Blogstats in a year of intense LLM "training"

Last year I noted the Top Six locations for C@W hits, as follows (in descending order):

Hong Kong / China / USA / Singapore / UK / Norway

2025 has been a bizarre year for blogstats.  Much as I'd like to think the upsurge represents long overdue global recognition for this blog, many other blogs have found the same and the widely accepted explanation is that the latest generation of LLMs, ever hungry for new "training" material, have been voraciously "reading" new bodies of text.  Obviously, the nearly two decades of flawless prose and compelling reasoning to be found on C@W make us a highly suitable educational experience for these eager students.

Anyhow, the "readership" has increased six-fold over 2024, with a mighty spike in June.  Someone in the LLM industry could doubtless explain this in detail.  Interestingly, while the spike has long since receded into the rear view mirror, it has left a pronounced tail in its wake**.  Presumably, the said LLMs keep coming back periodically to check on our latest gems: and I wonder if also their initial burst of reading resulted in a wider dissemination of C@W as a cited source, which ordinary ("human") www-browsers now access more than previously.

Anyhow, here is the Top 10 for 2025. 

1. Brazil - first by a good distance
2. Singapore
3. USA
4. Vietnam
5. China
6. India
7. UK
8. Japan
9. Bangladesh
10. Norway

Questions:

• Does this mean that electricity is cheaper in Brazil?  Or is it the cost of bandwidth?  I haven't heard of a rash of data centres being built there.

• To what body of hitherto untapped sources in the digitised world will the LLMs go for their next training binge?  What's left that is broadly literate, extensive, ignored thus far - and free?  The complete speeches of Stalin, Mao and Castro?  Might it be something of unspeakably awful content, such that LLMs will soon be effing and blinding like a docker and sharing pictures of nudified politicians ..? 

• May we hope that the wholesome diet of C@W wit, wisdom and literary excellence will raise the whole tone of AI output?  (*ahem*)

ND

PS: start thinking about your 2026 predictions ...

________________

** A company of my acquaintance that publishes a notable blog (as a promotional exercise) has suffered the opposite effect.  Its content having been extensively devoured in the same way, it now gets fewer hits, because AI searches give answers based on the LLM's reading and summarising of that blog.  These days, such searches do at least give links to the original source; but most lazy bastards go no further than the AI summary.  Which, to be fair, characterises much of my own use of AI - given the excellence of the summaries!  (Which is not necessarily a crass, circular assessment: when full reasoning & explanation are incorporated, you can pretty much judge that nothing salient has been omitted or "misunderstood".  And it's self-regulating, too: if the search is for an important purpose, obviously you dig deeper.)      

Some established business models are of course being actively trashed by the phenomenon - not to mention the upcoming fate of many professional, semi-professional and clerical-type jobs ... 

When the King invests in the bubble

Not only do human beings have a marked propensity to believe in perpetual motion machines, nuclear fusion, price forecasts and sure-fire betting scams wheezes, they love investing in a good bubble.  When a bubble really gets going - the classic case study being the South Sea variant - almost everyone with any spare cash piles in, accompanied by eejits who've even borrowed to do the same.  And by 'everyone', we mean (in the South Sea case) the King, the Court and, for good measure, Isaac Newton.  Yep, the suckers' propensity has little to do with brainpower or status.

So now we learn that a Trump family enterprise is getting into nuclear fusion, with the AI bubble in mind and the correspondingly vast putatively future demand for electricity.  Has the Orange One finally taken leave of his senses?   Well, note that the story says "the move would 'create one of the world's first publicly traded fusion companies'" which speaks more to making a killing on the share price rather than actually producing anything tangible like, errr, electricity.  So maybe it's quite canny.  What's more, who'd be surprised to learn that the Trump "50% share" is in the form of carried equity?  I'm just guessing here: but after all, his big contribution is surely to act as the pedlar of the snake oil** to his millions-strong legion of sucker-followers for whom he can do no wrong.

Needless to say, the announcement goes on to claim: "the combined company planned to begin constructing the 'world's first utility-scale fusion power plant' next year, with further plants to follow".  Well of course it does: what else would they say?  "Starting construction", as any dodgy building firm will tell you, doesn't need to mean anything more than clearing a site.

Publicly traded, hmmm ... does this mean we'll be able to short the Hell out of this thing?   

Merry Christmas to all !

