweak the Caline to taste). If all they got is digital shite, and I can't be bothered to take a combo, I'll take a little Warwick Gnome 130W head (nominally a bass amp) that's so small it fits into the side pocket of a guitar gig-bag, and I'll plug that into whatever cab they have there.

Live I play Fender Jaguars and Mustangs, and just the neck pickup. (I am lead vocalist as well as lead guitarist, and what with engaging with the audience, I have no time to think about switching). The pickups are hand wound for me by Jaime at The Creamery to my personal specs, but the neck pickup is loosely based on his Sonic 60 bridge pickup. I have other guitars with vintage PAFs, P90s, Filtertrons, Firebirds, Fender Cunife Wide Range Humbuckers, Lipsticks, and classic Fender single coils, so I'm well covered. (edit: I forgot the Goldfoils!)

Any other questions?

(I thought you were a synth person? Don't get me started on my analogue keyboard collection, or my studio outboard, or on my guitar and bass amps. And definitely don't get me started on my other pedals, I have way over 100 haha! And whatever you do, never ever get me started on guitars and basses - although I am sure that one member here will be relieved to hear that my most valuable guitar is an acoustic...)

[c-a-r-f-r-e-w]

Aug 24, 2023 22:04:37 GMT jen said:

Aug 24, 2023 20:43:28 GMT c-a-r-f-r-e-w said:

Yes, I take your point about latency. I have wandered about the Blonde, I’m quite keen on the Fender Twin sound though and some say pedals don’t tend to have the headroom to match the Twin?

What other analogue amp sims do you use Jen?

(I know what you mean about modelling pickups; in situations where you just have the one guitar, it can be handy, as can switching to different tunings without switching guitars if you use something like the GK pickup).

The AMT F1, it does a decent Fender Twin sim, and you can run it at 18v to get more headroom.

I also have the DSM & Humboldt Simplifier, an excellent Fender Deluxe (and switchable to AC30 or Marshall) sim, though it breaks up too early for me to use it on a live board, just like a real Deluxe.

Marshall JMP-1 - 19" tube preamp with optional speaker sim output. Does all your standard Marshall tones. Used it for years (live and on many records) when I was in my Gibson + Marshall phase.

Joyo AC-tone - decent enough, but when I want the Vox tone I'll generally take the trouble to mic up one of the two amps I have that do that.

I also had the Joyo version of the Tech21 Blonde/Caline Americana. Decent enough, it's like the Caline but has less headroom. I've given it to the other guitarist in my main band and he uses it all the time.

Also got the AMT Chameleon Cab - a very tweakable speaker sim. Good fun, but don't need it on a live board.

FYI, I always play live through an amp. I have several 90s USA Fender solid state combos, but if the gig has backline provided I can work with any solid state or tube amp (just need to tweak the Caline to taste). If all they got is digital shite, and I can't be bothered to take a combo, I'll take a little Warwick Gnome 130W head (nominally a bass amp) that's so small it fits into the side pocket of a guitar gig-bag, and I'll plug that into whatever cab they have there.

Live I play Fender Jaguars and Mustangs, and just the neck pickup. (I am lead vocalist as well as lead guitarist, and what with engaging with the audience, I have no time to think about switching). The pickups are hand wound for me by Jaime at The Creamery to my personal specs, but the neck pickup is loosely based on his Sonic 60 bridge pickup. I have other guitars with vintage PAFs, P90s, Filtertrons, Firebirds, Fender Cunife Wide Range Humbuckers, Lipsticks, and classic Fender single coils, so I'm well covered. (edit: I forgot the Goldfoils!)

Any other questions?

(I thought you were a synth person? Don't get me started on my analogue keyboard collection, or my studio outboard, or on my guitar and bass amps. And definitely don't get me started on my other pedals, I have way over 100 haha! And whatever you do, never ever get me started on guitars and basses - although I am sure that one member here will be relieved to hear that my most valuable guitar is an acoustic...)

