Technology evolves rapidly. It can be likened to evolution. Survival of the fittest. Not the best or the best designed, there’s a lot of awful technology out there, but the one that persists for any reason. Unlike evolution, which is undirected and a vertical process, (every change is on top of something already there), technology progress is tested for “fitness” as accepted by a market, and new ideas from their creators can redirect the path in a lateral orientation, (remember Edward de Bono’s “Lateral Thinking”?), starting all over again.
In the early 1950’s I remember Dad bringing home these amazing miniature records, only 7 inches diameter, a miniature of our 78 RPM discs. We could not play them until we replaced the radiogram, and were surprised to see also a button for 33 on it, whatever that was. But the direction soon changed to Philips’ cassette tapes invented by Lou Ottens, fitness being ensured by royalty free public release by pressure from the Sony Corporation on the inventors. And as soon as I had bought my compact tape players and made compilation tapes from my 33RPM long playing records, they were outdated by CD’s, which after investing in CD copying and burning equipment are now obsolete because all music is floating in from a cloud somewhere. No equipment necessary. The first transition was a culture shock, now it seems that nothing is worth trying because tomorrow will be a new, new-thing.
A great Silicon Valley entrepreneurial technology memoir I read was “The New New Thing” by Michael Lewis in 1999 published by W.W. Norton. It covered the makers and shakers of the Internet Boom, which was the new new thing at that time. Fast paced, humorous, vivid details of the personalities, the excesses, the anticipation and the transience. Wanton greed and gluttony is his pet thing in all of his books. It’s dated now, but worth a read. Several reviewers named it one of the best books of that year.
Of course the Internet is a real thing and is ubiquitous, but it is today a commodity and not a boom. Commodities are not luxuries, they are tools with a price ceiling and with the Internet reality came for the industry with a bang called the dot-com bubble. A commodity: “A product or service that is indistinguishable from ones manufactured or provided by competing companies and that therefore sells primarily on the basis of price rather than quality or style.”
It was not the first financial or economic bubble. Like a child’s party toy, bubbles pop up constantly. Pun intended. Because they are full of nothing but air and they burst suddenly after inflating to a critical pressure point. Some of the earliest were the Dutch Tulip Mania in 1630, the South Seas Bubble in South America a hundred years later when the term “bubble” first appeared, in our times the 1999 dot.com bubble followed closely by the housing bubble in 2000. No one learns, they have come faster and furiouser. Sounds like a movie franchise. That is because of the very fact of their avarice being driven by human psychology: speculation. It’s gambling and its contagious. Calling it the Tulip Mania hit the nail on the head and maybe the expression “mania” should be retained as a warning. The 1929 Great Depression is the quintessential archetypal.
The warning signs are liquidity, when there is too much money chasing too few assets, and inflation when the asset market valuation is far greater than the underlying true value.
The current New New Thing is AI. Artificial Intelligence. Regardless of the accuracy or appropriateness of the term “artificial intelligence”, and efforts by practitioners to alter, or correct, or malign, or challenge it, it sticks. Defining AI is like many things, evolution is again a typical example, something we kind of know intuitively, but find difficult to detail precisely when pressed.
But AI is not a bubble. Since about 2020 is has dominated the technology sector. Early pioneers and adopters are OpenAI with Microsoft involved now of course, Siri Inc now with Apple of course, DeepMind, DeepDream. The big players have come in quickly, companies like Alphabet Inc, Amazon, Apple Inc, Meta Platforms, Microsoft. And following the leaders are Claude, Gemini, Copilot. Later participants include for instance CloudFlare adding Workers AI; one of the largest cloud platforms, with according to them, about 21.3% of all web sites located on their servers.
And everyone is on board. Sam Altman, Elon Musk, Donald Trump, Mira Murati. And the money! Musk’s XAI and Murati’s Thinking Machine Labs each with $1 billion. $1 billion seems to be the new $1 million. From Wikipedia, “In 2024, AI patents in China and the U.S. numbered more than three-fourths of AI patents worldwide. Though China had more AI patents, the U.S. had 35% more patents per AI patent-applicant company than China.”
