TEARS OF RAGE?

Dick Pountain /Idealog 379/ 6th February 2026 : 10:07am 

It feels as though we’re currently living in a peculiar sort of limbo, because deep-down we all know that AI is bound to end in tears, but many questions remain about what kind, from whose eyes, at what flow-rate and salinity, whether caused by economics, politics or by ecology, of rage or hilarity? (On that latter point, I’m guiltily amused by the cruel wag on Substack who suggested that we should start calling Sam Altman ‘Sloppenheimer’).   

I’m acutely aware of the amount of space I’ve already devoted in this column, over many years, to expressing my scepticism of the strongest claims made by the AI crowd, and I don’t wish to bang on using those same arguments, but I might just have found a new one to bang on about. Of course, living in the world of Einsteinian space-time as I must, it’s quite possible that an upsetting event might occur in the finite period between me typing this and it appearing in your copy of PC Pro. I’m typing it on Friday February 6th 2026, and nVidia has just announced that it’s only going to ‘lend’ OpenAI $20billion instead of the $100billion they asked for (and perhaps need?) Were this to bring down the company by disrupting the bizarre circular funding model (‘I lend you the money to buy my products and charge you for them’) that might trigger a chain of crashes that will answer the economic question. Perhaps I’ll be practicising a new guitar arrangement of ‘Buddy Can You Spare A Dime’. 

The political question depends upon how many jobs AI will take away, which depends in turn on how smart it can get, which depends on whether you believe that AGI is arriving this year, or even soon. No-one, not even Altman, really believes that any longer, and it’s no longer just Gary Marcus howling in the wilderness but even Sutskever, even the god-father LeCun now admit it ain’t going to happen ever. The ecological question – how much of the earth’s surface will we allow them to cover with datacentres. using how much of our water and electricity? – will only be answered by old-fashioned politics (perhaps assisted by pitchforks, burning brands and barricades) when and if the USA ever returns to sanity. Such a return might be precipitated by the slop/enshittification problem, as major social media, advertising and other communications companies begin to suffer massive defections of customers disgusted by the torrents of AI slop that render their products unusable.

My new AI gripe is that the grotesque squandering of wealth which AI companies are currently perpetrating stems from a fundamental asymmetry between training and querying. Our current generative AI chatbots, agents and assistants operate by scraping the textual and visual assets of the whole world internet, concentrating these into inconceivably huge tensor spaces of weights, which once scraped are essentially 

static data structures. Answering queries from such a space is so inexpensive you can perform it on the most modest smartphone or laptop, but incorporating new knowledge into it is only possible by retraining the whole thing on a colossal datacentre crammed with millions of nVidia chips. Whenever you have a dialogue with ChatGPT, Gemini or Claude you’re sipping from a concentrated extract of one particular snapshot of the whole world’s knowledge, but what transpires between you and that LLM does not get written back into the model. This kind of asymmetry will be familiar to experienced computer programmers as the difference between ‘compile time’ and ‘runtime’. I needed the opinion of someone smart enough to grasp the importance of this distinction, so of course I asked ChatGPT. 

    ME: “Is it the case that facts imparted by me during a conversation do NOT get incorporated back into         the LLM?!

    CHATGPT: “Yes — that is essentially correct with one important nuance. Facts you tell me during a         conversation are not incorporated back into the Large Language Model itself. The underlying LLM is         fixed at the time it’s deployed and does not learn or update its weights from individual conversations. It      cannot absorb new facts in the way a human does. This is desirable for Safety: it prevents poisoning,         manipulation, or accidental corruption. Stability: a model that changed continuously would be                     unpredictable and impossible to validate. Privacy: your individual inputs aren’t being absorbed live             into a shared intelligence.”

A candid, lucid and compact summary, typical of what LLMs are good at: they make smart librarians, even teachers, but not innovators. Frantic competition between the AI baronets to create the AGI God-Server is a wasteful, dangerous product of ignorance and immaturity. The world needs far fewer actual training datacentres, subject to government inspection and regulation, to which everyone should have free access as a public good as part of the education system…

[Dick Pountain does treat ChatGPT like a person, but one who’s even nerdier than his real friends]

AWFUL LOT OF COFFEE

Dick Pountain /Idealog 378/ 6th January 2026 : 10:03am 

I’ve loved coffee for most of my life, which is mildly unusual in a tea-drinking nation. During childhood my mother made a sweet beverage from ‘Camp’, a jet-black treacly extract (with colonial soldiers on the label) which was diluted with hot milk. I discovered real coffee only in my teens when the Wimpy Bar first arrived in town: brewing coffee with real cream in a little jug on the side, and managed by a Jamaican youth of our age, up from London, who also brought bluebeat records and became our hero. Later, having moved to London to study, I discovered espresso (known as ‘froffee coffee’ to sceptics) and later still lived in Italy for some years and rediscovered proper espresso. But what, I think I hear you ask, does this have to do with computers?Well, an awful lot actually. Caffeine is a brain stimulant of a tone and strength particularly suited to difficult mental problem-solving (which programming certainly is), far more so than amphetamines which merely lead down rabbit holes. The Hungarian mathematician Alfréd Rény was addicted and once said  “A mathematician is a device for turning coffee into theorems”. During the boom of the 1980s, the birth years of our current main operating systems, US programmers found coffee insufficiently caffeinated and demanded a supercharged additive called ‘Jolt Cola’, about which the less said the better. My own first encounter with a US tech giant was at a regional Hewlett Packard HQ, where I noticed that at each main intersection of its many corridors stood an elegant automatic espresso machine that delivered shots in little paper cups for free. Why, even PC Pro’s own Sasha Muller reviews coffee makers for the Guardian. I make no more excuses and rest my case. 

But of course computers-and-coffee nowadays means YouTube videos about the apparatus and technique of squeezing espresso shots. This, like photography, is an activity with so many crucial independent variables that it opens up a space for fetishisation, argumentation (often heated), OCD-level attention and artistry. And no-one has a stronger claim to be its Pope, Dalai Lama or Cartier-Bresson than James Hoffman. In a couple of previous columns I’ve mentioned my amusement at his style, but not intended as insult: I admire his highly successful combination of wry humour and scientific rigor (even his extravagant coiffure I’ve learned is in tribute to the late, great David Lynch). His recent animated feature on the laxative properties of coffee is both hilarious and a template for a whole new way to teach biochemistry…

While I’ve learned much from Hoffman’s videos about the importance of bean roast, of grinders and grind size, of pressurised versus non-pressurised coffee baskets, I am however not fully converted. As a qualified organic chemist I appreciate his sound explanations of the extraction process - which rise above the equipment-fetish of many of his imitators - but I’m also a pragmatist who baulks at spending hundreds, let alone thousands, on hardware, and has no time for roasting my own beans. I’ve had the same faithful DeLonghi Dedica espresso machine for over a decade, running Lavazza Black ground coffee from the supermarket though it – but watching Hoffman has prompted me to attempt some inexpensive ways to up my game. I bought a non-pressurised basket on Amazon and modified the Dedica’s portafilter to accept it by removing the plastic insert (I still refuse to hacksaw the bottom off it and agonise over drip patterns). I noticed an improvement in taste, with less bitterness and acidity, but also a weaker shot. 

I also bought an inexpensive hand grinder with adjustable burrs to grind whole beans, but that proved too tedious even after I added a £10 Chinese electric mini-drill to drive it. At around the same time greed-flation struck the price of Italian ground coffees which shot up by 50% in weeks, and I started examining Sainsbury’s range of single-country ground coffees, settling on their Fairtrade Colombian as quite tasty. The problem is they’re ground to be suitable also for cafetières or pour-over, way too coarse for my non-pressurised filter basket and so yielding an even weaker shot. But, given that I’m not fully inducted into the cult I decided to commit the ultimate heresy by regrinding the coffee myself with a non-burr grinder! A piece of hardware I’ve owned ever since my biochemistry days is a vintage US blender, a Waring, capable of liquidising whole rats, golf-balls, possibly even golf-clubs. I tip a 230gm bag of ground Colombian into its graduated cup and whizz until it’s reduced by 20% in volume (takes around 30 seconds) then tip it back into the repurposed Illy tin I use for storage. I may never reach coffee nirvana but it does taste pretty good. 

[Dick Pountain does also enjoy tea, particularly Chinese Lapsang souchong or Pu’er]

         

LIVE LONG AND PROSPER?

Dick Pountain /Idealog 377/ 24th November 2025 : 11:39am 

Mankind has always hankered after immortality… Sorry about that, but I’ve been overdosing on social media for several weeks and began to worry that my column isn’t bombastic enough. And it’s not just social media either, because last week The Guardian ran a piece about a US investment boom into ‘transhumanist’ technologies like Musk’s Neuralink that claim they’ll enable us to upload our brains and download them again into new bodies. Concerned US scientists say that this is starting to divert funding away from sensible medical research. I’ve written plenty here about the delusions most software engineers (even billionaire ones) harbour about brain function and so am not going there again, except to say that it isn’t static data, it’s an evanescent flux, geddit? 

What interests me more is the replacement body aspect, prompted by watching Benicio del Toro’s excellent new film of ‘Frankenstein’, followed by Alex Garland’s equally excellent ‘Ex Machina’ for the third time on TV. Mary Shelley was neither a software engineer nor a neurologist and hence avoided such delusions by getting Victor Frankenstein to use a real, wet brain from another human, along with all the other necessary bits collected from fresh battlefield slaughter. Garland’s egomaniacal Ethan makes a brain for his gorgeous female robot Ava from a lump of a special jelly, which he trained in LLM-style using the output of all the world’s mobile phones (which he stole with the full connivance of the phone companies). It’s as well to remember that the phrase ‘deus ex machina’ originally referred to the crude tricks that playwrights of old used to fill in gaping holes in their plots, and Garland’s movie isn’t just a remarkably prescient prediction of AI lunacy but also a cool and sardonic satire of the industry. 

Of course the biggest problem for all transhumanist fantasies is powering the creatures. Shelley evaded this one neatly too, because her Monster just eats food like the rest of us (though the deeply unpleasant Victor F. in the new film probably skimped on his rations). Ava on the other hand appears to plug herself in like a smartphone every night, and a crucial plot-point is that this gives her enough knowledge of the power system to create the electricity blackouts which drive the story to its gory climax. It’s not obvious where she stores charge during the day though since her transparent midriff, while curiously titillating, reveals nothing resembling an adequate battery. And that brings us to the delicate matter of what these creatures are created for. 