That’s a very interesting reply Jen, lots to ponder, and am suitably intrigued by your gear collection but I have to dash now and shall reply tomorrow if that’s ok (and indeed, I’m partial to a synth but started out on drums, then got into guitar before discovering keys…)

[c-a-r-f-r-e-w]

jen

Thanks again for your reply Jen. That’s worth knowing about the AMT, that you can run it at 18V. I shall have to give that a closer look. Are your different amp sims/preamps mostly for the studio, or do you take more of them out live at times? 

Good to know you had a Marshall phase! I had a Marshall phase too, though it was a Strat and Marshall, and had a 100 watt head and 4 x 12 cab at one point, but I was soon to discover the Funk (and that Clapton had switched from Marshalls to Fender amps). It’s still a great sound for some things… have you been influenced by particular guitarists? 

Do you greatly prefer tube to solid-state amps, or are you not so fussed? It’s cool you get your pick ups from the Creamery, and am impressed with the selection you use! I keep meaning to try gold foils, quite partial to a bit of Ry Cooder. Quite like mini-humbuckers at times…

I like your idea of taking a head and plugging it into a cab when you get there… I have a Rivera combo that is not exactly light.

Regarding your other pedals, what do you use for reverb? And of course feel free to let me know about your analog keys! Btw, what interface do you use for recording?

[jen]

I have the Caline Americana on all of my pedalboards, I use it in the signal chain going into my live amps, I find it a great tool to sculpt the sound of whichever amp I amp I am forced to use. I do the same when I use an amp of my choice (atm I especially like my Fender Stage 100 and my Fender Deluxe 112, both 90s USA made solid state amps). The other sims only get used in the studio when I want a quick and dirty solution (generally I will mic up a real amp, I've got lots).

Guitar influences? Originally Steve Jones (Sex Pistols) and Daniel Ash (Bauhaus). Now I have my own style.

I like both tube and solid state amps. When I record I like to split the signal and have one of each. My favourite combination atm is a Kustom Coupe 36 (tube) + Fender Stage 100 (solid state). But I am prone to experiment, I have a lot of amps. But the 90s solid state amps are real bargains. I regularly gig with a Fender Deluxe 112, you can easily get them for £100 used, they are loud, reliable, and have "that" Fender clean sound. I refuse to play modelling amps. Not just because of the latency. They sound quite good, but that ain't good enough for me. But the technology is in its infancy, I'm certain they'll get better...

Goldfoils: to be honest, I don't get the hype, but they are good fun. I have two weird beasties with them, a Silvertone and an Italia. Strange and quirky guitars. I used them at a live show when we were playing a "Vampire Night" haha! I don't like Ry Cooder. I am not especially impressed with mini-humbuckers. For me, they react just like PAFs but with a different frequency response... you can use EQ for that. I don't have any ceramic pickups (just AlNiCo or CuNiFe for me). I think ceramic pickups are shite.

Hey! Some of those Rivera combos sound really cool, but yeah, they are all really heavy. My tiny Warwick head is really practical, negligible weight. I'm currently putting together a 1x12 cab with a solid pine body and a neodymium speaker, it's going to be very light weight. I'll be very interested to find out how it sounds...

Reverb pedals (naturally I prefer amp reverb)... the best is from Anasounds, you use it to drive a real spring reverb tank. The company offer three different tanks (I have two of each!). I use these for recording. For live I'll use either a Mad Professor Silver Spring Reverb (Fender amp style), an EarthQuakerDevices Ghost Echo (spooky spring reverb), or a Digitech Polara (all round high end reverb with Lexicon algorithms).

Interface: everyday - Audient iD44; serious multitracking - RME Digiface with Ferrofish Pulse 16 AD/DA. I also have an Otari 8-track reel-to-reel tape machine and two Revox 2-tracks!

The analogue keys can wait for another day... I am pooped! (I had a live show tonight.)

What guitars do you play?

[leftieliberal]

Vox article on why generative AI is just a 'party trick' for now
www.vox.com/technology/2023/8/19/23837705/openai-chatgpt-microsoft-bing-google-generating-less-interest

[leftieliberal]

One question I have always wondered about is why life uses one form (L-amino acids and D-sugars), but not the other (D-amino acids and L-sugars) even though abiological processes should have produced both equally. Recent work suggests that the answer may be magnetism.

www.quantamagazine.org/magnetism-may-have-given-life-its-molecular-asymmetry-20230906/

The study didn’t directly explain why life’s preferred nucleotides are right-handed and its amino acids are left-handed, Ozturk said. But these new findings suggest that the determining factor was the magnetization induced by the Earth’s field. Athavale noted that even if the crystallization process happened in 100 primordial lakes, Earth’s magnetic field would ensure that they all produced precursors with the same handedness rather than a mixture.

