AI faces closing time at the cash buffet

It is the season of overindulgence, and no one has overindulged like the tech industry: this year, it has burned through roughly $1.5 trillion in AI, a level of spending usually reserved for wartime. Between 2001 and 2014, the wars in Iraq and Afghanistan cost the US an estimated $1.5 trillion to $1.7 trillion in direct …

[doublelayer]

Re: Gold rush

But after the frontend has handled the clicks, something still needs to get the information and process it, and Java is still often used in that. Certainly not universally, but just because something is a web UI doesn't tell you anything about what language does most of the work which will be whatever the team that built it wanted it to be, and if that team was from one of the many companies that have been building Java-based backends for business applications for twenty years, they're often not changing.

[MachDiamond]

Re: Gold rush

"So how's that going?"

New rush, new shovels needed. Companies need to innovate as fast as they can to stay in one place. The issue is that a behemoth such as IBM has so much inertia that it needs to see what the future will be a decade in advance to get aimed in the right direction to arrive there when the party starts.

[Ken Hagan]

Re: Gold rush

"So how's that going?"

We'll, IBM still exist despite it being ages since they last led the field in anything. That may seem a low bar but, given their enthusiasm for retiring their older staff, is there anyone at IBM today who was working for the company back when they led the industry in anything?

[simonlb]

Re: Gold rush

It's going to be interesting to see what happens to these companies when - not if - the bottom falls out of the AI market and they are committed to only delivering AI related products that are no longer wanted but they are tied into multi-billion dollar contracts to take delivery of the AI chips from the chip manufacturers.

[Tron]

Re: Gold rush

GAFA's legal teams are likely to have clauses that allow deals to be cancelled, whilst a lot of the cash is recycled between big tech companies. The mugs who just signed on the dotted line will lose the lot. The xPU suppliers will sell, sell, sell and then just ease up, having banked serious wonga. Governments waste [borrowed] billions on a regular basis. They will just draw a line under it and move on. Any parasites running businesses based solely on AI will fade away. The AI features will just not work any more. Any pure AI companies that can wrap up quickly and neatly will walk away with what is in the bank and flog their IP to the big players, who will licence it for component products (natural language interfaces, distributed AI). Those that believed their own hype will make like Lehman Brothers and be cited in future MBA courses. Everyone will stop talking about AI, as they did about the Metaverse, pretending it never happened. Then they will ramp up for another 'Next Big Thing' scam.

[MachDiamond]

Re: Gold rush

"Any parasites running businesses based solely on AI will fade away. The AI features will just not work any more."

If you can take somebody else's tool and start a business while that tool remains the property of that other company, you don't really have a business. Somebody else can come along the next day, do the same things and make piles of money through not having a bloated management structure. The company that owns the tools can go belly up, change those tools or see what you're doing, do it themselves and slow down your work until your customers all migrate to the bigger faster shinier.

[ChrisElvidge]

Re: Gold rush

And do you really think E Musk will lose his shirt? More likely that "Musk's money" will end up offshore and untraceable (he's a crook, after all) and the people who lent him the money will be left holding the can. This will include all the smaller shareholders.

[Sorry that handle is already taken.]

Re: Gold rush

Quite. The only company turning a profit from any of this LLM hype is Nvidia. Even Microsoft has stopped reporting on it directly.

Nvidia's at the point of attempting to incubate and cultivate its own customers so I assume it knows the end is nigh and is just trying to squeeze out a few more dollars in the meantime. The banks offering loans underwritten by depreciating GPUs will be left holding that bag.

Meanwhile I hope the DRAM manufacturers have a plan B for when OpenAI can't service its contracts.

[Yorick Hunt]

I had to re-read the heading.

Thrice!

My brain kept seeing it as "AI Faeces..."

[Excused Boots]

Re: I had to re-read the heading.

"My brain kept seeing it as "AI Faeces...””

Well, there’s a good reason for that....

[swameister]

Re: I had to re-read the heading.

arse-i-faclal intelligence?

[weladenwow]

Re: I had to re-read the heading.

AI?

?

More likely:

BDSM Built on Database Search Manipulation

BDSM2 Double-Down version of BDSM

BDSM3 Belatedly Desperate Survival Mechanism version of BDSM2

PipPip!

[Anonymous Coward]

Somebody tell Benioff...