Victor Frankenstein made it quite clear that his purpose was unambiguously to ‘conquer death’, rather than anything more prosaic or economic, so Mary Shelley didn’t pursue what Victorian industrialists would have used his work for had it succeeded. Garland though is quite clear about 

what’s on Nathan’s mind, because he gives Ava functional genitalia and a heterosexual ability to entrance susceptible males. Nathan is basically a dirty middle-aged mogul. A reasonably convincing static version of Ava might be cobbled together today by connecting the latest generation of Chinese Sex Dolls - which I know about solely via a YouTube video I hasten to add - to ChatGPT: they wouldn’t be able to walk and would require mains power, but I’m sure there’s a sizeable market for whom that wouldn’t matter. 

During her final fight with Ethan, he rips off one of Ava’s arms, but having dispatched him with a big old kitchen knife she calmly goes to his wardrobe, unplugs an arm from one of his previous girlfriends hanging up there, and snaps it onto her stump. Having wrestled more than I wanted with USB-C 3.2 recently, I find Nathan’s modular nervous system, which can suspend and restore pain, touch and motion to a severed limb via a bayonet snap-on connector, pretty damned impressive. Which leads to the final, caustic, irony of the movie where having escaped Nathan’s mountain eyrie, Ava fulfills her wishes by people-watching at a busy traffic intersection in San Francisco. Where she’s going to recharge herself that night, or obtain any future snap-on appendages is rather glossed over….  

So much for the logistics of body replacement. Much physiological research goes on into the ageing process, hoping to discover ways to mitigate it, but my gut feeling that while this might extend lifespans by 25-50% it will eventually hit a wall: we’re made of inescapably perishable materials which need to be continually replenished, and that process must eventually yield to entropy and error-accumulation. But why would anyone want to live forever anyway? The real enemy is not so much death as boredom. History teaches that the pleasures of the flesh, of caviar and Lamborghinis, even of having one’s ego inflated by ruling the world, will eventually pall.  

[Dick Pountain has a very long, but finite, attention span]

SUBSTACKED

 Dick Pountain /Idealog 376/ 6th November 2025 : 10:10am 

This column has been running for thirty years now, and has never ever felt like a chore. I garner ideas from tech problems with my own computers, from films, websites, books, papers, and I remember them using some software tool that has changed repeatedly over the years, from Idealist to Google Keep, from saved HTML to Pocket (and now Instapaper). But however they’re recorded I expand on them, write then send a file by email to a publisher who prints onto paper using heat-set web-offset lithography, sells the result to you (and to advertisers) and pays me for writing it. That’s just how the world worked, but in recent months it’s begun to feel like a rare privilege, and that’s because I started using Substack.

I opened my Substack account back in 2018 when it was a different beast altogether, intended for online self-publishing of lengthy works like books with a built-in payment mechanism. I tried it with little success, eventually used Kindle instead and then forgot about it until quite recently. In the meantime Substack expanded its features to include blogs, podcasts, and Twitter-alike short notes, and then in 2025 it took off like a rocket. I discovered this when Facebook, which I’d been using for years to keep in touch with friends and post photographs, became a swampy mess of AI-generated guff (mostly about space travel, pseudo-science and historic artefacts in my case) which even tools like Fluff Busting Purity failed to control. I resolved to try-out both Substack and the newly launched BlueSky as possible alternatives, and discovered that Substack was filling up at an astonishing rate with high-quality content about politics, economics, AI, science, literature and more and then realised why that was happening. MAGA. 

Much of the serious material now appearing on Substack is there because it’s written by American academic and other writers who got displaced from jobs, consultancies and other positions in the extraordinary rampage that Elon Musk’s DOGE army perpetrated on American academia and media in the first half of 2025. And having lost valuable sources of income, they look to Substack to garner paid subscriptions and stay afloat. Not all of this new blood was DOGEd-out but their traffic enhances the perception of Substack to such an extent that people who still have jobs feel a need to be there too. And it’s not only in the USA because media shenanigans in the UK at the BBC, Observer and elsewhere have sent stars like Carol Cadwalladr there too. All of this creates a remarkable new ecosystem which I’m only just coming to terms with, and which often threatens to Sub-merge me (see what I did there?) 

There are several downsides. The first is that Substack’s user interface is a nightmare, even compared to Facebook. Its history of shifting focus 

by bolting on new subsystems makes finding things a pain – I never know whether something I post will appear in the app or emails (and ditto with replies). The second is that when everyone is trying to make a living, almost all articles have paywalls of varying height and permeability. A few (me included) leave everything free; some make a few important pieces free; some make everything paid; but most offer you a teaser, perhaps half the article, perhaps three-quarters, at which point the [Subscribe] button pops up to read the rest. 

A third problem is that if I do subscribe by choosing a free option (not always available), that automatically enrolls me into their mailing list for new posts. Now a lot of the political and economic material is very useful to me with my other, book-reviewer’s, hat on, and so I end up subscribing to the likes of John Ganz, Tim Snyder, Noah Smith, Gary Marcus, Brian Merchant, Carol Cadwalladr and several dozen more, all of whom send email notification of new posts, some daily. Found out how to list all my (free) subscriptions from that demonic UI – though have now forgotten again how I achieved it – then totted up how much it would cost to upgrade them all to paid, which took me a couple of hours and turned out to be £335 per month! I actually pay for just two (not saying which), put up with studium interruptum for the rest and regularly weed the notification emails that keep refilling my inbox like Ground Elder. 

This is not a sensible business model and the seekers of subs must eventually get culled either by Darwinian selection or by monopolistic consolidation. I notice that it’s 17 years since I last wrote here about a sensible solution, namely micropayments collected by internet providers and distributed via the tax system, but nowadays that would sound far too much like China’s social credit system. 

[Dick Pountain wouldn’t mind if you paid him £8/month on the side to read his column, but doesn’t want to push his luck] 

 

 

 

AUDIO DENTIST

Dick Pountain /Idealog 375/ 5th October 2025 : 09:53am 

I love music and I love soldering. Here is a story that has both. I’ve explained my opinions about hi-fi here several times: I like good sound quality but I’m not a hi-fi nut, don’t buy oxygen-free cables or onyx cartridges or gold-plated anything (and definitely didn’t colour-in the edges of my CDs with a green marker pen). I won’t spend thousands on any component, and I wrote here how my listening was transformed by connecting a cheap Fosi Class 4 Bluetooth amplifier to my vintage Castle speakers (legacy items from when Dennis Publishing used to publish Hi-Fi Choice magazine). 

All sound from my Chromebook Plus gets routed via Bluetooth through these speakers and I remain delighted by its quality, but a problem arose. I have to attend regular Zoom committee meetings, where we discovered that my Bluetooth remote arrangement was causing irritating echo effects for the other members, so I decided to resort to headphones. I purchased a set of moderately-priced JBL wireless headphones and that solved it, but it also created a slight annoyance for me because the dreaded Bluetooth wouldn’t automatically switch itself, so I had to manually disconnect the Fosi amp in Settings and select the JBLs.

Now with my other, sociological, hat on I’m a keen observer of the doings of GenZ youth, which recently include discovering a preference for vintage digital compact cameras over smartphones and for wired headphones over wireless ones. Aha I thought, in my Santa’s-Grotto-of-retired-digital-artefacts upstairs (covered in a recent column) there must be some of those, and there were indeed two – Sony MDR V100s and Sennheiser HD 201s that I hadn’t used for decades. I tried them, both worked and sounded surprisingly good, but the Sennheisers sounded terrific. So terrific I decided to do the pseudo-scientific thing, an A/B/C comparison with the Sony and Bluetooth JBLs. The result shocked me as I far preferred the Sennheisers. The JBLs were louder and had more bass, but somehow they, and the Sony, were less ‘engaging’. 

Engagement is difficult to discuss without descending into woo-woo. The Sennheisers are very light (plastic and aluminium pressings), oval in shape so they completely encase my ears rather than resting on them. Their sound-stage is better balanced, with bass that’s not so pronounced but ‘right’. (As an aside, I personally judge sound reproduction by just two instruments, acoustic piano and double bass: for my A/B/C test I used the tracks ‘Mademoiselle Mabry’ by Miles Davis and Janacek’s ‘On An Overgrown Path’ played by Josef Páleníček). The Sennheisers induced what we used to call a ‘drugless trip’, where you engage so far that you feel there.  

Problem was, those Sennheisers were made in 2005, the era of plastic-from-hell that turns into chewing gum with age. The outer insulation on their far-too-long cable was rotting and peeling, to reveal not the expected red or black plastic-covered conductors but a gleam of bare copper! I tried insulating tape, Gorilla tape, even silicon sealer, but ended up with an ugly, sticky mess. Wanting them so much emboldened me to replace the cable myself. Gemini taught me that those inner conductors are called ‘tinsel wire’, microscopically thin strips of copper coated with an insulating lacquer that you can’t just scrape off, spirally wound like guitar-strings around a central textile cord. Soldering them is a real challenge and reader, I took it! 

From Amazon I bought two short cables with female RCA plugs going to bare wire, along with a longer cable from 3.5mm jack to twin male RCA plugs. Snipped off the old cable, took out four tiny screws to disassemble each earpiece, only to find a single black-box holding all the gubbins from which emerge two horse-hair-thin tinsel wires lacquered red and blue. YouTube explained how you create of large blob of molten solder and plunge the wire into it: if the temperature is just right that burns off the lacquer (with a puff of smoke), adheres to and tins the copper; if the temperature is too high it burns off the copper too; too low and it doesn’t adhere. I chopped up the old cable to practice and practice, only suffering one painful burn, and then did the deed. 