ND

__________________

** The so-called Swansea Bay Tidal Lagoon, a dreadful scam that ran 2012-2018 before HMG finally put it out of its misery, provides a sobering lesson on such matters.  The promoter, one M. Shorrock, was evidently very good at getting people to part with their money.  Some of his investors were high-rollers (who definitely ought to have known better, shame on them) but he also persuaded hundreds of small investors to take a punt, frequently on the heartwarming pitch that it would provide lots of jobs and cheap electricity for the benighted town of Swansea.  In this way he raised, and spent, somewhere between £37m-£50m before the scheme went bankrupt.  You won't be surprised to learn that many of these millions went to others of M.Shorrock's companies in the form management fees and loan interest paid at 20%.  He has a very nice house in Gloucester.

Oh, and for his own shares in the enterprise (he held approx 23%), how much did Shorrock himself pay?  Answer:  £70.  Not £70,000.  £70 (seventy).  PS, had the scheme ever reached financial close (which, had HMG given it a 'green' subsidy, it probably would have - the banks were lined up), Shorrock would have been paid £14m.

It is not known whether he currently acts as financial adviser to D.Trump.  Possibly not, since he was last sighted selling solar panels in, errr, Vietnam.

Miliband & the perpetual belief in perpetual motion

There's something very deep in human psychology that encourages people to believe in magic swords, universal elixirs, price forecasts, perpetual motion machines etc etc.  It's out there - and it just needs that little bit more human ingenuity to bring it to reality.

Ed Miliband is (we are told) a human, and clearly suffers from this syndrome.  Here are some of the things he believes in:

• a "net zero electricity grid" by 2030
• nuclear fusion as a practical source of power
• "negative CO2 emissions" from burning trees for power and burying the CO2
• hydrogen as a wonder-fuel of very wide practical application
• himself as a dominant force in a post-Starmer UK government

A man can dream: but he's spending our money on the first four of these (and indirectly, one might say, on the fifth).  In a harder-headed world, they would disqualify him from the fifth.  

Sadly, I don't think they do.    

ND

Enron's plan for property-price derivatives market

Canary Wharf: cornerstone 

Following on from the post about the 'predictions market' & how various attempts to make financial markets in superficially prospective areas have sometimes come unstuck (water; bandwidth, weather): I'd mentioned that just before the Big Collapse, Enron was planning a property-price derivatives market, meaning futures / forwards at the outset, and ultimately options.

The rationale for there being demand for such a thing was this.  Many individual and commercial entities, as well as outright investors, can find they have a lot at stake as regards the variability over time of property prices in general, and the differences between property prices in different regions (technically, a source of 'basis risk').  Simple examples at the personal level: someone who hasn't yet sold their current property but has committed to buying a new one - needs a hedge against prices dropping while they find a buyer.  Someone who needs to move from London to Manchester for a couple of years but expects to return to London thereafter: needs a hedge against London prices outstripping Manchester over that period.  Someone who wants to lock in an attractive price they've seen the identical house next door fetching when it sold last week, but doesn't plan to move just yet: needs a hedge against local prices falling.  Etc etc etc.  

And of course once a market is established, speculators and punters can pile in: unlike weather (see previous post), people often really do have strong opinions about whether the property market is overheated or underpriced.

So how was Enron going to get the show on the road, back in 2001 at the time of the Collapse?  They put some of their best people on it.  Regionally specific price indices already existed - the sine qua non for derivatives.  Key to any market is liquidity, in turn requiring market makers and critical mass: and, with some aspects of derivatives, the ability to cash out into the physical.  They had a strong relationship with the Halifax (then a big property player and publisher of indices) and planned to start with the London commercial (office space) sector - and to ensure physical delivery, as an opening gambit they were going to buy Canary Wharf !

Sadly, we will never know how this would have panned out ...

ND

Kalshi billionaires & the "predictions market"

As fans of open markets, generally speaking, what do we make of the soi-disant "predictions market"? in which the firm Kalshi has made paper billionaires out of its youthful founders (one of which is the youngest ever female paper billionaire).

Well, first of all, good luck to them: I assume their investors are consenting adults, hopefully with a brain cell or two to rub together.  Investment bubbles help make the world go around: where would we be for railways if there hadn't been an "irrationally exuberant" railway-mania boom in the 19th C?  Just so long as it is private cash, hopefully not pension money, that someone can afford to lose 

But secondly, this is just a slightly exotic betting platform with good PR, right?  And will they ultimately prove correct in that it's possible to "monetise every difference of opinion"?