Or is it a bubble? Is it a mania? OpenAI had revenues of US$3.7 billion in 2024 with a loss of $5b, advising the market that they expect to stack up “stunning” losses of $74 billion over the next 3 years. The money being pumped in is staggering. Newcomer Anysphere $2.5b valuation on revenues of $100m. Last year fifty AI companies raised $142,45b billion. The technology, whether Large Language Models, Generative Transformers, Machine Learning, Deep Learning, Natural Language Processing, Scene Analysis, are energy devourers, cloud based, are error prone, subject to civil, copyright and other law suits, unregulated legislation at present, and other challenges. Time will tell.
Immediately before the AI Boom, was the previous new new thing. Virtual Reality. Its history and profile are identical to the Internet, the Housing and the AI booms. Is that an omen for AI?
Ross Rubin from Tech Radar says that VR’s Golden Age is over, and there was not much gold there. “It was promised as a new age for businesses. Virtual reality, augmented reality, mixed reality – in whatever shape it took, this was going to be the 21st century game-changer. No more staring at screens or using a mouse. That’s ancient, 20th century thinking, that is.
“This new reality would see the advent of true hands-free computing and unparalleled remote experiences, wherever in the world professionals were based. From prototyping to healthcare diagnoses, it heralded a new age. And, like those stuck in Casablanca, we waited. And waited. And waited. Now, it’s clear the death knell has been well and truly rung. Microsoft HoloLens is dead. Google Glass, once hailed as a path to improved productivity, doesn’t exist in any meaningful way. Now, Meta’s own Quest Pro has bitten the dust.”
-Does this all sound familiar?-
Rubin continues, “During the last decade, though, no company did more (or certainly spent more) to popularize virtual reality – the classic, immersive flavor of spatial computing – than Meta.
“While the company never characterized the Quest as a video game console – and neither Meta nor the leading console makers referred to the other as major competition – Meta’s developer events would invariably highlight new games as the headset’s leading third-party content.
“The enterprise use case was rarely, if ever, put to us.”
-To me this resembles one of my biggest gripes around technology development. Technology pundits assuming everyone is like them and wants what they want, to the point were often preceding failure no verification of market size is ever undertaken.-
“There are a number of reasons for the lack of up-take in the professional (and, indeed, gaming) world.
“Those grand promises, it turned out, were all mist. As tangible and as real as dreams. That overpromising of the capabilities inevitably led to user backlash when it was discovered they couldn’t effectively deliver what businesses needed.
“Features like X-ray vision, simply didn’t exist. And what did exist felt as limited as the field of view offered by many devices. Software compatibility remained a perpetual problem – the VR/AR software itself was often unreliable and unresponsive. Integrating that with existing IT architecture was time-consuming, costly, and borderline unscalable.
“There were issues with the comfort of headsets during prolonged use – eight hours of a working day using one was all but impossible, assuming the motion sickness didn’t get to you first.
“Then there was the cost. Developing and manufacturing headsets like the HoloLens was high. For consumers buying them, they were positively prohibitive.
“That has a major knock-on effect across the wider VR/AR market, such as gaming.
“For example, Quest unit sales have been a small fraction of smartphone sales, making it harder for developers to drive revenue with free-to-play games (at least for exclusive titles). Even Microsoft’s Xbox Series X/S consoles, hardly best-sellers themselves, are estimated to have outsold all Quest headsets despite launching after the first Quest.
“Given that companies such as Meta and Xreal are only now making the jump to in-lens displays, it seems that, until smart glasses can produce something reasonably close to the Quest experience, the momentum is elsewhere.
“That said, Google may provide a silver lining.
“Because it licenses Android XR to multiple hardware manufacturers, those might include companies more focused on professional business use, particularly with PC vendors coming into the Android fold with the Chrome OS successor “Aluminium.”
“Late last year, TechRadar’s editor-at-large Lance Ulanoff tried out a pair of prototype Android XR smart glasses. In his article said, ‘I think we’re just about done with our dalliances with too-expensive episodic-use immersion devices. The time for AI-powered AR glasses is now here.’”