Reassembling and finding they worked, I punched the air in triumph. Completed the job by discovering that my Chromebook’s two USB-C ports support high-def audio. Back to Amazon for a UGREEN adapter from USB-C to 3.5mm jack that contains a 32bit/384kHz DAC giving more volume and (perhaps) slightly better definition than the audio jack socket. The signal path from recording studio to my eardrums is long and convoluted, via TCP/IP, Wi-Fi and various different AC-DAC-and-backs, but at least I’ve managed to extract one irritating Bluetooth…

[Dick Pountain doesn’t have ‘golden ears’ but he does have dry earwax]

 

 

  

  

LOVE AND HAIDT

Dick Pountain /Idealog 374/ 3rd September 2025 : 10:36am 

It becomes harder and harder to scrabble grains of online pleasure, amusement or edification, but it remains possible (just) on YouTube. I particularly enjoy two grizzled performers, Rick Beato (whose interviews of musicians like Rick Rubin, Guthrie Trapp and Tom Bukovac are priceless) and Jon Stewart, satirical political commentator whose late night Daily Show kept many of us in stitches during the GW Bush presidency. Stewart attempted to retire in 2015 but he’s back, presumably lured by the grim shenanigans in the White House, presenting a new YT version of the Daily Show on Mondays, but on Thursdays an in-depth podcast called The Weekly Show, where a recent guest was the social psychologist Jonathan Haidt.

Haidt is currently controversial for his best-selling book that alleges that smartphone overuse is damaging young peoples’ mental health, but I know and admire his work from reviewing an earlier book called ‘The Righteous Mind’ (2012), an experimental study of the way peoples’ moral outlook affects their political behaviour. Haidt is a leading light of the ‘Intuitionist’ school of psychology which holds that not all our behaviour is rational, and in particular moral judgements like disgust are hard-wired to bypass the reasoning parts of the brain. (His thought-experiments to test this are highly amusing but unsuitable for a family magazine like this, involving incest and molesting chicken dinners). The reason I raise his work in this column is another book I just reviewed, Karen Hao’s ‘Empire Of AI’, an inside glimpse into the rise of OpenAI and ChatGPT. 

Hao documents three important facts about the company: unanimous agreement that AGI (Artificial General Intelligence) is possible and the only worthwhile goal; belief that AGI will be achieved by endless ‘scaling’, cramming tens, thousands, millions of Nvidia GPUs into their servers; and a split, right from the very start, between those who think AGI will be great and those who think it will be deadly (many of whom left to set up Anthropic and Claude). Now I’ve stated here several times that I believe AGI is neither desirable nor possible, for reasons that depend upon the work of Haidt among others. If it’s not achievable we won’t face the worst of many imaginary harms like enslavement by robots, but it makes the current monomaniacal hyperscaling futile, dangerous and horribly wasteful. 

Impressive, amusing and addictive as current LLMs and GPTs are, they fall far short of general intelligence because they’re not alive. Unlike Nvidia chips, living beings need food, safety and to reproduce themselves, and these imperatives structure our thought and behaviour profoundly. Billions of years of evolution equipped us with ‘emotions’, chemical computational sub-systems that detect and seek to satisfy needs. The US/Portuguese neuroscientist Antonio Damasio postulates that when we store a memory of an event it somehow gets imprinted 

with the emotional/hormonal state at the time (via biochemistry that is barely yet understood). When we retrieve it later to help understand some future event, these emotional markers act as weights (like parameters in an LLM) and contribute to the outcome of our decision. So images and words can never be entirely neutral, they carry subconscious emotional connotations of varying strengths. AI models lack needs and fragile bodies, and hence purpose. Actually smartphones, which can ‘see’ and ‘hear’, know their location and orientation and can travel the world in our pockets (so long as we remember to charge them) are way closer to human experience than ChatGPT is. Equipping one of those highly-capable Boston Dynamics robots with a fully autonomous AGI must remain science fiction so long as GPTs require aircraft-hangar-sized supercomputers and consume megawatts of electricity. Our own bodies have a mitochondrial ‘battery’ in every cell, enabling us to think and/or reproduce ourselves on around 2000 calories a day…

Cognitive psychologists and economists like Haidt and Kahneman have revealed that emotional modes of intuitive thought aren’t reducible either to symbolic logic or Turing computability, and that these mechanisms drive attraction and enmity, friendship or bias and prejudice. They underpin crucial affective human virtues like empathy, wisdom, justice, courage, honesty, compassion and generosity without which any aspiring AGI would merely be a sociopathic silicon solipsist. And most importantly, intuition is vital for creative reasoning, causing those unprecedented leaps between vastly differing conceptual spaces that make up the mind of a Newton, a Mendeleev or an Einstein. The training data for connectionist AI models contains only representations of mental states – text and pictures scraped from the internet – and what emotional weight it does aggregate is mostly bad news, a swamp of hateful and obscene human communication that costs the AI corporations big bucks to hire human beings to painstakingly disinfect, in procedures they call ‘alignment’ and RLHF (reinforcement learning from human feedback)...

[Dick Pountain likes ChatGPT, but in a purely platonic way]

 

COLLECTABLES

Dick Pountain /Idealog 373/ 08 Aug 2025 10:52  

I don’t really have the collectors’ instinct. When I was a kid my father was a serious stamp collector and I briefly made a feeble effort to be one too. I was slightly more interested in my album of labels from exotic canned goods, but that petered out pretty soon too. I find that in adulthood I’ve accumulated nine guitars but each of those was bought to play, then superseded but not sold, so it doesn’t really count as a collection. I have owned ten motorcycles over sixty or so years, but only ever one at a time. Books don’t count: I started accumulating those as a student and continued as a book reviewer, but all were obtained to read and never sold (there are around a thousand of them, none rare).

When I’m not writing about computers here I review books for a political journal, mainly ones about political economy and sociology. An author who had a big influence on me was the French sociologist Pierre Bourdieu, whose best-known book ‘Distinction: A Social Critique of the Judgment of Taste’ examined taste as an act of social status building, drawing on huge amounts of data gleaned from quantitative surveys, photographs and interviews. Two former associates of his, Luc Boltanski and Arnaud Esquerre in 2014 published a paper called ‘The Economic Life of Things: Commodities, Collectibles, Assets’ which is the best, most interesting account of collectability I’ve seen. B&E describe the way that all the things we manufacture, purchase and use pass through three phases of ownership, which they label the ‘standard form’, the ‘collectible form’ and the ‘asset form’. Consider for example a lemonade bottle from before WW1 with one of those stoppers that’s a caged glass marble. A mass-produced item, made as cheaply as possible, reusable and returnable for a penny. It served its purpose of containing and dispensing lemonade, then eventually got thrown in the bin (standard form). Years later someone found it on a tip and it ended up in an antique shop, sold for £15 to a middle-class couple as a kitchen ornament (collectible form). Their friend was a famous film director and it became a key prop in a very successful movie, and ended up sold for £5000 in a sale of memorabilia at Christie’s (asset form). An ancient Roman olive-oil jar might follow a comparable trajectory, but with prices several orders of magnitude higher. A drawing done on a napkin to pay for supper by a famous painter ditto, but its asset form might be in millions. In the asset form, things are no longer used, may often not even be displayed but stored in a vault, a hedge against inflation or financial crisis, a store of value.

What has this to do with computers you may be wondering. Well nothing actually, and that’s the point. Computers, along with much of the rest of the merchandise of the digital world, seem to defy B&E’s classification scheme by being stuck forever in the standard form: they end up in a skip, then get ripped apart to recycle a few chips and some gold-plating. The aesthetic appearance and quality of workmanship of such goods is so low that very, very few people want to collect them, and what’s more those few who do face insurmountable problems in keeping them 

working, due to the rapid and haphazard evolution of firmware, software, ports and cables, storage media, the lack of effective documentation, and the rapid disappearance of smaller manufacturers prior to the monopoly era we now inhabit. 

If you detect a faint tinge of animus in that last paragraph you’re correct, it’s because I have a room upstairs full of digital junk accumulated over my 40+ years of computer journalism that I can’t get rid of (and which those nearest and dearest to me would love to see dumped in a skip to reclaim the room). I can’t bring myself to do that. Along with some quite notable historic hardware – first-gen IBM PC and Macintosh, Acorn Archimedes, Newbrain, Epson HX20 – there are shelves full of software both famous and obscure that I have a hunch I may perhaps be the only person to still have, given my privileged status as recipient of review materials. The early history of the UK personal computer scene is sitting up there, and no-one appears to want it. I’ve tried all the various computer museums people have recommended, and none are interested in collecting the lot (I no longer have a car). All the reasons I mentioned above render it enormously hard for them to get this stuff working, and once they do it’s hardly entertaining. I do feel though that someone ought to document this history before skips claim it all. 

[Dick Pountain will hold onto the draw containing every Psion Organiser]

 

TOO DARNED HOT

Dick Pountain /Idealog 372/ 07 Jul 2025 01:15

I’ve been watching the rebellious mood that’s growing among Microsoft Windows 11 users with a degree of (not very nice) complacent amusement, as someone who dumped Windows in favour of a Chromebook more than eight years ago. Actually my defection was as much an accident as an example of prescient wisdom. When Dennis Publishing moved to new offices, then CEO James Tye took the quixotic decision to deploy Chromebooks to all, I took the opportunistic decision to borrow one and, being totally brassed-off with Windows 8.1, immediately became hooked. The Asus I bought for myself still works well and had been a source of great pleasure but for one problem – Google stopped supporting its version of ChromeOS with automatic updates about a year ago, and I started encountering apps that demanded an OS update I couldn’t procure. So I splashed out £229.99 on a new Asus CX3402 Chromebook Plus, which has a better screen, twice the memory, an 8-core Intel CPU and 10 years guaranteed updates. Migrating to the new machine was as arduous as usual: charge the battery, switch it on and wait 10 minutes for my online life to come down from the cloud. There were two extra chores though. Because in all matters digital I’m very far from being a trusting person, I also keep data I consider crucial on a local 128Gb memory stick that lives permanently in one USB port, so I had to unplug that and plug it into the new machine. Another quirk of mine is that I don’t really like touchpad cursor control and so use a Logitech Wireless Mouse whose dongle lives in another USB port.