A story.  At the back end of the 1990s when Enron had triumphantly succeeded (against strenuous opposition) in becoming a highly profitable market maker in all manner of commodity contracts that nobody except them thought was remotely possible - indeed, some academics declared was a priori impossible - a lot of people in the energy sector and much-larger financial / insurance sector were looking for the Next Big Thing to commoditise, create derivatives on, and trade.  I could tell you an amusing but irrelevant story about the attempt to do this with water: and at the time Enron went bust (2001) it had a clever plan, never realised, for doing it with property prices.  But the one that's relevant here is weather derivatives.

It's clear enough that a great number of companies and individuals often have something of value to them that is riding on what the weather will be.  Ice cream manufacturers, energy companies, holiday concerns both buyers and sellers, brides-to-be, county fairs etc etc.  They often buy insurance, from long-established specialist insurers.  So what?

Well, in principle we can identify all the technical factors required to establish financial derivatives.  Financial issues at stake in very large measure, impacting on a huge plurality of players.  Gains and losses related directly to uncertain but measurable outcomes.  Data on said outcomes that are objective and not open to manipulation or influence by the prospective winners & losers.[1]  So the weather-related derivative products were devised and a very large number of players from several sectors piled in with traders, marketing teams, software, complex stochastic analyses etc etc - all the paraphernalia of traded markets.  All these costly resources created, they sat back and started trading - mostly with themselves.  but surely, the genuine "natural counterparties" / end-users (entities with something at stake), the 'locals', the outright punters, would come along in due course?  What's critical in any such market is transparency (easy), clearing & systems, but above all else, LIQUIDITY.  

Well, nope: despite this huge global investment the weather derivatives market market never took off on anything remotely like the predicted and expensively planned-for scale.[2]  Why?  Because no bugger has any view whatsoever on what the rain is going to do next August 13th at noon, in any regions other than the Sahara or Antarctica (where there's no end-user business to be had).  Not even inveterate gamblers of the Sky Masterson variety.  Somebody may well have a horrible exposure to rain at that precise time, but no broker or market maker will be able to find anyone willing to take the other side of the bet - when there's a perfectly good insurance-based alternative.  (For those interested in the technicalities of financial risk management, insurance is a completely different paradigm to hedging, with different applicability.)

Why is this different to, say, the price of oil on August 13 next year?  Because there are very large numbers - sufficient numbers! - of people willing to take the other side of almost any number you care to put out there.  If I say $50/bbl, there will be loads of people with theories that say it'll be less, and loads who'll say it will be more.  The proof of this is the liquidity of the forward market for oil.  And any number of other commodities, as well as financial variables altogether more abstract.  Ditto big sporting events.

But weather?  Nah - or so it turns out.  Could this have been foreseen?  Some folks did, and saved themselves a bunch of time, money and effort.  But no so many.  Meanwhile, the perfectly healthy weather insurance market continues on its merry way.

So: a market in differences of opinion?  Well, sports and political betting are well known to be ultra-fertile ground for these things[3].  But if Kalshi is to break into new territory, it seems to me it needs to think of something a bit more interesting than "Will Trump attend another UFC event this year?" - and a heap less open to, *ahem*, manipulation.   (Guess who might be taking the other side of that bet?!  I think we can fairly be blunter: that kind of thing is really, really crass.)  

Nobody can rule out Kalshi actually finding a rich new seam of prospective punting.  But (a) this is a crowded field, and (b) it won't be hard for a load of other players to pile all over it.

Did I say "paper billionaires"?  They'd better try to cash out PDQ - as regards young(ish) female billionaires, I suspect Taylor Swift is the more solvent ...

ND

_______________

[1] I am not interested in entertaining conspiracy theories about how weather data is systematically manipulated  by the Deep State or the Green Blob or whatever.  Obviously, forecasts are forecasts, and weather forecasts are substantially better than forecasts of commodity price & other financial variables forecasts which are seriously manipulated by vested interest all the time. 

[2]  Some AI disputes the notion that it never took off, suggesting it's a $25bn market.  That's derisory.

[3]  Is it unkind to laugh at Alasdair Campbell's son's sure-fire sports-betting syndicate?