... that maybe, if they themselves used their precious "production ready" AI agents, then my wife's company massive Salesforce "transformation" (how I hate that term) wouldn't be such an utter, utter shambles. If they are, then according to her, she can't tell the difference between "AI" and some army of underpaid off-shore drones.

[Bluck Mutter]

digital transformation!!!!

<RANT>

That's the one I hate the most.

True digital computing arrived in 1945 with ENIAC so all computers today are digital.... so what is a "digital" transformation?

No business is running their processes on an abacus and only mum and pop stores MIGHT still do accounting and inventory management with a pen and paper so despite the definition of Digital transformation being " the deep integration of digital technology and strategies into all areas of a business"... any large business is doing this already so what is there to transform?

It might have been a legitimate phrase in the 50's and 60's but not now.

But of course you need fancy words that the C Suite can't easily comprehend to ensure they open the corporate wallets for the new magic pixie dust of AI (previous iterations include Big Data, blockchain, SOAP, 4GL's etc)

</RANT>

Bluck

[Bebu sa Ware]

so what is a "digital" transformation?

I always imagine a finger puppet being non·consensually digitally penetrated.

[vtcodger]

Re: digital transformation!!!!

FWIW, "Digital Transformation" seems to mean. "We can fire all the workers and replace them with real cheap machines and the money will just roll in (from whom one wonders if the consumers are all out of work?) and we'll be rich as kings."

One strongly suspects that may not play out exactly like the sales pitch claims.

[Mike VandeVelde]

Re: digital transformation!!!!

It's when the paperless office finally arrives. Along with robot butlers and flying cars and cold fusion, and actual artificial intelligence. All just a few years away. Still.

[spold]

Re: digital transformation!!!!

ENIAC needed rewiring for new programs. The Manchester Mark 1 was the first stored program computer (although the preceding Manchester Baby proved the concept). The Mark 1 lead to the development of the Ferranti Mark 1 - the world's first commercial computer.

[doublelayer]

Re: digital transformation!!!!

I won't claim the term is good or defend those who use it, but there are more levels of what can be considered digital. If a computer is involved but doesn't do the work, processes could be considered not digital. For example, if the process for doing something is that I send person A an email, they do manual tasks, they send me an email, then the computers are nothing more than a faster memo distributor. There are many businesses who still have things like this around even though they could be automated so people don't have to do mechanical processes on data. There are at least more businesses with things like that than those trying to do it all using only paper.

[Groo The Wanderer - A Canuck]

The American investor community is notorious for obsessing over bubbles. It has happened time and time again in my 6+ decades on the planet, and it is happening now. And when the bubble bursts, the bankruptcies are rampant and huge. You can count on a significant chunk of corporate America to come to the federal government, cap in hand, pleading "You have to bail us out - we're too big to fail!"

[abend0c4]

One of the problems to obsess over is the circular nature of the investments with compute-providers heavily committed to the slop factories that buy their services. The CDS spreads are starting to tick up, particularly, it would appear, in the case of Oracle - whose shares Ellison père is offering as collateral in the attempted takeover of Warner Bros. by Paramount, headed by Ellison fils, just one of the ways in which the potential contagion is becoming entwined with other sectors of the economy. Given the febrile nature of contemporary US politics, it could all get quite unpleasant.

[Steve K]

Indeed…

Vendor financing (Enron/Nortel) and the Lloyds reinsurance spiral are parallels one could draw…

Ed Zitron has written on this for a while, and even mainstream media are starting to notice.

[Headley_Grange]

Lottery Ticket

They look back at the dot-com boom and instead of remembering the ruined companies and disappeared pension funds they see Google, Facebook, etc. and the billions they could have made if they'd invested with the benefit of 20-20 hindsight. So they're hopping on this bandwagon and figuring that that a tiny probability of making billions is worth a punt today.

[spireite]

Bang, and the dirt, sorry, AI is gone

Like many, my employer is trying to leverage AI for extraction and summary stuff.

The problem the industry has with this whole AI thing is treating it like the messiah, or should that be messiAIh.

At some point, this will come crashing down.

[[e-mail]]

Business guys making technical, enterprise architecture decisions.

My years working as a consultant working for major corporate customers taught me the following:

1) The corporate business guys who sign off on technical decisions have little knowledge or interest in technical issues.They delegate to business guys down the chain of command who also have little knowledge or interest in anything other than moving up in chain of command.