I periodically backup the contents of the USB stick, a tiny metal-cased one from Integral, for which purpose I swap the Logitech dongle for another backup memory stick. A few weeks later I was doing such a backup when I noticed that the sticks had become very hot. Not warm but hot, hot enough to make me flinch on touching them, hot enough to worry. So I went online to the source of all wisdom which is Reddit, where I discovered that hundreds of people were reporting the same experience, not only with Chromebooks and with a variety of brands of stick, but in all cases when using them in USB-C ports. The consensus was that it’s OK, it’s because of the huge capacities of the current generation of sticks and the poor ventilation of the smaller cases. I pretended to believe that and to live with it for a few more months, until I started to get unreliable behaviour from the toasting stick. First it started unmounting at random times, though it always came back after unplugging and replacing. Then during a backup session, ‘copy failed’ messages and directories going missing from listings, at which point I panicked. I dug out my original old Asus where I confirmed that the contents of the stick were in fact intact, and that it didn’t get hot, and I carried out the backup there on the cool older USB ports. 

Something clearly had to be done because it’s become part of my work practice to keep this tiny, unknockoutable, USB stick permanently in place, and changing to some huge protruding one wasn’t acceptable. My first recourse was a heat-sink, cobbled together by wrapping the Integral stick tightly in aluminium cooking foil held in place with a 15mm binder clip with the handles detached. This worked, dissipating enough heat to reduce it to just warm to the touch, but it was too inelegant for me to live with. I couldn’t find any hard info online about which brands were most liable to overheat, but my own collection accumulated over the years revealed Sandisk, Patriot and Tab all got just as hot. As I was dolefully scrolling down the endless Amazon list of sticks, one from Samsung caught my eye because it looked nicer: the same shade of grey as my computer, short, fat and shiny. I ordered one and discovered that it’s just as fast, and barely gets warm… 

What moral to draw from this story I’m not really sure. I’m not a semiconductor engineer and can’t find an adequate explanation online from anyone who is, as to why such a huge discrepancy in performance exists between brands. Does the difference lie in the chips themselves, the design of the cases, the electrical interfaces or a problem in USB-C sockets? Such silence is deafening and disturbing when data loss is a distinct possibility. But of course the computer is becoming very much the poor cousin to the smartphone, for which such sticks are not relevant and for which professional standards of data hygiene are barely relevant either.    

[Dick Pountain still occasionally dreams that he’s trapped inside the Windows Registry]

COLOR ME OLO

Dick Pountain /Idealog 371/ 10 Jun 2025 12:15

I’ve expressed my feelings about science fiction before many, many, probably too many, times in this column. A big fan in my 1960s teens, a bout of illness in the ‘70s let me binge all the greats – Vonnegut, Ballard, LeGuin, Dick, Pohl, Bester etc – overdosing so badly that I never wanted to read sci-fi again. Looking back now as emotionally-retarded pseudo-intellectual sci-fi fans appear to be taking over the world, I think perhaps my immune system was telling me something. However this allergic reaction doesn’t apply to the closely related genre of fantasy (or gothic, or cosmic) horror. I still can cringe a little to M.P. Shiel’s ‘The Purple Cloud’, William Hope Hodgson’s ‘The House on the Borderland’, or the entire oeuvre of H.P, Lovecraft. 

One of Lovecraft’s stories, ‘The Colour Out Of Space’, struck me particularly hard. A meteorite lands in a tiny community in the New England woods, containing globules of a weird colour that isn’t in the solar spectrum: it has unpleasant effects that consume all living animals, plants and humans and turn them into grey ash. Several unsuccessful attempts have been made to film this story, hard work given that it’s not in the technicolor spectrum either: perhaps the nearest anyone has come is Alex Garland’s 2018  ‘Annihilation’ which clearly shows Lovecraftian influence and employs a digitally-produced shimmer in place of a new colour. I suppose Lovecraft’s story is a sort of parable about environmental destruction, but that’s not what explains its hold on me – I’ve always been fascinated by colour, studying its chemistry, reading up on all the various systems, appreciating great paintings and creating my own digital art as a favourite hobby.   

All of which explains my enormous excitement on reading in the Science Adviser newsletter about new research at the University of California Berkeley, which ‘creates’ a new colour that lies outside the gamut humans can perceive, but can be seen by shining a laser onto ones’ retina (don’t try this at home kids). First some background is required. You already know that our eyes, and hence also digital display devices, perceive or present colours as mixtures of the three components red (R), green (G) and blue (B). That’s because the retinas of our eyes contain three types of colour-sensitive cell called ‘cones’, sensitive to different wavelengths: long (L), medium (M) and short (S). Objects illuminated by natural sunlight stimulate L, M and S cones to different extents which gives us the experience of different colours. Red light primarily stimulates L cones and blue light S cones, but since M cones respond to the middle of the range, overlapping with both L and S, there’s no component of sunlight that stimulates M alone. 

The authors of the Berkeley paper (https://www.science.org/doi/10.1126/sciadv.adu1052) set out to investigate a new system for describing colour perception, by using a laser to illuminate retinal cells one at a time: 

“We introduce a principle, Oz, for displaying color imagery: directly controlling the human eye’s photoreceptor activity via cell-by-cell light delivery. Theoretically, novel colors are possible through bypassing the constraints set by the cone spectral sensitivities and activating M cone cells exclusively. In practice, we confirm a partial expansion of colorspace toward that theoretical ideal. Attempting to activate M cones exclusively is shown to elicit a color beyond the natural human gamut.”

Calling their new system Oz was of course a trigger for me, having cut my journalistic teeth on that notorious hippie journal where my articles were printed in every colour of the rainbow on a background of every other colour of the rainbow. But I digress. Their new colour, from stimulating M cells alone with a laser, they named ‘olo’. It can’t be reproduced in paint, ink or screen, so you’ll only ever see it sitting in a dentists’ chair with a laser strapped to your head: 

“Subjects report that olo in our prototype system appears blue-green of unprecedented saturation, when viewed relative to a neutral gray background. Subjects find that they must desaturate olo by adding white light before they can achieve a color match with the closest monochromatic light.”

Thankfully olo has not so far shown any inclination to suck the life force out of living beings and reduce them to grey ash. The best approximation is a light turquoise, a colour you might glimpse fleetingly when watching a big wave break on a rocky headland in bright sunshine. Astronomers for a while believed the whole universe, by mixing all the light together, would be a light turquoise but that turned out to be a bug in their software, and it’s now ‘cosmic latte’, a light beige (hex triplet value in standard sRGB #FFF8E7). 

[Dick Pountain wonders whether Olo could become ‘the new Pistachio’]

GOING NEURO

Dick Pountain /Idealog 370/ 05 May 2025 01:31

I’ve written many, many sceptical words about AI in this column over the years, railing against overconfidence and hype, hubristic pursuit of AGI, deepfakery and content pillage, but nevertheless I do believe AI – once we’ve civilised it – is going to be hugely  important to science, economics, robotics, control systems, transport and everyday life itself. Given the political will, public concern about misinformation, invasion of privacy, and theft of artistic data can be regulated away, but there would remain one colossal stumbling block, namely energy consumption. 

When AI corporations consider purchasing mothballed nuclear reactors to power their compute-servers the absurdity of AI’s current direction ought to be visible to everyone. The current generation of GPT-based AI systems depend on supercomputers that can execute quintillions of simple tensor arithmetic operations per second to compare and combine multiple layers of vast matrices holding encoded parameters. Currently all this grunt is supplied using the same CMOS semiconductor process technologies that gave us the personal computer, the smartphone and especially the computer game – the Nvidia chips that drive most AI servers are descendants of ones originally developed for rendering real-time 3D games. The latest state-of-the-art GPUs have a watts/cm² power density around the same as an electric cooking hob, and the power consumption of AI server farms scales exponentially, as the square of the number employed (order O(N²) in the jargon of complexity theory). 

In their 1978 bible of the CMOS revolution ‘Introduction to VLSI Systems’, Mead and Conway devoted a final chapter to the thermodynamics of computation: we’ve long known that logic operations and memory accesses always consume energy, whether in silicon or in protein-and-salt-water like the human brain. However the human brain has far, far more neurons and synapses than even the largest current AI server farms have GPUs, yet consumes around 20 Watts as opposed to AI’s 50+ Megawatts. Understanding what’s responsible for this immense efficiency gap is crucial for creating a more sustainable next generation of AI, and the answer may lie in new architectures called ‘neuromorphic’ because they mimic biological neurons.

Individual metal-on-silicon-oxide transistors aren’t six orders of magnitude more power-hungry than biological neurons, so other factors must be responsible for the huge difference. One factor is that biological neurons are analog rather than digital, and another is that they act upon data in the same place that they store it. In contrast the CMOS GPUs in AI servers are examples of von Neumann architecture, with processing logic separated from memory, and program code from data. But the MOSFET transistors they’re made from are inherently analog, operated by varying voltages and currents, so the digital data they manipulate gets continuously converted back and forth between the domains, at great energy cost. 

Neuromorphic AI hardware designers try to bring data and processing closer together. Intel introduced its Loihi 2 research chip, with 128 neuromorphic cores and 33MB of on-chip SRAM, which communicates via trains of asynchronous voltage ‘spikes’ like those in biological neurons. Steve Furber (of ARM fame) works at Manchester University on a neuromorphic system called Spinnaker that has tens of thousands of nodes each with 18 ARM cores and memory, also using spike-based communication. These schemes do reduce data access overhead, but they remain digital devices, and to approach biological levels of energy economy will require a still more radical step into purely analog computation that exploits the physics of the chip material itself. 

The US firm Mythic’s AMP (Analog Matrix Processor) chip employs a 2D grid of tunable resistors whose values encode the weights for an AI model, whereupon it relies on Kirchoff’s Laws to in effect multiply-and-add the analog input voltages and perform convolutions. However AMP is still fabricated in CMOS. A more radical next step would be to implement this resistive analog computation using low-power ‘spintronic’ memristors – devices in which the orientation of magnetic spins represent bits as in modern hard disks. One way to implement non-volatile memristors is by FTJs (Ferroelectric Tunneling Junctions) formed by sandwiching nano-thin magnet/insulator/magnet layers which can be fabricated using existing semiconductor processing. These devices can be written to and switched cumulatively like real neurons, and read-out non-destructively using very little power. 

