Summary
Nvidia's shares dropped nearly 10%, removing around $280 billion from its value—the largest single-day loss for a US stock. This decline highlights market's tendency to create hype, and scepticism regarding AI's transformative potential. Politicians, particularly from the UK Labour Party, have overestimated AI's capacity to cut budgets and spur growth, overlooking its potential to cause unemployment and its actual limitations in technological adoption and utility.
Shares in Nvidia, the leading AI chip maker, fell almost 10 per cent yesterday, erasing around $280 billion of its supposed value—the biggest single-day loss ever for a US stock. The total loss over the last three days has exceeded 14 per cent.
Why does this matter? There are three reasons.
First, it shows the folly of using share-based savings for pension arrangements when most companies - even quoted ones - do not deliver long-term value.
Second, it shows the capacity of markets to create hype.
And third, it suggests that markets have realised that AI is not going to deliver on its promises. It may well prove to be useful, but it is not the shocking transformation that has been claimed.
This then matters because there are politicians - mainly linked to the UK Labour Party for the purposes of this analysis - who have believed all the hype around AI. It is AI that they believe can cut government spending budgets and AI that they believe can deliver private sector growth, even though they seem wholly unaware that AI would also be the source of considerable unemployment if all their hopes for its power to replace humans and so increase what they call productivity were to be fulfilled.
I know AI is going to be a part of life. To some extent, it already is. But I do not believe the hype. Most tech delivers vastly less than it promises for one simple reason, which is that people cannot work out how to use it. They might also not want to do so. This will be as true for AI as it is for all the other tools that we have on the IT we already own that we never use.
I have no idea whether the stock value of Nvidia has been over-hyped and is suffering a necessary correction. But I am certain that gullible Labour politicians have, under the influence of the Tony Blair set-up that appears to have massively undue influence over our current government, fallen for AI as the answer to all their problems when it is anything but that.
The fool and their money are easily parted. Some will have lost out badly on Nvidia. We all may lose out badly from the naivete of Labour politicians who think that everything the market has to offer glistens when that is far from the truth.
Thanks for reading this post.
You can share this post on social media of your choice by clicking these icons:
You can subscribe to this blog's daily email here.
And if you would like to support this blog you can, here:
Buy me a coffee!
Blair seems to have lost all contact with reality in his belief in AI. His suggestion is that instead of GPs we should have a chat box !! This should ensure the elderly die off quickly- if only from apoplexy trying to get help from a chatbot.
https://www.politico.eu/article/tony-blair-nhs-chatbots-uk-gp-healthcare/
God help us all / help us to afford the private GP fees.
Blairite Techno Cargo Cults ?
We will be saved from ourselves, and from the environmental and economic devastation that has been wreaked over 250 years of industrial societies, by technological solutions originating in corporate innovation and then by consumer capitalism.
Techno-cargo cultists like Blair have an unshakeable belief, that approaches religious levels of fervour, in the imminence of a new age of blessing, to be initiated by the arrival of a special “cargo” of techno fixes invented and developed by almost ‘supernatural’ techno sources.
We just need to trust cultural heroes like Steve Jobs (RIP) or big business entrepreneurs like Elon Musk.
They will save us all with their techno-cargo solutions.
Meanwhile we can all carry on with western consumerist lifestyles pretty much as normal.
Blair and his fellow techno-cargo cultists see a society administered by a well meaning technocratic priesthood dedicated to solving the problems of human suffering and we just need to trust them. They know best.
The new techno cargo cult world will deliver whatever material resources we need, and oversight of the technocracy will be by problem solving elites, including people like Blair or consultants appointed by them administering AI or whatever techno-fix.
Personally, I see Tony Blair more as Kaa the Python singing “Trust In Me” to the idiocracy he has cultivated.
@ AliB
“Blair seems to have lost all contact with reality …..”
If you recall, this happened while he was still in office as Prime Minister.
Like Thatcher he was barking before he left office.
I agree ref: A.I. hype.
Some musings. For my sins I am involved in a project that combines a renewables control system (supervisory control and data acqusition – SCADA) with an A.I. that takes (weather forecasts and turns that into a SCADA scheduling system (with the added complication of feedback from the SCADA so the two converge rather than diverge). I’d class the A.I. as quasi level 3 (= machine learning). I’m designing the SCADA and a uni’ is doing the A.I. Neither system is computationally intensive & could be quasi-standalone. That said, the real-time nature of the SCADA adds a layer of complexity. Does the system need semiconductors with 3 nanometer architectures? probably not.
I suspect that part of the A.I. hype was driven by the prospect of self-drive cars (and the image processing that goes with that) – which has faded somewhat (although the Chinese seem to be pushing hard in that area). The Ukraine war is leading to “self-drive-drones” & military apps will tend to be a driver in this area. The problem is not hardware, its the software which takes time to develop – & perhaps that is where things will slow down.
For those interested in novel developments: IMEC is a good place to start – the world’s leading semiconductor R&D site – established in the 1980s by…….. the Flemish gov – funny that.
Blair’s name appears now so frequently in the media and linked to the government, that I wonder who is really running the country and setting the agenda. To me, it is making a mockery of elections and our political system.
AI has seemed for some time to be a false god. I saw a remark that one mathematician (can’t remember who) made that AI large language models was just a bunch of very sophisticated statistical algorithms. As such, it can’t actually produce good original ideas.
But what also leapt out from Richard’s piece were the figures – $280 billion =10% of value, so the company was valued at $2800 billion! Surely, this figure must have been a clue that the value was over-hyped.
OK so there has been a hiccup.