“Of course, Google hasn’t hit a lot of home runs with Android in categories beyond smartphones, and PC vendors including Acer and HP were burned by Microsoft’s Windows Mixed Reality headset initiative.
“But if brands like Dell, HP, Lenovo, Acer, and Asus can get the ball rolling with pro-grade business tools and enhanced AI, there’s potential to sustainably match or exceed what Meta spent big to create and is now moving quickly to abandon.”
So, from cylinders to 78s to 45s to 33s to tapes to CDs to Spotify. VR from headsets to smart glasses, to what? Embedded brain inserts?
While bubbles burst suddenly, they take some time to build, the decade from 1711 to 1720 for the south Seas Bubble, the decade of the “Roaring 20’s” for the Great Depression, the decade of the 90’s for the Internet dot.com, a decade from mid 1990’s to mid 2000’s for the Housing Bubble, and it has only been 5 years of chatbots. During that phase, all looks, sounds, feels, good, momentum builds, everyone is on board. Maybe on board a hotter than air balloon?
Thanks Alan. I do not share your gloom for AI. VR is a given and in hindsight is a mere blip in the technology journey. I see AI as totally transformative as was the industrial age and the dawn of the internet. I was at the cutting edge of LCD projectors having had a career with Bell & Howell 16mm projectors and Kodak 35mm slide projectors. Before that as a younger person the B&W darkroom was my fun place as I embraced analogue (film in those days) photography… something I am ready to jump into once again. I mention these because they come and go. They have evolved somewhat like your music analogy. AI I predict will not be the same. I liken AI to having access to electricity or the light bulb (or LCD). It will not be a bubble because our only option is to embrace it for how it will manifestly change everything we thought was stable. Enter quantum physics and quantum computing and the outcome is almost impossible to imagine. Exciting times ahead at a time when civilisation as we know is on a precipice.
Hi Geoff, welcome to the fold, and thank you for your first comment. You picked up on my reserve. Readers, Geoff is a neighbour of mine and a significant expert in IT, and current with AI proficiency, so his comments will be worth considering. Actually I do think AI will be here to stay, just like the internet is and like VR will be in some form or another. Many of my friends in other discourses are aware of my skepticism, not of AI per se, but what it is purported to be. Something about “bee’s knees” rankles me. I think AI is great, I am a pioneer in it of sorts, but it is a lot more than chatbots and the current wave to me. Speech recognition, pattern recognition, task planning, problem solving, all the things our brains do. So one gripe is that AI has almost become synonymous with just chatbots. At least in the public eye, and that disturbs, lowers AI, I feel. I have described the chatbots as (just) the use of current search technology with improved natural language generation, so the search results are presented as a human-like conversation. But what it does do is amazing, no doubt. I have two concerns, and the jury is still out, hence my post about is it a bubble? Despite what we know about the brain, we know little in totality. SO it is early arrogance to compare current AI performance with what we do with out grey matter. And the other is the hype and propaganda. Here in Bangkok there are a series of TV ads suggesting people forgot a $100k 4 year college education and take a 4 month AI course and make more money. Yeah! For someone. That is what caused the dotcom bubble, half educated wannabes setting up million dollar startups that failed basic business methods, all collapsing. So I think AI will be around forever, will improve, maybe even become the General AI system everyone wants, or does not want, but become a rational, even tempered commodity.
When I was working in speech recognition (and had young children at home) I made a rough calculation that a growing child has heard several million phonemes uttered by adults by the time they learn how to speak. Large Language Model hadn’t been thought of back then but when I first heard of them it clicked.
Certainly huge exposure is key. In changing the synaptic connections that create learning, or in training in LLM which creates the data base the RobK refers to in his comment. I travel and live and work around the world, and with my adult knowledge of language and languages, I find it difficult to learn new languages because of just that, the lack of huge exposure.
I see AI as a mass data handling/ organisational tool. The question is what level of resolution is the ultimate or optimal and for what purpose . Im guessing only time will tell.