The Dutch physicist Johan Mentink used a recent Royal Institution lecture (https://youtu.be/VTKcsNrqdqA?si=ZRdxeyP4B-hfUw3X) to announce neuromorphic computing experiments in Holland that employ two-dimensional cross-bar grids of memristors, organised into a network of  ‘Stochastic Ising Machines’ that propagates waves of asynchronous random noise whose interference yields the spike trains that transmit information. The Dutch researchers claim such devices can potentially be scaled linearly with the number of synaptic connections, reducing power consumption by factors of 1000s. I love the idea of working with rather than against noise, which certainly feels like what our brains might be doing… 

[Dick Pountain’s neurons are more spiky than most]

INTERESTING TIMES?

Dick Pountain /Idealog 369/ 07 Apr 2025 12:06 

Some 26 years ago (May 1999/ Idealog 58) I opened this column thus: “It's said there's an ancient Chinese curse, "May you live in interesting times!" Actually it's been said by me at least twice before….”  

Well, it looks time to trot it out for the fourth time, though nowadays I can check its truth using my old friend ChatGPT: “No record of this phrase (or anything close to it) exists in classical Chinese Literature: the earliest known attribution is from British sources in the 20th Century, a 1936 speech by Sir Austen Chamberlain (brother of Neville) who claimed to have heard it as a Chinese curse but it seems more like a Western invention made to sound exotic or wise.”  So, I’ve been peddling fake news for a quarter of a century, but I don’t feel too guilty as everyone seems to be doing it. 

The financial turmoil created by President Trump’s tariff barrage purports to be against a whole world ripping off the USA through unfair trade, but we all know it’s really about China. Many of us believe Trump has made a monumental blunder that will ultimately help China to economic dominance: The Wall Street Journal expressed it thus “What a fabulous change in fortunes for the Chinese leader. Mr. Trump has taken an ax to the economic cords that were binding the rest of the world into an economic and strategic bloc to rival Beijing – and at precisely the moment many countries finally were starting to re-evaluate their economic relationships with China.” 

I’ve covered semiconductor fab, Taiwan invasions and DeepSeek in recent columns so let’s not go there again, except to guess that Trump’s shenanigans could cause interest rate, futures and bond market chaos that may bring down the intricate house-of-cards finance of the AI bubble corporations (already under siege from IP lawyers). Instead I’d rather talk about how the ‘interesting’ times are affecting my own everyday activities. 

For many years my online presence, apart from my own website, consisted of Flickr for posting photos and Facebook for chatting/arguing/posing with friends. No longer, as my online self has been shattered into a dozen fragments, none of which have quite the same scope or satisfaction as before. Facebook started to deteriorate for me a couple of years ago as old friends left and un-asked-for content increased, but since Zuck did a Musk on it by removing moderation it’s becoming intolerable, and as a result I’ve begun to work on building a following on both BlueSky and SubStack. 

BlueSky is full of left-leaning refugees from the steaming pit that X has become: lots of excellent, sympa content, in fact too much to read it all and unanimous enough to risk boredom. I joined SubStack years ago hoping to get paid for some of my stuff but that didn’t work out so I forgot it until now, when it appears to be changing into something different. It’s becoming a social platform to rival Facebook, an alternative refuge for X-iters and I actually find it more interesting than BlueSky, but with one huge reservation - its structure and user interface remain totally baffling. Is it a mailing list or a website, a forum or… what?  Do I add posts or notes, and where will the comments arrive? My efforts in computer-generated music are now scattered among a host of platforms including SoundCloud, YouTube, NoisePit, BandCamp and GrungeDump (I may have invented one of those) and it remains stubbornly antiviral on all of them. YouTube is still my main source of entertainment, from genial luthiers to hilarious espresso gurus, Rick Beato’s music interviews to Jon Stewart’s Weekly Show. I even watch some movies there cheaper than other paid platforms (recently found ‘Lonesome Dove’ for free). 

The ‘interestingness’ seems to be spreading from online matters to offline. In recent months my Chromebook finally ran out of support (bought a new Asus CX340, cheap, way faster and nicer). BT announced that it was killing off my analog landline early, meaning a new hub, and that my mobile account should be moved to EE. While trying to surf such unwelcome disruptions several websites started playing up – I became adept at navigating poorly-implemented two-factor authentication schemes that trap you into endless loops of passcode tennis, and discovered a new game called ‘hunt the human’ while traversing the maze of AI chatbots that firms now erect in the name of Help…

Shall I end on a cheerful note, that things can only get better? It’s getting ever harder to believe that. Once the DOGE-days are over, assuming some kind of sanity is restored, then the craven way the big Silicon Valley corporations crowded onto Trump’s rattling gravy-train will haunt and taint them for years to come. 

[ Dick Pountain pronounces DOGE as ‘doggie’, like that creepy Shiba Inu dog meme] 

 

KIND OF BLUE(TOOTH)

Dick Pountain /Idealog 368/ 05 Mar 2025 04:38

I’m literally a one-man-band, by which I mean that I make music using an assembly of electro-acoustic gadgets that permit me to do without the collaboration of other human musicians. (I hasten to add that those gadgets do not include drums or cymbals strapped to my legs, like the blokes who used to entertain cinema and theatre queues). 

I love music, from a wide range of genres, and I’m picky about quality reproduction for other peoples’ music I enjoy. I’m far from being a hi-fi nut (and indeed quite sceptical of the excesses they indulge in) but I do run a nice-sounding system based around a Fosi Class D amplifier connected to vintage British speakers, which sucks in music via wire from a vinyl turntable or CD player, and via Bluetooth from my Chromebook, Samsung tablet and smartphone. (I can also listen to those via Bluetooth headphones and earbuds). It would be handy to incorporate the sound from my Panasonic smart TV via Bluetooth too, rather than SPDIF, but the brute refused to pair with my amp so I bought a tiny cheap Bluetooth sender/receiver from Amazon and plugged that into the set’s headphone socket. It worked fine but the cursed TV then hogged the Bluetooth and had to be unpaired before any other source could use  it, so I’ve learned to live with its (actually quite acceptable) Dolby sounds, and so now had a Bluetooth dongle to spare.

I tried plugging this into my various electric guitars via an adapter, to play them wirelessly: it works but is totally unusable due to latency. Meanwhile in another corner of my music room stands the very analog, one-man-band conglomeration through which I play those guitars: two acoustic, two electric, one bass switched into a small Marshall amplifier that’s encrusted with five effects pedals and a small Zoom drum machine. This is all connected via standard ¼” jack cables, and It’s only taken barely 15 years to arrive at a satisfactory topology, the brainstem of which is a 3-channel passive audio mixer, barely bigger than a Yorkie Bar, which the late-lamented Maplin emporia used to flog for £20 (I bought two). One channel is taken up by the drum machine, a second one by the multi-effects box which the other instruments go through, so one channel remained empty and it was simply irresistible to plug in the Bluetooth dongle. Paired with my Samsung tablet, a whole new world opened up... 

Among the pedals is an Akai Headrush echo/looper I bought 20 years ago after seeing KT Tunstall play one on the telly – this lets me store and replay short clips of music and overdub them with more layers. By logging on to Spotify on the Samsung I can play any tune and store a chunk of it into the Headrush to play over and add to – great fun with slices of Bill Frisell or Julian Lage, or Ron Carter bass riffs. I also have Volcanic Mobile’s MIDI Sequencer for creating backing clips using its piano or brass sounds, and also my own Algorhythmics programmatic system (which I’ve described here before). I can compose parts using BandLab, a popular free Android DAW, with sampled instruments other than guitars. In fact any software that can make a noise on my tablet, phone or Chromebook can now be routed into my guitar amp and chunks saved and looped to play over – which even includes my various voice synths and text-to-speech readers, or mic recordings.

However what I really wanted was to create original soundscapes for use as backdrops to improvise over, an ability that I discovered quite by chance in Pegboard (https://semitune.com/pegboard/). Its publishers Semitune describe this app as “an advanced mobile polyphonic wavetable synthesizer with a virtual analog filter, 12 standard modules and 6 effects modules.” It’s driven by two separate wavetable oscillators that you can customise via a graphic interface in which you drag envelope shapes and make them evolve over time in complex ways for really rich sounds. Play them back via an onscreen keyboard, either piano-style or an array of accordion-style touch buttons whose layout you can alter to represent a chosen scale or mode. Pegboard isn’t a sequencer so you can’t save whole tunes, just the sounds themselves, but that’s perfect for my purpose, to play short phrases into my looper (it can also work as a MIDI controller, a feature I’ve yet to master).Pegboard is free to play with but you must upgrade to Pro for £18.99 to save sounds, which I very quickly did. Think of my guitar as a pen and Pegboard sounds as washes I draw over to make sound pictures for my one-man band’s audience of one…

[You can see Dick Pountain’s one-man-band rig at https://www.facebook.com/share/v/15xD1dS1UD/ ]

 

  

DEEPCHEEK?

Dick Pountain /Idealog 367/ 06 Feb 2025 09:12

I used to think that a monthly column was a fairly relaxed schedule compared to, say, a daily newspaper, but no longer. I’d decided to do this one about how China upset the USA by doing AI on the cheap, but now every ten minutes I feel a need to check online for whatever new geopolitical atrocity has just overshadowed that. Nevertheless I’ll start with a nod to the original plan, how China pulled down the knickers of the US AI bubbleheads.  

I won’t dive deep into tech details of how Deepseek succeeded in doing what ChatGPT does for a fraction of the price, or how it rocketed to the top of Apple’s mobile-app store hit parade, nor how it did so by parasitising the US AI bros’ own data in just the same morally and legally unsavoury way they got it from us in the first place. No, instead I’d prefer to harp on about something I’ve been harping on about for at least 20 years, namely how the whole AI industry deludes itself because, being lead by sci-fi-addled nerds (one of whom now appears to be the de facto POTUS), it has a severely limited grasp of biology and philosophy.

Two columns ago I forcefully expressed my opinion of OpenAI’s plans for continued expansion in order to achieve AGI (Artificial General Intelligence), which they claim would confer human-level reasoning. One objection was its colossal, antisocial, power requirements, but my real objection is that I don’t believe AGI is even achievable by simply crunching more data. That’s for reasons of biology I’ve explored here many times, namely that though human intelligence expresses itself through language – by manipulating symbols which is all any computer can do – that’s neither its only nor its most important source. 