Joyce noted that there’s a “cool little twist” if the magnetic field gave such a bias: If life started in the northern hemisphere and favored molecules with one handedness, then it would have shown the opposite handedness if it had arisen in the southern hemisphere.

It does suggest that if we find extra-terrestrial life there will be only a 50% chance it will be of the same handedness as us.[1] and it may be that a planetary magnetic field is a requirement for life.

[1] life originating beyond the Solar system; life within the Solar system could have come from Earth by meteorites, we have found meteorites from Mars on Earth.

[leftieliberal]

Microsoft Research on AI
While most people will associate Microsoft's work in this field with OpenAI (which they do fund) here is some work they have been doing themselves:

www.microsoft.com/en-us/research/blog/phi-2-the-surprising-power-of-small-language-models/

What is interesting is that just by ensuring that the input data is of high quality, the "textbooks are all you need" approach, they can get better results from smaller language models than others are getting from larger language models. It illustrates that the ChatGPT approach of "throw the whole internet at the wall and see what sticks" is far from optimum.

[leftieliberal]

Hyperloop company shuts down. As with so many of Elon Musk's ideas, he underestimates the difficulty in turning an idea into reality. Tesla's Autopilot is another example.

[alec]

leftieliberal - I as thinking about Tesla's Autopilot and connected thoughts a couple of weeks ago, when stuck in some snow. More accurately we weren't stuck, but the numpties ahead of us that came out to 'look at the snow' but didn't have a clue how to drive on it were.

It was one of the high level roads over the Pennines near us, with a couple of steepish banks to get over, and with 2 - 4 inches of decent quality snow. Everything was OK, cold, quite grippy, so no real issues, so long as you stayed off the brake on the downhills and got your speed up for the climbs. We waited at the foot of one of the banks to ensure we had a clear run up without any other drivers on that section, so we could tank it up the slope. Several dummies went past me. crawling slowly, and got completely stuck. We pushed them down in the end and showed them how to drive in snow, and got home without any bother, but when I was sitting waiting, I did ponder how driverless cars are going to ever be able to understand how to cope with such conditions. You needed to look around half a mile ahead, not start with any other vehicle in sight, get your speed to a level that is technically not safe for those road conditions, but essential to make the hill, etc etc. I'm not sure how driverless cars will deal with that.

[leftieliberal]

Dec 22, 2023 15:37:54 GMT alec said:

leftieliberal - I as thinking about Tesla's Autopilot and connected thoughts a couple of weeks ago, when stuck in some snow. More accurately we weren't stuck, but the numpties ahead of us that came out to 'look at the snow' but didn't have a clue how to drive on it were.

It was one of the high level roads over the Pennines near us, with a couple of steepish banks to get over, and with 2 - 4 inches of decent quality snow. Everything was OK, cold, quite grippy, so no real issues, so long as you stayed off the brake on the downhills and got your speed up for the climbs. We waited at the foot of one of the banks to ensure we had a clear run up without any other drivers on that section, so we could tank it up the slope. Several dummies went past me. crawling slowly, and got completely stuck. We pushed them down in the end and showed them how to drive in snow, and got home without any bother, but when I was sitting waiting, I did ponder how driverless cars are going to ever be able to understand how to cope with such conditions. You needed to look around half a mile ahead, not start with any other vehicle in sight, get your speed to a level that is technically not safe for those road conditions, but essential to make the hill, etc etc. I'm not sure how driverless cars will deal with that.

There are plenty of places in North America that have those conditions in the winter.

[lens]

Dec 22, 2023 15:37:54 GMT alec said:

leftieliberal - I as thinking about Tesla's Autopilot and connected thoughts a couple of weeks ago, when stuck in some snow. More accurately we weren't stuck, but the numpties ahead of us that came out to 'look at the snow' but didn't have a clue how to drive on it were.