I see it in a similar vein. Data handling in the sense of search, then a consolidating the search to eliminate redundancy and repetition, then generating a natural language version of the search results. I don’t think it makes any associations, cause and effect things, or new insights, but I am still thinking about that. I think it finds those associations from where someone else has done it. Mostly I am not completely familiar with how they work. I’ve studied it a bit and see some of it, but some maths is beyond me.
AI, artificial intelligence, is all the rage. People in the computer business use it daily and are mostly sanguine about possible downsides. Before these cutting edge technologies become widely known and used by all and sundry, there were almost invariably long periods of gestation and development, and, when functioning, often a period of exclusive, even covert, use by privileged parties.
When the personal computer entered the market, computers had already been used by the likes of IBM and others. During WW II, apparently, punchcard machines, early versions of the computer, were used by Nazi Germany to enable the authorities to manage the war effort more efficiently. War is now impossible to wage without the use of computers, internet and all that. AI is being used for war, both for defence and offence. It stands to reason that they who can afford it, use it, and they who cannot afford not to use it have used it at their earliest opportunity and to the limits of its, AI, capability. These capabilities have been growing well beyond Moore’s Law, and there is no telling when and where the use and growth of AI will slow down or stop. My guess; if the application of AI eventually starts slowing down, those of us then alive may well live in a world we would not recognise as being fit for humans to live in. If the use and range of AI were to suddenly stop growing, something drastic will have happened, and the world may not be fit for humans to live in.
The uses to which AI can be put are virtually without limit, I think, but if one were to ask any of the AI bots now …, you don’t know how correct the answer might, eventually, turn out to be. Currently, AI is driving a wedge into the communication space between people amongst themselves, but more so, it erects a barrier between the authorities/the controllers and the public, where the communication is increasingly one-way from the top down, influencing, censoring, subtly threatening. Since automation and computer-answer phones, the public has also been limited in talking back to an annoying extent. If that has increased bureaucratic efficiency, it has not necessarily led to more efficient delivery of governance.
Though the PC and the internet have enormously increased two-way communication between people and peoples in almost all configurations of relationships, one-way communication from controllers/top-down users has been greatly enhanced compared to earlier methods of top-down communication. AI as a means of top-down communication through the internet and computer use both by the “consumer” personal user and the authorities, the latter now includes all top-down computer/internet/AI users, is likely to expand to, for all intents and purposes, a currently incalculable degree. Which leads us (all) directly to geopolitics.
The uses to which AI can be put for geopolitical purposes is surely enormous. AI has sofar been used for geopolitical purposes, but one cannot, without research, guess to what extent. One can guess, however, that much of that use by the authorities is classified, and will be more so in the future. It’s not only the enemies as narrated in the mainstream media which are and will be kept in the dark here. Politics at any level and all is at heart a matter of influencing and controlling people’s and peoples’ minds. Like religion before it, from which politics evolved, politics is about top-down controlling of people and peoples. The idea that politics is a two-way affair is only relatively recent, except perhaps in tribal situations. In the West it can be seen to have started in a limited way in the Greek city states and more comprehensively with the pioneers of Protestantism: The likes of Bruno Giordano, Jan Hus, Zwingli, Luther, and so on. The Thirty Year War is an exmple of the demo trying to communicate from the bottom up. The Thirty Year War was a bloody and protracted conflict; not as protracted as the One Hundred Years War between the fledgling nation Holland against Spain, say, but it took a greater toll on the population, mostly centred on southern Germany. The struggle by people to have a say in how and to what extent they are controlled by authority has been long, unrelenting and bloody, so far.
So, the public, as we are now mostly designated, or “citizens” by, amongst others, the EU (real citizenship in the once sovereign EU member states having effectively been abolished), Joe, Jill and the alternative genders as public are guaranteed to be kept in the dark about how we are ruled from on high. Some of these measures to keep us in the dark are no doubt sensible, so as to facilitate certain administrative and/or other essential safeguards. However, with AI as a tool the extent to and the means by which the authorities control the public is much more comprehensive than previously. We can safely assume the risk of overreach and wrongdoing is equally much expanded.