We’re animals who have been equipped by evolution to succeed at living in a physical world, achieved with the help of many (more than five) senses to sample what’s going on around us. We build, continuously update and maintain a mental model of that world. We have needs – including to eat, drink, reproduce and avoid predators – which are intimately entwined into that model. We’re born with some built-in knowledge about gravity, upness and downness, light and shade, convexity and concavity, that control the model in ways of which we’re not conscious, but which deeply affect our symbolic processing of that world. We’re by no means just ‘rational’. AI has learned how to pretend to be intelligent only by plundering our symbolic representations of the world, texts and pictures, but knows nothing, and can know nothing, of our embodied experience. Sure, it could build imitations of emotions and needs, but they’d just be more static representations, not extracted from the real world by living in it.  

Historically computers arrived thanks to advances first in mathematics, and then electronic engineering, so it’s hardly surprising that the intellectual atmosphere in which they’re embedded is more influenced by science fiction than by philosophy, anthropology or cognitive psychology. It may well be too late now for AI practitioners to go back and do the necessary reading, since they’ve reached a level of megalomania that convinces them they already know it all, and have just achieved power over the world’s most powerful nation to prove it. 

Were I to be asked to set the tech bros some homework I’d recommend first of all my favourite philosopher George Santayana and his theory of ‘animal faith’, which enables us to navigate life’s uncertainties by deploying our intrinsic knowledge and not ‘overthinking’. That leads directly into the more modern version by Nobel Laureate Daniel Kahneman and his discovery of two modes of thought, the fast, imprecise-but-often-good-enough one, and the slow symbolic one which is all that computers can mimic. Then I’d suggest perhaps George Lakoff and Mark Johnson’s ‘Metaphors We Live By’ which explores the actual content of our intrinsic embodied knowledge and how it modulates our language. Oddly enough, smartphones are already more embodied than GPTs because they have senses (hearing, sight, touch, spatial orientation) and hunger (for charging). Fortunately they can’t yet reproduce.

Having absorbed all that, then perhaps they might dip into Chris Frith’s ‘Making Up The Mind’, the best explanation I’ve read of how the brain creates and updates the mind using fundamentally Bayesian rather than Cartesian mechanisms. That ought to convince them they don’t (or most likely don’t want to) know it all, but the final step would be by far the most difficult, to get them to take democratic politics seriously and divert their megalomaniac schemes toward improving life for the majority of the population rather than a feckless techno-elitist minority. Of course they may prefer to go to Mars, which would provide a rigorous education in embodiment…

[Dick Pountain hopes that Elon doesn’t read his column]

POD PEOPLE

Dick Pountain /Idealog 366/ 05 Jan 2025 03:05

It’s January, when columnists feel obliged to reflect on the past year and who am I to refuse, though I’ll try to be different by not saying that 2024 was the year of AI. Instead I’m going to say that for me it was the year of the podcast. That’s partly because I got exposed to AI rather early via Stable Diffusion in 2022, and was bored stiff by the end of 2023. But it’s also because online services that had kept me amused for years, like FaceBook and YouTube, started sliding down a sloppily slippery slope into irrelevance during 2024. Feeds filled up with unwanted sponsored guff and AI-generated fluff, real friends abandoned platforms to be replaced by reels and clickbait that spread like digital cockroaches. In response I began to view more podcasts. 

Just as reels were shrinking down to 30 seconds of inane pointlessness, podcasts started expanding into 3-hour epics. Of course our own excellent PC Pro podcasts, crafted by Barry, Tim, Jon, Lee and Rois, adopt a manageable one hour format, probably the optimum length for normal attention spans, but several other podcasts I consume started at that length then got carried away. Back in 2023 an old friend recommended an article about The Velvet Underground, of special interest to me as the first piece I ever had published was about my experience of working at Max’s in New York in 1970 while they were the house band. This piece was on a podcast called “A History Of Rock Music In 500 Songs” by Andrew Hickey, and it was three hours long… 

Rather to my surprise I listened to all of it and was riveted: Hickey’s taste, depth of research, even his bluff Mancunian accent kept me enthralled. This episode on “White Light/White Heat” was only number 164 of the 500 he plans, in chronological order, but I was hooked and started listening from the beginning – number 1 was on Benny Goodman Sextet’s 1939 “Flying Home”, the first record with electric guitar, played by Charlie Christian. Andrew’s early episodes ran around 30 minutes, soon zoomed past the hour and now are regularly split into two or more parts – as for example The Beatles and Rolling Stones – reaching three hours plus. Thanks to his immense research efforts they remain quite engrossing. He’s now at episode 177 and intends to finish with a song from 1999 (which may take another 25 years at his current delivery rate). 

Another mega-podcast I’ve listened to all through is Paul Cooper’s superb “Fall Of Civilisations” about the rise and fall of empires throughout human history. He has an advantage over Andrew Hickey in that they’re fewer of them, mostly long in the past, and he’s covered most of them in 19 episodes. While not an academic historian, Cooper like Hickey has invested huge research effort and is an excellent presenter, making every episode informative and exciting without resorting to sensationalism. Some online niggling about historical accuracy is only to be expected, but his interpretations are largely convincing, not grossly ideological biassed, and the video version of the podcast (free on YouTube) illustrates his arguments with a well-curated montage of photographic, film and literary evidence on par with the work of Adam Curtis. Turns out that my favourite dead empires were the Nabataean and the Pagan.

Cooper’s series, available in both audio and video, raises the question of when is a podcast actually a vlog, but I don’t much care. Among my favourites is a series of 80+ YouTube interviews with living musicians by the veteran jazz guitarist and producer Rick Beato, which is probably neither or both but his interview with Rick Rubin is priceless.  

Have I ever podcasted myself? Only once because I don’t much like the sound of my own voice. It happened this way: in 1990 my brother-in-law Pip Hills and I took a road trip to Prague in his 1937 Lagonda saloon to witness Václav Havel’s inauguration as president of the Czech Republic. Following this trip another friend, Mark Williams, commissioned us to write about it for his magazine The Classic Motoring Review and subsequently the Scotch Malt Whisky Society, which Pip had founded in 1983, asked to reprint our article in their magazine and accompany it with a podcast. I charily agreed, and since I don’t possess a professional-grade microphone let alone a studio, performed my part over my Chromebook’s mic, using an audio editor called Lexis (my Android replacement for the wonderful Audacity with which I had 20 years of experience). I managed a usable take after two attempts, even including a snatch of music by Smetana at a pivotal point. Judge for yourself from the link below whether a career in voice-overs beckons…

[Dick Pountain’s Prague trip podcast is at https://unfiltered.smws.com/unfiltered-01-2024/smws-adventures-prague]

TRUMP OF DOOM?

Dick Pountain /Idealog 365/ 09 Dec 2024 10:48

I’ve been writing this column for over 30 years, during most of which I’ve deliberately tried to keep my political opinions out of it, apart from the occasional nod and wink about my lack of faith in free-market dogmas. However there are, very occasionally, world-historic events of such importance that to avoid mentioning them would be a sign of ignorance and cowardice. The last such event was the destruction of the World Trade Centre on the 11th of September 2001, and I did permit myself a column on that. Well, to me the re-election of Donald Trump on 6th November 2024 is another such event. 

I have of course been commenting on the rising power of Silicon Valley moguls – corporations like Microsoft, Apple, Amazon, Facebook, eBay, HP, Twitter and more – who built the industry whose products we document in this magazine. And during those whole 30 years I was writing under an unspoken assumption that these moguls, having emerged from the post-1960s counterculture, were fundamentally inclined toward ‘liberal’ (in the American sense) values. The two Steves Jobs and Wozniac were once ‘blue box’ phone phreaks, Google was started by two Stanford students in a friend’s garage under the motto ‘Don’t Be Evil’. And in the interest of full disclosure, PC Pro itself was created by a company founded by Felix Dennis, once editor of Oz magazine on which I too worked. 

In order to remain neutral, over the last couple of years I’ve refrained from expressing alarm as it became clearer that my assumption was being overturned. It started to look really shaky when Elon Musk, owner of Tesla, bought Twitter and proceeded to corrupt it from a vital news conduit for journalists of all persuasions into X, a conduit for previously-banned hate speech and pro-Trump propaganda. Then a week or so before the November election Musk came out for Trump and appeared prancing on platforms with him. Meanwhile Jeff Bezos, owner of the Washington Post as well as Amazon, forbade its editors to endorse any candidate, while Mark Zuckerberg announced he’d made a “20-year mistake” and “political miscalculation” (coded language for dumping the Dems). 

What has induced such a hand-brake turn in these billionaires’ opinions? A stock price rally following Trump’s victory increased their collective fortunes by $64 billion overnight but that’s merely chump change: Musk spent $250 million to finance Trump’s election campaign, a sum he earns every 15 minutes. Trump is promising to oppose internet regulation and prosecute journalists who investigate or criticise too much, but I think even those aren’t sufficient bait. These moguls already had everything except power to rule, which is now on offer.

The other promise Trump makes is to dump the Democrats’ (already feeble) policies toward climate change mitigation, turning the USA away from the Paris Agreement and Net Zero. This very well suits a second generation of moguls – the AI barons. My own attitude to AI has changed somewhat over the last few years. I’ve been sceptical of earlier claims that silicon tech will soon produce intelligence equivalent to or greater than humans, a goal now renamed a AGI, but I’m enormously impressed by the strides made in language and perceptual processing (I did after all let ChatGPT write a guest column for me). 

Three AI problems are rapidly becoming visible. The first is that those who really know (as opposed to simply hyping a stock-price bubble) are  as sceptical as I am about whether merely adding more GPU and training data will push GPTs across into AGI: there are already signs of plateauing or even degeneration through data pollution. A second problem is the absurd, even obscene, amount of electrical power consumed by the huge processing arrays that support the current generative AI models. Pronouncements from OpenAI about their future energy needs are beginning to sound frankly deranged – restart old nuclear power stations to marginally improve AI services which are, let’s face it, really only souped up search engines rather than solutions to any physical-world problems. Building a new clean energy infrastructure to mangle words and bitmaps rather than provide clean transport, heating and air-conditioning is actually psychotic. 

The third problem is that if Trump humours his new silicon buddies by employing their current, flawed, AI products to displace huge numbers of human jobs, he’ll likely trigger an economic crisis that leads to social unrest or even breakdown. This magazine is called PC Pro, the first P standing for Personal. We grew out of a 1980s technical revolution that put computing power into the hands of individuals and decentralised power away from the mainframes of state bureaucracies. The ambitions of the AI brigade concentrate processing back into gargantuan data centres that threaten data democracy itself.  