But we must remember that politics is now all about making the unacceptable acceptable.
Enjoy the moment but stay vigilant I say.
They will bounce back.
Your comments and conclusions don’t just apply to AI, Richard, but also what used to be referred to as the ‘new technologies’ (IT/ICT) that supposedly drove and enabled ‘electronic government’, ‘Web 2.0’, the ‘digital revolution’, and various other catch-all labels invented to signal the arrival of the ‘information age’.
Over my 20+ years as an academic, from the mid 1990s, I was actively in involved in research in the UK and across the EU (via EGPA: the European Group for Public Administration; COST: Cooperation in Science and Technology, and, lastly, the EU’s Horizon 2020 scheme) which examined the use of ICT (information and communication technology) in its many forms and applications across local and central government. And never once did we find an application or project that delivered all – or in many cases even 50% – of the benefits (e.g. improving the efficiency and effectiveness of governments operations and actions) claimed when a technology or technology based project was purchased.
Furthermore, in every project that I can recall looking at the outcome was worse where the design and delivery of a project was outsourced. Indeed, this was well known both by those in the pubic sector tasked with commissioning projects and those in the private sector who designed and delivered them, not least because through the 1990s and 2000s many of those on the private sector side had previously worked in the public sector but had been ‘outsourced’ along with the technology as privatisation became the endemic (supposed) solution to the search for cost savings (or as the 1990s slogan used to run: the search for the three Es – economy, efficiency and effectiveness – of which one was way more important than the other two, despite claims by politicians that this was not so).
As you might imagine, I could write a great deal more on this subject but will resist the temptation. All I’ll add is that back in the day there was a fairly constant tension between ‘techno utopians’ – mainly from the US, as you might imagine – and ‘techno sceptics’ – mainly from Europe (I actually think we were ‘realists’, because a lot of the researcher I came across actually knew how organisations worked, rather than focusing on the technology). Tony Blair and his ilk are solidly in the former camp and have infected all around them accordingly. This is, of course, not entirely unrelated to the fact that he and his ilk have been heavily influenced (infected?) by US ‘techno utopian’ thinking, and spending too much time on the wrong side of the Atlantic – point Colonel Smithers has drawn our attention to on numerous occasions through comments across your various blogs. Sadly, I see none of this changing with Starmer and New Labour 2.0.
In principle I am a techno utopian – I live the stuff
I also know how hard it is to use it well in almost every situation I come across
I like your analysis
Richard, with this government (and probably a lot of previous ones) we know what this is really about; finding a magic silver bullet to avoid real investment in people and organisations because the cowardly state politicians still cling to the neoliberal model of government.
So AI will be touted as a way to miraculously increase the productivity of the CS, NHS, local government etc that have been under funded for years. The private sector will deliver fabbo AI to the lazy fuddy duddy public sector.
Of course as Ivan points out, these techno solutions never really do. And as you point out, it requires real work to use the technology effectively.
If you want a good illustration of the limitations of AI, then ask chatGpt how many r’s are in the word strawberry. It will get it wrong
It did
Bit I am not sure that proves AI as a whole is useless
It just says as yet it is pretty unreliable
No it doesn’t prove it is useless, but it’s an illustration that it isn’t “intelligent” in the normal sense of the word.
The way these text based models currently work is to break words into numerical tokens and then predict an answer based on what they have learnt about their likely sequence. They don’t actually attempt to count the number of r’s (they would struggle because they work in terms of of tokens and not characters)
So like all models they need to be understood before applying to real world problems. Within their domains of expertise they can be very useful, but those domains need to be well defined and it should be remembered that the models aren’t “intelligent” but merely guess the most plausible answer based on their training material (which may not be appropriate, as in the ” strawberry” example)
Thanks
I had assumed the Nvidia share price was slumping due to this news puncturing their over inflated expectations bubble:
“The Nvidia AI antitrust investigation is ‘escalating,’ reports Bloomberg”
https://www.theverge.com/2024/9/3/24235233/nvidia-doj-ai-antitrust-investigation
Commentary:
As the Biden administration is stumbling towards its preordained winding down, the anti-trust actions that it instigated are beginning to come to the fore.
As is so typical in American politics, voters will have to vote for the continuing administration in order to not have an incoming administration scrap the visibly anti-oligarchic and pro competition actions.
Plus ça change, plus c’est la même chose.
Just been rereading Galbraith on power.
His view of US antitrust legislation,
and especially enforcement over the last century is/was that it has not ever made a blind bit of difference in limiting the growth of US monopolies, and was almost exclusively tokenistic.
Galbraith was writing before Clinton’s corporate liberalism further undermined anti trust regulation.
He was right
It is simply down to the company losing its monopoly position and control over innovation and manufacturing.
The dumb attempt to hang onto IP and treat the Chinese and others as either not capable or not allowed to compete at the top end of the supply chain of chips. Whether it is 5G (and plus) or AI.
Huaiwe and other manufacturers have jumped the gap in both innovation and production, within a couple of years instead of decades as they were expected to fail under the dumb attempts at controlling their free trade through sanctions.
It’s similar in all such ‘high tech’ industries of the Collective West. Just part of the crumbling walls of the sandcastle, the supposed superiority and monopoly of the imaginary Sillyconehead kids and the mega tech corporations they front to retract continuous rents from the whole world.
AI is fed information, analyses it, then gives a response.
On that basis it could replace CEOs and the heads of Think Tanks like Blair’s more than front line workers. 😀
Isn’t AI a solution in search of a problem, as somebody smarter than me once said?
Yes