........... I'm not sure how driverless cars will deal with that.

I suspect no worse than "the numpties" you encountered a couple of weeks ago! That may sound sarcastic, (and to an extent it is!) but there's a serious point.

The key question is not whether driverless cars will start off perfect (I guarantee they won't) but whether they are better than an average driver *most of the time*. Because if so, it's likely to lead to an *overall* decreased number of incidents, which should be seen as a good thing?

More specifically, and in connection with your point above, maybe the critical thing will be how they deal with the situation you describe. Will they just blindly "have a go", or default to "OK, human driver - you're in charge now". (I hope it's the latter.)

There's a lot more to the whole driverless car debate than just driving in snow (though you bring up a good example with that). A more mundane thought would be if I was (say) at the supermarket, and told my driverless car to take me home. OK - via GPS it finds the way back, but then what? Does it know whether to go in the garage, in the drive, and whether to go in forwards or backwards? (Easy to unload shopping one way - difficult the other.) Or to stay in the road because my daughter is driving back soon? And if the latter, avoid parking in one spot in the road because I know that would irritate that particular neighbour?

My own thought is that Tesla with "Autopilot" has probably got the name right. A basic aviation autopilot is not the same as a "pilotless plane", rather it relies on a pilot to do the difficult bits, typically mostly at start and end of the flight, then just keeps the plane straight and level in the cruise and takes a lot of the drudgery out of the pilots workload. And that's what I'd initially expect with a car. Set your destination and let it take you most of the way there, managing *most* traffic situations, but with an expectation that there may be occasions where it says "excuse me, but could you help me out here?"

That's at least the first generation. Who knows where we'll be in 20 or 30 years time? The Tesla Model 43 driving itself up a snowy hill criticising "that numpty human driver alec - why do they let humans drive cars themselves?"  as it whirrs past.  

[alec]

lens - good points.

One big question though, which I think is yet to be resolved, and which does have consequences for how we view this technology, will be who is held responsible if there is an accident? We like to have a human in control, because there is an obvious line of responsibility. But if it is a software issue, a lack of experience in terms of the driverless system meeting an unforeseen circumstance?

I have no idea what the answer is, but I think your general points are highly valid - they are likely to be safer overall, and 'autopilot' is a good way to use them.

Have a good Christmas.

[lens]

And to alec, then yes indeed re legalities and responsibility. I had intended to mention exactly that point.

Of two things I'd be pretty sure. A framework will be developed, but it won't be without a lot of debate and expensive lawyers. I believe such is already well underway.

It's worth thinking of a couple of other circumstances which have parallels. One may be the introduction of compulsory seat belt wearing. Look at the stats and it undoubtably did lead to a reduction in deaths and injuries - but there were instances where seatbelt wearing had a negative effect. (Though well outnumbered by the instances where the effect was positive.) But AFAIK noone who had such a negative outcome (eg trapped longer in a car) was able to bring a case against a seatbelt manufacturer. They could only face legal action in the event (say) that the mounting or mechanism was faulty in manufacture.

Vaccination may be another comparison. Again, a situation where *overall* an individual in a population is statistically far more likely to have a health benefit if vaccinated - but for a few individuals the vaccine may have a harmful effect. To any single person you really have to go with the odds - the benefits *should* outweigh the risks.

And I think driverless cars will go the same way. A recognition that *in general* the safety benefits will outweigh any potential risk, and the legislation controlling this will reflect such to encourage it. Possibly the main area of interest (at least in the early generation) will be how a system fails, when it undoubtably will occasionally. And I'd hope the fail-safe it would adopt would be to say "help" to the driver when something happens beyond it's experience.

Finally, I think it's worth looking back and to remember that personally it's only about 30 years ago that I first ever saw anyone use the internet. Only 30 years! And less than 20 years since I had my first digital camera. I may not be around to witness it, but it's why I wouldn't make too many bets about what self driving cars will be in 20-30 years time. 10 years may seem a long time whilst living it, but is little in a historical perspective. In 1920 electricity in the home was a rarity, in 1930 it was commonplace.