Various conspiracy theories, many of which have done the rounds for a lomg time, now look to be coming to fruition. I am here assuming that the authorities/controllers did not get their inspiration from the conspiracy theorists. It is natural that they who are in a position of control are people who have made effort to reach such a position and, once there, would make consistent effort to maintain their hold on such positions (of power). They who are in control do not need the public to give them the inspiration, though the way the public reacts to being controlled determines the ways, means and methods by which the authorities do control and maintain their hold on power, the aims being essentially the same.
Each country and people have their own tailored recipe for control. This is a problem for the authorities at the coal face of mass control, as shown in the EU, for instance, where its elites are trying to weld together a Europe of separate peoples still anchored, even hard-bitten, in their differences. In this problem, despite their differences, globalists and the delegated national/supra-national authorities are one, supported by supposedly opposing, but subservient, political operatives who also, ideologically and collectively, strive to impose a centralised one-way top-down regime. We can discern a merging of top-down and bottoms-up centralist ideologies in the all-encompassing geopolitical space. There is now nothing in the way of human endeavour or interest left alone by politics. Nothing is sacred, except the power held by the authorities. Everything is subjected, to be collared and subdued for the purpose of central control. Does the public, the lowly consumers of services rendered by the totalisation of authority ruling over them, have an answer to that, either through AI or otherwise? How would that answer be communicated from the bottom up if dissent is punished, as it already is, by whatever means the authorities think they can get away with? Medical doctors, scientists, bureaucrat whistleblowers, dissenting activists, etc., have already had their basic civil rights unlawfully denied in the West. Where/how will this end, one wonders. We could be in for a dystopian future, Orwellian, but incomparably worse.
Religion, politics and all ideologies are basically one of a kind, at all levels of human society. Lately, I have been thinking in terms of moral law versus natural law in the now looming presence of global governance instituted by means of a globalist geopolitics mostly hidden from the public and not openly acknowledged. Social democracy in its modern form can be said to have had its genesis or roots in Europe. It is grown from a shared moral basis, the origin of the more humane, or less inhumane, version of which must be sought, imo, in the regions along the eastern shore of the North Sea. Be that as it may, social democracy is, again imo, now gone. I put its demise at around the late 1970s, but, certainly, the planning from on high to strangle it I think was well before WW II, perhaps as early as the Russian 1917 revolution (likely to have been facilitated from the West). This revolution would have shown the authorities-controllers in the West how to guide the process of social democratification at home to suit their purposes .
Divide and Rule! Open warfare is only the top of the iceberg in that regard. Financial crises, medical emergencies and so on, are part of the package, induced, elicited, encouraged or otherwise. How essential is all this for responsible government?
Some would now wonder what this all has to do with Artificial Intelligence. Bliss for them that have great faith in the authorities, no matter what happens. We are looking at enormous changes coming our way in the very near future.
AI is artificial, and no better than what determined and ruthless criminals can make of it. If its controllers cannot or will not police it, except to suit themselves, it will be bleak indeed for the rest of us.
Starting with the first technology revolution, the agricultural one – not always thought of perhaps as a technology change, but it most definitely was, the insight and innovation to use the knowledge that animals and plants could be domesticated, tamed and bred, planted when and where desired, and the specifications for achieving all of that – there is evidence that technology revolutions create the elites, or controllers as you mention them Jacob. Each one since then has centralised power and control in more and more established higher levels of the resulting social structures. The AI revolution will do the same. Each new power structure is different. The latest seem to be the Sam Altman, Elon Musks of the AI world with some carry overs like Mark Zuckerberg and Geoffrey Hinton, but emerging new names like Anjli Jain and Semyon Dukach.
Pre-agricultural peoples typically lived in egalitarian societies. Their social organization was primarily based on hunting and gathering, which did not support hierarchical structures. The establishment of kingship and complex social hierarchies generally occurred with the advent of agriculture. As communities transitioned to settled agricultural lifestyles, they began to form more structured societies, leading to the rise of leaders and kings.
https://en.wikipedia.org/wiki/Neolithic_Revolution
Well; equating the AI revolution with the agricultural revolution without taking into account the long and bloody struggle for the European peoples to arrive at a system called social-democracy is a non-sequitur, unless you predict, or at least allow for, a return to feudalism. You did not, but may claim it is implied in your comment.