2) At the bottom of the chain is the least experienced business guy who listens to salesmen. The best salesman wins. The decision seldom has anything to do with "what is the best technology" or heavens forbid, "what is good for the business". Instead it generally depends on salesman like-ability. Also, new stuff is more fun to sell. Also, if the junior guys is the first person to champion something new they will have a leg up on moving up the chain.

3) Once top business guy has made a decision based on the decision made buy the most junior guy, admitting a mistake is unthinkable since it might result in losing his position in the organization. Blaming the mistake on the junior guys doesn't always work. This is why projects that are failing continue for long after it is obvious that they are a waste of money. As long as the dead horse continues to be beaten, the top business guy keeps his position - and maybe has time to find somewhere else to go.

There is no chance that corporate consumers of AI will admit their mistakes prior to the bubble bursting. The most they will refuse to pay for price increases for something that isn't working.

[TimMaher]

Re: “dead horse”

If I am a bestial, necrophiliac, flagellant… am I flogging a dead horse?

[Pickle Rick]

Re: “dead horse”

I'd agree, and similarly posit that a slight inclination of the cranium is as adequate as the spontaneous closure of one optic to an equine quadruped devoid of all visionary capacity!

[Anonymous Coward]

Re: “dead horse”

You started it it, so : "I used to be a necrophiliac until some rotten c*** split on me".

_{Split:to snitch; to inform upon; to grass.}

[amanfromMars 1]

If only you could believe the sad rad bad mad tales which are being told ...

How very interesting not, although so understandably so, that the news cycle and its mega media operations might suggest investment and interest have outpaced technology and society whenever currently such is only truly considered by that and/or those in the know as experimental speculative seed funding.

Get with the NEUKlearer HyperRadioProACTivIT ProgramMING, Ladies and Gents, or reap the whirlwind of change and abject depression and recession that accompanies the decay of processes paralysed and stagnating in the maniacal offices of failed state administrations.

[Anonymous Coward]

Re: If only you could believe the sad rad bad mad tales which are being told ...

Yeah, it's all polar nights and AI Winter Blues imho ... a tad of light therapy here, vitamin D there, outdoors exercise, and voilà ... good as new! ;)

[amanfromMars 1]

Re: If only you could believe the sad rad bad mad tales which are being told ...

Yeah, it's all polar nights and AI Winter Blues imho ... a tad of light therapy here, vitamin D there, outdoors exercise, and voilà ... good as new! ;) ...... Anonymous Coward

And/But a great deal better than just new this time around again, AC, whenever being led so completely and strangely different? :-)

[IsSoftware]

The AI Infrastructure Boom: Are We Building on Solid Ground?

As a consultant, I believe no infrastructure—whether software or hardware—should be built without a clear, measurable business case. Yet today, much of the AI infrastructure race feels driven by sales momentum rather than strategic necessity. History reminds us that periods of ballooning tech investment often promise transformative returns but deliver far less.

What concerns me now is the scale and interconnectedness of this boom. Unlike past tech bubbles, today’s infrastructure spending and cross-corporate contracts resemble the systemic risk we saw in the 2008 housing crisis. If this bubble bursts, the fallout won’t be limited to tech—it could ripple across global financial markets as liquidity issues unravel.

The question we should all be asking: Are we investing in sustainable value, or inflating a self-sustaining illusion?

[Artagnan]

Could the phenomenon of AI be the result of a generational problem?

Modern computer scientists find it difficult to resist the siren call of marketing and are accustomed to accepting what they do not understand...

This is not how the foundations of their dreams were built!

[MachDiamond]

So far

It seems to me that most AI is throwing everything at the wall to see what sticks. I'm not seeing companies picking a good use for AI and working towards models that's specific to that task. If I am working on a piece of electronic gear, a smart assistant that can track voltages and signals and let me know if something isn't what it should be could be handy. It would speed things up if it could show me nodes on a image of the PCB so I'm not spending too much time trying to find frickin' C47! If there's also a repository of issues reported, problems found and repairs that resolved those problems with a cost breakdown and source of parts, even if that source is a donor board since Apple won't sell me parts, I could be persuaded to use AI. Until then, it can bite me (at no charge, of course).

[Pickle Rick]

...trying to find frickin' C47!

I think there's a C-47 at the Imperial War Museum if you're having trouble...

[Mine's the one with the Jane's Guide in the pocket!]

[MachDiamond]

Re: ...trying to find frickin' C47!

"[Mine's the one with the Jane's Guide in the pocket!]"