[Dick Pountain is busy gathering followers on BlueSky (@dick-pountain) as an act of Xtermination]

IN PRAISE OF PDF

Dick Pountain /Idealog 364/ 07 Nov 2024 12:58

Besides all that arty stuff I wrote about in last month’s column, I also review books. Not novels, self-help or pop science books but rather heavier subjects like political economy, evolutionary and social psychology. Some of these tomes get big (I mentioned a 1000-page job back in column 315) and note-taking for such a behemoth is hard work. That’s why I was delighted to discover the notation abilities in Kindle ebook editions, which I’ve also mentioned in previous columns. This ability to bookmark pages, search for keywords, add comments and cut-and-paste notes and quotes has become essential to my way of working. 

Regrettably or otherwise, Amazon’s dream that Kindle editions could take over the book business hasn’t materialised, with sales stagnating or declining and, more importantly for me, few of the academic publishers I deal with still producing Kindle editions. However they do all produce PDF versions, used internally for proofreading and so on. Hence when soliciting a review copy I now always ask for a PDF too.

The facilities provided in a PDF are sometimes sparser than those in Kindle Reader, depending on the settings used by the publishers and designers when outputting them, and so the capabilities of PDF readers have become a vital issue to me. I started out as everyone does with Adobe Reader, but found it less and less satisfactory over the years and not just because of Adobe’s grabby pricing structure – it’s become way too slow to navigate and search my biggest books. Also in column 315, I described discovering a far better PDF viewer, the Chrome extension PDF.js which is a free GitHub project built with HTML5. I’d been using this happily for several years until recently a Google change to ChromeOS stopped it working, and my Chromebook is too old to support the OS update now required. Before biting the bullet and buying a new computer I decided to try out all the PDF readers in the Playstore.

What a horror-show that turned out to be. There are dozens, maybe hundreds, of them nearly all crap. It appears every wannabe Android developer from Lapland to New Zealand writes one as his apprentice piece, then bungs it online to reap 2-star reviews. All assume you’re using a phone rather than a Chromebook, all offer their own wretched file management, and few can see external USB drives at all, so I soon became Olympically fast at the ’install-swear-uninstall’ cycle. Scouring the forums yielded recommendations for phone users only, few understanding requirements as demanding as mine. Then one glorious day in October I found Readera, which does almost everything that Kindle does (and a few things more) at remarkable speed and with a very comfortable UI. Not only does Readera read every ebook format you’ve heard of, on Windows, Apple or Android, but it does it for free without ads. And not only does it see external USB drives but its own PDF manager with cover thumbnails is actually superior to Android’s. I used the free version for several weeks but they lured me into spending £14.49 (one-off) on the Premium upgrade which offers colour-coding for notes, which I’d started using under Kindle but then lost with the Chrome extension.

An unexpected side-effect is that I’ve started using PDFs for all sorts of purposes besides book reviewing. I’ve always known you can print documents to PDF rather than a physical printer via the system print dialogs in Windows, ChromeOS, Android (and I imagine iOS too) but I’d never found much use for this until recently. I started having problems saving web pages to Pocket for various reasons and so the penny dropped – just Save To PDF instead and store the file locally rather than in Pocket’s cloud, USB storage being so big and cheap.Some websites offer a download option, but for those that don’t, save-to-PDF often works and retains all formatting and colour pictures. 

I must stress that I’m not a pirate who re-sells or wishes to profit from such copies – they’re just aides-memoire for my writing of both book reviews and this column. For many years I used browser bookmarks but weblinks can break or just vanish and the bookmark hierarchies in successive browsers became more and more unwieldy. My local Idealog folder has subfolders like AI, Bio_Neuro and Quantum full of PDFs from journals like Nature, which I can find quickly to swot up or quote from. 

What I’d like is a good grep-like utility that searches for text within multiple PDFs. Adobe Reader does this but its file management has irritating limitations, so occasionally I fantasise about writing one in Python (and bunging it in the Playstore to reap 2-star reviews) but the feeling soon passes…    

[Dick Pountain thinks PDFs are Pretty Damn Fine] 

ARTY FACTS

Dick Pountain /Idealog 363/ 05 Oct 2024 03:05

When I’m not writing this column, which let’s face it is most of the time, I perform a variety of activities to keep me amused. Apart from walking, cooking, reading, listening to music (live and recorded), playing and fettling guitars, I have a couple of computer-based ‘hobbies’, namely making computer-generated music and computer-generated pictures (non-moving). I recently restarted work on my Python-based computer composition system Algorhythmics – which I described here back in issue 306 – after a four-year rest. What prompted me was viewing a YouTube video about the Indian mathematician D. R. Kaprekar and some interesting numbers he discovered, so I set-up Algorythmics to compose a short piano sonata around two of his numbers, and it sounds like a fairly pleasing mashup of Ravel, Janáček and Satie.

I’ve been documenting my efforts at visual art in this column for over 20 years as I marched through successive generations of paint software from Paintshop Pro to Photoshop Elements to Sumo Paint to Artflow. I loved art at school and can use both paint and pencil reasonably well, but I’ve never been tempted to use either seriously since I discovered the computer (any more than I’m tempted to write articles with a quill pen since I discovered the word-processor). The crux is editability: once you discover the infinite flexibility of digital imagery it’s hard to give up this ability to experiment, redo and correct without wasting paper, canvas and paint. Images on a screen certainly lack the texture of proper paintings, but then I’m strictly an amateur with no realistic ambition to make a living selling my work.

I’ve also been into photography ever since the 1960s and my computer explorations began by processing snaps to make them look like paintings, which taught me how to use layers and blend-modes to take total control of both the colour and tonal content. Later I began using this knowledge to create purely abstract images.

It won’t have escaped anyone in the habit of watching Instagram, Facebook or YouTube reels that there’s a remarkable revival of abstract painting going on right now among the social-media savvy younger generations. Unlike me they’re not working in the digital domain but rather in the messy and expensive domain of wet paint. Under various labels like fluidart, spinart and poured art they’re making dynamic action paintings – Jackson Pollock style – by pouring multicoloured acrylic paints onto a canvas, manipulating it using palette knives and then spinning and tipping it to produce a finished image. Often they incorporate silicone-based additives that introduce cell-like patterns of bubbles, and the results have a very particular biological look. Some of them are very attractive indeed and nearly all are highly decorative. I doubt that many of these folk consider their products to be high art but they are highly saleable, and that’s on top of any revenues that derive from successful social media hits, which is just as well as the costs in canvas and acrylic paint must be considerable.       

I'm very keen myself on early 20th-century modernist abstract art, especially Vassily Kandinsky, Paul Klee and Sonia Delaunay. I don’t set out to imitate their works but merely play around using digital processing on a starting image, which could be a photograph, a clip from a website or a digital image that I draw by hand. Mostly I just use a mouse nowadays (I’ve had several Wacom tablets in the past) since my starters are so simple. Another way I sometimes start is by using Zen Brush 3 on my Samsung Galaxy Tab, a delightful finger-painting program with extraordinarily realistic watercolour bleeding effects, and then send the result to my Chromebook via Android’s Nearby Share. Then I spend some time layering, blending, smudging and slicing until I see something I like, which does indeed tend to mean something that reminds me of one of my modernist mentors. 

I’m not at all tempted to go in for fluidart myself to make money, even though those attractive canvases are more readily marketable than digital prints. That’s not only because I don’t have a garage in which I can splatter paint up the walls, but also because whenever I watch these artists at work on Facebook, more often than not I find the earliest part of a new work the most pleasing but then they keep adding too many colours and over-do it. So, while watching one particularly spectacular piece I hit ‘Watch Again’, then hit || to stop it at an early stage I liked better, took a screenshot and used that as a starter for my own piece! This potentially raises a novel legal issue about copyright and plagiarism: I froze a moment in time that didn’t make it into the final painting, so was I really stealing…. 

[You can see six of Dick Pountain’s abstracts at https://www.facebook.com/dick.pountain/posts/pfbid02UBtGRbAU7aTLSPYjeyTebxJUjMJ6EME6cKd5iqYBsYcdbaPCPrUNxZNqJhE48rSKl and hear his Kaprekar tune at https://soundcloud.com/dick-pountain/kaprekar-sonatina]

  

SMELL U LATER

Dick Pountain /Idealog 362/ 05 Sep 2024 11:5

I have no qualms in claiming that I have a better (or least better trained) sense of smell than the average citizen. That’s partly, maybe mostly, because I studied organic chemistry in the 1960s. During my first few weeks of working in the cavernous Victorian college lab I was instructed to learn the odours of a dozen commonly encountered chemicals, and advised to employ smell as the first step in recognising any new compound. I can often tell an aldehyde from a ketone by sniffing, and became briefly addicted to ionone, cinnamaldehyde and menthol in succession, carrying little specimen tubes in my pocket. I’m sure this method is no longer taught, on health-and-safety grounds, as there are many substances nowadays that can kill at one sniff. 

In later life this training came in handy when my brother-in-law Pip founded The Scotch Malt Whisky Society and was writing a book that needed to categorise the nose of various famous spirits. Of course odour is now a huge business, not merely for perfumes as it has been for centuries but for those hundreds of flavourings contained in most supermarket foods which are manufactured in a huge chemical works in New Jersey. But smell has barely impinged upon the computer business so far, apart from the smell of burning insulation which most of us quickly learn to recognise (and investigate…)

I wrote semi-humorously in an earlier column about the possibility of a ‘sminter’ loaded with an assortment of smelly ‘inks’ that could be triggered via internet messaging, and I even got a letter some years later about an (unsuccessful) attempt at one. Even Hollywood attempted a brief stab at a Smell-o-Vision movie (‘Scent of Mystery’  by Mike Todd Jr.,1960 in case you’re interested) but the obstacles in both cases were the same, that smell is a chemical, not electronic, signal that moves at the speed of breeze rather than light – and you can’t just switch it off quickly either…

But a far more serious obstacle is that while the components of human light perception are threefold – red, green and blue retinal cells – the components of smell perception are vastly more numerous. Our noses contain at least 400 different chemical receptors and individual smells are recognised by trillions of combinations their outputs, which release a plethora of proteins that are still not entirely understood.  