For years I have made mention of neo-feudalism taking shape in the West. Looking at Europe, I would say it has made a solid base there, and in Australasia. Whether the European peoples, with or without the assistance of other peoples, are able to stay the advance of neo-feudalism I cannot tell, but I do believe there is a window of opportunity for the electorates in Europe to wake up and return to social-democracy. Will they? Can they? I don’t know.
I don’t think I am espousing anything Jacob. I’m simply saying that technology takes jobs but adds new ones. I have not commented nor wanted to comment on the social implications.
That said, the long and bloody struggle that might arise from the AI revolution is yet to come. A close simile would be the financial oligarchs who caused the “sub prime” problem. Elitists who never were held to account. But the social consequences were significant. Foreclosures, families unhoused and ripped apart, consequent legal and health problems from broke and homeless, suicides. There is definitely a neo-feudalism movement in the USA with King Trump at its head.
But what the heck, the rich have always had the power. And will do anything to hang on to it, regardless of any social movement. Mind you, modern communications and networking give the masses more leverage than ever before. No ignorance these days (unless it is self inflicted). That is why one of the first things autocrats do when there is a modern uprising is cut off the internet. One day it will be to cut off the AI.
Off-topic, but about economics and the politics of economics: Not only banks take your money. In Holland, a tax haven for multinationals and very easy on ex-pats working there; there has long been a tax regime for residents and Dutch citizen non-residents which is not just punitive but criminal. The government of the day, no matter its coalition makeup, taxes people with private assets on the value of these assets- the wealth tax. Economic policies decided by these governments drive inflation, and consequently these assets increase in value every year. The rate at which these assets are taxed also increases year by year. The system is based on inflation.
Next, the plan is that the annual increases in valuation, usually based on council rates, will be taxed at 36%. If the assets go down in value the year after, when people may be forced to sell these assets because they have to pay tax on fictitious income, tough luck. At any rate, holding on to private assets become very expensive and one is paying tax on fictitious earnings which may never be realised when the assets are sold.
Maybe one doesn’t need AI to work out what the result will be , but there will be a large majority in that electorate which will have no private assets and which will be totally dependent on regular paid work or government hand-outs. More to the point, private savings are being forced into funds and shares. The middle class will be strangled by and by. So, people who object can go take a hike. You don’t do social issues, Allan, but AI will certainly become an all-encompassing social problem. I don’t see how one can isolate social issues from everything else, or vice-versa.
So, you can see why I think democracy is a thing of the past in Holland.
Good thing: I don’t have to live there.
Most regimes collapse because the powers that be get carried away with their authority, taxing more and more to compensate for poor administration or to fund personal greed, to the point of retaliation. Using the positive feedback loop of inflation to insidiously ramp up taxes is one nefarious technique.
Retaliation by the proletariat is itself an avatar of democracy – by definition.
About pre-agricultural societies being egalitarian: I don’t know about that. These peoples would have been nomads. Small nomadic groups would have a simple hierarchy. The bigger the tribe, the less egalitarian these people would have been. The Huns and the Mongols were nomadic hunters and keepers of life stock. They were not part of the agricultural revolution you are referring to. A little research may be a dangerous thing, but I stand to be corrected. The Huns were beaten, at long last, by European forces. The Mongols, apparently, were not, but simply faded away. However, unless the neo-feudalists with their AI find ways to manufacture food to eat with the resources they are able to lay their hands on, they are dependent on farmers producing it and armies to protect themselves, the land, the food factories and the natural resources these neo-feudals need to survive.
Maybe the techno-AI neo-feudals will eventually move to Mars. Maybe they cannot do as much damage there as they are bound to do here. Never mind the footprint of these data centres, the resources these centers need in terms of power and water are fast becoming a huge burden on the rest of us. Will that be the hay bale to break the AI techno-kleptocratic neo-feudalists back? We’ll see.
It may not ne valid to transfer current, modern ideas of elitism and power mongers and hierarchies on what may have existed in simpler civilisations. I am sure there would have been elites in a form that suits their times. Think of sharmans and tribal elders and bullies.