OMG, what size pockets do you have?!

I had to make custom bookshelves to fit the ones I have.

[spireite]

Re: ...trying to find frickin' C47!

That is his pickup line, trouble is brewing if the bird isn't called Jane

[Pickle Rick]

Re: ...trying to find frickin' C47!

"Did you just call me 'Jane'?"

"Yeah, it's my pet name for you..."

"No, it isn't."

>Feck!<

[Pickle Rick]

Re: ...trying to find frickin' C47!

I do have one of these. Granted, it's not a "proper" Jane's guide, and it would just about fit in some pockets, although one's stride might alter slightly!

[MachDiamond]

Re: ...trying to find frickin' C47!

"I think there's a C-47 at the Imperial War Museum if you're having trouble..."

I've got a bucket full of them in the garage.

{C-47 is shorthand for clothes pins (clips) in the movie industry}

[Pickle Rick]

Re: ...trying to find frickin' C47!

To crib directly from you: OMG what sized bucket do you have? Or garage for that matter!

> {C-47 is shorthand for clothes pins (clips) in the movie industry}

I did not know that, cheers! The etymology is interesting too: unknown, but I'd like to think the 'correct' origin is the reference to the Skytrain/Dakota.

[ecofeco]

Slight correction

The war cost $4 TRILLION+

https://www.washingtonpost.com/world/national-security/study-iraq-afghan-war-costs-to-top-4-trillion/2013/03/28/b82a5dce-97ed-11e2-814b-063623d80a60_story.html

https://costsofwar.watson.brown.edu/sites/default/files/papers/Crawford-US-Costs-of-Wars-through-2014.pdf

Other than that, honestly good article.

[ecofeco]

It not going to end well for the tech douche bros

https://www.digitimes.com/news/a20251223PD203/alibaba-data-adoption-qwen.html

China has reached parity.

https://www.digitimes.com/news/a20251223PD203/alibaba-data-adoption-qwen.html

The kicker? Their models are cheap to free and many can be run without a huge data center.

[EdSaxby]

Trojan Horse?

Does anyone other than me think that AI is just a trojan horse?

The tech oligarchs all can clearly see that AI's revenue stream is abysmally low, but yet they are continuing to build out a massive amount of data centre capacity.

It's also not coincidental that the Palantir, Oracle, Meta etc are all snuggling up to an authoritarian government who pays little regard to the niceties of personal privacy and the law, and would bail them out in the blink of an eye.

Welcome to your future... that capacity will get used for every piece of data on you... AI or no AI.

[Winkypop]

Imagine the future

If you want a picture of the future, imagine a boot stamping on a human face - for ever.

Orwell was right.

[CapeCarl]

Re: Imagine the future

"Love is War"

"AI BS is Truth" // 2025 update...Well "Double Think" does sound binary, but with both values in play simultaneously

[rogerroger]

Tech debt

Like most failed digital transformations, which is most of them, simply giving operational business teams a load of money to improve/optimise their own existing process with a new tech, just sinks money into tech debt and business complexity.

Nestle doesnt grow by giving the box packers, mould stampers, and quality control counters a load of money to improve their production line station. It designs new products which open up new markets of customers, or it designs a transformational new end to end manufacturing process which enables it to create and deliver faster and more efficiently.

The key word here is Design! And it's something that most businesses are very lacking in, at least inwards facing, even if they are imvesting in some UX for customer experience. Getting Sales, finance, HR teams to come up with AI solutions is just a waste of money.

[theblackhand]

Re: Tech debt

I thought Nestle grew by buying up water rights in areas they operated and charged everyone excessive pricing because "the idea that water is a human right "extreme." (Quote from Nestle CEO Brabeck-Letmathe in 2019).

Having said that, it is a useful comparison because the big 7 seems to want a monopoly on AI and by extension, water, electricity, data centres and leading edge semiconductors and the rest of the world doesn't matter.

[retiredFool]

The thing that could put a fork in AI

I'm surprised no one has thought of it. The epstein file release I think is now at over a million documents. No one individual is going to be able to digest that. But AI could munch it down in what an hour. And then congress could ask the AI questions such as, AI, give us in descending order the "friends" of epstein who are pedophiles. I expect AI would spit out a rather embarrassing set of names and that list would terminate AI development/usage in at least the US and UK. I imagine even more embarrassing questions could be asked. I'll leave that as an exercise to the readers.