But when you hear the word trillions nowadays, you’re usually either talking about a GPT (or at least about NVidia’s market cap). Understanding smell perception, like protein folding and DNA sequencing, is a perfect candidate for AI to analyse, so it comes as no surprise to learn (via an article in Nature https://www.nature.com/articles/d41586-024-02833-4) that many teams are working toward this end, with ample financing from industries. 

The problem has several aspects, which include: predicting the smell of a molecule from its structure; decoding the output of human odour sensors for particular compounds; and automating comparison of smells of different mixtures by identifying their components. The current hot variant of AI – the Generative Pre-trained Transformer (GPT) – works using the mathematics of parameter spaces: identify the important parameters of the subject to be analysed, apply tensor calculus to create a multidimensional space with a dimension for each parameter, and then map training examples into this space. For graphical AIs like Stable Diffusion and MidJourney such spaces already have trillions of parameters for identifying shapes in visual worlds.  

One immediate problem for applying this to smell is getting training data: odour receptors, whether human or animal, are hard to study, often won’t work outside the creature, fragile and the amount of protein released is minuscule. Two receptors from insects and two more from mice have been deciphered in the last year, leaving just 400+ more to go. A team at Duke University in North Carolina is using AlphaFold and machine learning to screen millions of chemicals for binding to two synthetic receptors they’ve engineered. A very important motivation for such work is to use smell recognition in diagnostic medicine by identifying odour molecules produced by disease processes (dogs are doing this already). Precisely how and where odour nerve signals are processed in the brain is perhaps the leading-edge study right now. 

Real progress is being made and AI may soon speed it enormously, but smell remains the least understood of our senses, and least amenable to digital manipulation. It’s so subjective that human tasters and perfumiers will retain an advantage over automated solutions for far longer than most professions, and I don’t expect to have to consult my laptop when I’m mixing our own custom bath oil from my little box of tubes of neroli, rose, ylang-ylang and sandalwood oils (plus several other secret ingredients).  

[Dick Pountain believes that a rose by any other name would smell mostly of β-phenylethanol] 

ARTYFICIAL INTELLIGENCE

Dick Pountain /Idealog 361/ 08 Aug 2024 01:04

Who’d have thought that AI could become so boring so quickly? I feel a rather urgent obligation to devote a column to what’s been happening, but it appears to be happening faster than I can type. Last month NVidia’s stock rose faster, and then fell faster than a SpaceX test shot. Then Amazon reports that its Kindle self-publishing platform has been flooded by such a torrent of AI-generated bodice-rippers that it’s having to impose a limit of only three books a day on its customers.

Because Amazon stops short of banning AI-generated content altogether, ChatGPT is rapidly becoming a cancer on the body of the publishing industry in more ways than just crap e-novels. When I’m reviewing a book I often look online at other people’s reviews, not to plagiarise them but to see other opinions. Recently though I’ve witnessed spammers using ChatGPT to spatter the net with worthless AI-generated paraphrases and summaries of best-selling real books, which make it almost impossible to find any proper critiques. And it’s not only words. The hand-crafted goods site Etsy recently told The Atlantic Magazine that it’s being swamped by AI-generated tee-shirts, mugs and other merchandise which employ ChatGPT to optimise their Google search rankings and crowd out real producers.    

In a previous column I was worried about abuse of AI deep-faked photographs to compromise political opponents, but that’s turned out to have been somewhat wide of the mark because it requires a certain political seriousness on the part of the perpetrators. What’s happened instead is that Midjourney and its ilk are now enablers of pure fantasy and pop-surrealism. They allow everyone and anyone to produce memes and professional looking posters that make merely spray-painting slogans on a wall feel like something from a previous century. 

When I first tried out Stable Diffusion a year or so ago I was amused by the way its limitations generated such hilariously surreal images, but I’m not laughing now. The recent UK wave of far-right activity against immigrants and asylum seekers has been organised online via Telegraph, The-Platform-Which-Used-To-Be-Called-Twitter, TikTok and other social media, and an important part of this rallying process is a new genre of surreal nationalistic propaganda memes. Popular content components of such productions are squadrons of Spitfires, St George Cross flags, knight-crusader figures in mediaeval armour and British Lions (often wearing tee-shirts but no trousers and playing cricket), all meant as symbols of that old Britain they believe has been stolen from us. GenAI tools enable them to churn out infinite combinations of these icons in glorious Marvel-comic colour and for minimal effort. It’s worse still in the USA where these same tools are being used to depict Donald Trump with a six-pack and Hulk-like musculature, occasionally with the golden wings of an archangel and a blazing sword. Visual satire has a long history from Cruikshank, Rowlandson and Gillray, to Georg Grosz, Otto Dix and Ralph Steadman, but it always required a graphical skill that has now been entirely eliminated, and the purpose is no longer satire but adulation.

It feels as though we’re currently in the ‘phony war’ phase of a marathon battle between states and AI companies over regulation of the internet. The AI side continues to bluster about transforming the world’s economy with soon-to-be invented AGI, while also admitting that it will need to use about half of the world’s electricity supply to do so (and investing in fusion research). The state was until very recently clueless about the threat posed by widely available generative AI tools – though the wave of violence in the UK seems to have awakened them smartly – and they’re also quite chary about imposing content regulation, over quite legitimate concerns about freedom of speech. The owners of online content – publishers, television and film companies – form a third force standing on the sidelines watching the impending battle. They’re furious that the AI companies have already scraped a sizable chunk of their properties without payment, but also acutely aware that there might be a profit opportunity here somewhere (who wouldn’t like robot authors and actors that you don’t need to pay?). 

As for us poor authors, artists, actors and other creators we can only just watch aghast, while some members of the general public perhaps see possibilities to gain quick entrance to the so-called creative world without the bother of arduously learning a skill (hence all those Amazon three-a-day novels). How this will all pan out is beyond anyone’s (even GPT 4.5’s) ability to predict. Too many variables, like who becomes US president in November, and too many hero/villains like Trump, Musk and Altman with hidden and volatile agendas. My guess is that the stock market might just call a halt to hostilities, and quite soon…              

[Dick Pountain only uses ChatGPT as a party-trick to horrify arty friends]

FULL CIRCLE

Dick Pountain /Idealog 360/ 07 Jul 2024 11:12

Astute readers <aside class=”smarm”> which of course to me means all of you </aside> will have noticed from various other features that this is the 30th Anniversary Issue of PC Pro, and since this is a monthly magazine, and since there are 12 months in a year, and since this is Idealog 360, the corollary is that I’ve been writing here since the beginning. I used the word ‘corollary’ there because it suggests a mathematical proof, and that is a ham-fisted way of introducing my theme for this month, which is mathematics. 

360 is a special number to me not simply because it represents 30 years, but because when expressed as an angle in degrees it represents a full circle, a return to the beginning. Another way to look at a full circle is in radians as an angle of 2π which I find more congenial because π is an irrational, even transcendental, number and I like to think of this column as being sometimes irrational and occasionally even transcendental (which you astute readers may have noticed). 

What I’m tiptoeing around here (in this nauseatingly arch manner) is a confession, namely  that I’m only posing as a computer nerd, that I’m actually a mathematician manqué, a math sheep in hacker/wolf’s clothing. At school, way back in the early 1960s, maths was my top subject in which I got a distinction at S level. I had to choose between reading chemistry or maths at uni but was seduced into the former by the lure of stinks and bangs over pencil and paper. My introduction to computing did come very early, in 1962, as part of a school team who built a prize-winning computer out of ex-RAF radar set parts, but that computer was analog, not digital, and all it could do was solve sixth-order differential equations and display the result as green squiggles on a cathode ray tube (which only real maths nerds could appreciate).

Math-nerdship never left me even once I discovered ‘real’ digital computers. At college I only ‘used’ London University’s Atlas to process the statistics for my biochemistry experiments. After Dennis Publishing (or H. Bunch Associates as we were called then) bought Personal Computer World in 1979, as the only maths nerd in the room I was delegated to take home a Commodore PET and learn how to program. I discovered that I loved it, but math-nerdship continued to steer my journey because after Basic I learned Pascal, Forth and Lisp, rather than C which would have been the obvious choice were I to want to make a living from coding (which I didn’t and don’t).   

Elsewhere in this issue you’ll find our nominations for the most important milestones in computing over the last 30 years, so rather than recap those here I’ll instead name a few of my favourite milestones in computer-related math-nerdship. Thanks to the internet everything is computer-related now, so I follow developments in maths through YouTube videos, Wikipedia articles, Royal Institution and TED talks, but most of all through the excellent, non-profit, Pulitzer-Prize-winning online magazine Quanta. 

Launched in 2012 to promote public understanding of mathematics, theoretical physics, theoretical computer science and the basic life sciences, Quanta is funded by, but editorially independent of, the Simons Foundation. James Simons is a mathematician, educated at MIT and Berkeley, who started out working on pattern recognition, string and quantum field theory, then went to Wall Street and used his maths as a  ‘quant’ investor to become the 51st richest person in the world.

My favourite recent Quanta pieces have been by Philip Ball on [The New Math of How Large-Scale Order Emerges | Quanta Magazine] and one on Dedekind [How the Square Root of 2 Became a Number | Quanta Magazine]. Number theory is the part of maths that still entrances me. Irrational and transcendental numbers like π have infinitely many, non-repeating digits after their decimal point which makes them a little awkward to handle. Dedekind found a stunningly elegant way to pin them down, by splitting the number line into everything below and everything above the one you want. 

I still write programs – in QPython on my Chromebook – though nowadays they’re almost always about maths, playing with palindromic numbers or fiddling hopelessly with the Collatz Conjecture, or just solving a puzzle from Quanta magazine. I watch tons of YouTube videos that use clever visualisation tricks to explain p-adic numbers and their relation to the Riemann Hypothesis. The great thing about maths is that it doesn’t require a lot of apparatus, just a brain plus some sand and a stick, chalk and a blackboard or pencil and paper (or Python and a Chromebook). And there’s always a chance of being that amateur who makes a significant discovery…   

[ Dick Pountain is quite satisfied with his slice of the π ]