Fizzy Machinehead Sora Pop

I almost wrote “Sora-poop.”

Here, “Can local data centre protests pop the AI Bubble” is a short from TheNerdReich that was prophesying a bubble collapse for data centre finance. I had to be slightly contrarian, just for the helluvit, but also because I suspect there is something here for one or two readers, like [ANGTraders](@Ang Traders ), [ProfKeen](@Steve Keen) and Ghost to comment upon (not to me!, to your own readers).

I would not bet on a bubble collapse so fast. When most funding is Federal (keystrokes) there is no “pop” until the Federal government turns off the spigot. (It is not “ tax payer funded” — it is the exact opposite, except this funding funds only the techbro tax evaders.) Same thing as the 2000 NASDAQ crunch, with Clinton-Gore running a Federal surplus = non-government sector deficit.
      To maintain sales to pay loans the Valley Machineheads only need to raise their rents. There is no need for the endless growth of data centres for them to profit, that dynamic is a floating-point-operations rate arms race among the big-tech firms. They will tend to eat themselves up until one or two big firms have the monopoly. Their bubble will thus not have collapsed. It is a self-sustaining fizzy drink when Congress is backing the spend, but only the biggest bubbles survive, the little guys collapse.

Be kind … I wrote that in the minute, no second thoughts. There may be nuances I am missing, bit in-the-macro I suspect this is about how things play out. It is a classic way so-called “disaster capitalism” works. It is tremendously wasteful. You have to burn a whole lot of companies in order to wait for the monopolist to arise from the ashes. But is it all that far off the way theorists of Capitalism say things are supposed to work? Not too far off. Their theory will say what is left over is a whole lot of medium sized firms all competing healthily for the consumer’s “tax dollar” (that’s right, the tax payer funds the private sector. Their government scorepoints that might otherwise have been used to extinguish their tax liabilities.) But in the real world the tendency is towards monopoly. Is three to ten big firms colluding & cooperating on price gouging the same as lots of medium sized firms in competition?

(When we do get to such monopoly consolidation in a vital sector (so not underwear manufacture, say) I think it should be practically a biding Law of the Land that the big monopoly firms become properly publicly owned, not by stock-holders, but by the State. Of course, this is impossible under neoliberalism when the State is practically owned by oligarchs. But under some form of actual democracy I think it should be the Law —if we are to permit private ownership at all. Maybe it should all be worker-owned, all firms, without exception, and hence also a binding law, but also impossible under neoliberalism.)

Also, not that I care one iota for naïve tech investors getting screwed by the oligarchs. Maybe do not put your life savings into libertarian fever dream promises? (Why do people still invest in crypto, by the way? What’s that gambling all about? Not enough local Casinos to sink your savings in helping the mafia launder their dirty money?) I just do not want the epidemic of balance sheet degradation to cause my rents to soar as the ᖇᓵⓒᕼ ꕗ𖧥𖨚𖢧𖧥𖦪𖦧𖨚 seek to return to their usual five 60 day vacations a year normality.

Prophecies of Madness

The prophets who were predicting superintelligence have egg-on-face … but still will not admit it. Like cold fusion, or even blistering solar temperature hot fusion, the breakthrough is always “just a decade away.” (To be fair, Ray MachinePerv Kurzweil still has 14 years left.)

Only in hot fusion I think one year it will be true. It is the height of hubris to think we can build a virtual friggin’ star on the Earth. But this is possible hubris. Hence fantastically good science, even if it proves far less efficient than solar + geothermal. Why? Just because is is cool. (Yeah. Yeah. Don’t nerd-scold me. It is not cool if it diverts millions of resources away from the needs of poor communities, ok.)

Abundant but costly (in terms of irradiated waste and embrittlement) fusion energy might have a use-case outside of electricity generation, say in low energy physics research. Yeah, low energy! Plasmas are much cooler than the collisions at the LHC. we are not going to be exploring Beyond the Standard Model with fusion. It is a study of the fourth state of matter. Which has its own beauty for those of magnetohydrodynamical bent.

But that’s beside my point today.

The relationship is one point. There is no bubble in fusion funding, because no one took the idea of building a friggin’ star on the Earth too seriously. So the number of start-ups is small. The amount of fizz is low.

With algorithmic neural-nets it is different of course, because some of the most incredible fraudulent philosophers FN-1 in history have lived in our time — Dennett, Hofstadter, the Churchland’s etc. FN-2 — all predicting machines can think, if given enough ooomph. They know nothing of the human soul, thus it is forgivable their ignorance would spawn the “AI” industry fever dreams of “intelligence” singularities. I could perhaps forgive them if in every article they had published they had added the caveat: “This could all be ꕗꖹꝆꝆꕷꖾꕯꖡ.”

NB: this is not across-the-board. A decent skeptical article like Thomas Nagel’s “What is It Like to Be a Bat?” is so wholesome and skeptical of machine sentience that it needs no “could be ꕗꖹꝆꝆꕷꖾꕯꖡ” caveat. Also, Hofstadter wrote a couple of beautiful books, Le Ton Beau de Marot among the best, better than the nerd-indulgent fake Zen of GEB imho, so he at least should know better.

It is not for me to forgive them though. Who knows how many poor communities their diseased ideas published in innumerable journals will end up by proxy killing? Several million I would guess.

The relationship is that fusion research should know no end of funding. It can continue on shoestrings and breadcrumbs. Slow is fine. We do not need more electricity to live decent lives. We can solve the climate crisis by all cutting back on non-essentials. Same for machine algorithm string emission technology. FN-3

This is what the 𖥐𖦪𖠢𖢧𖥣ꛘ𖥕ꚶ𖨚 Abundance guys do not comprehend. They want neoliberal growthism because they are greedy. They cannot stomach the idea the median USA lifestyle is unsustainable, but that also below the median USA lifestyle, with a few simple changes in habits, a family can be even happier, and less wasteful. It is not rocket science. It is a solved problem that the world seems to not wish to know. You indeed cannot countenance such knowledge when you measure your success against your rapacious next door neighbour. You need to stop using such measures.

(At least first check to see if your next door neighbour with the Tesla Roadster or BMW really is happy. My bet is anything that they think they still do not have enough government scorepoints.)

No Bubble Fizz

I think fusion research is rather flat. Possibly in 10 years time so will machine algorithm fantasies of superintelligence. It is the pain we might go through in the meanwhile that worries me, not the financial collapse.

Nathan Tankus was writing about this recently. But I cannot afford to subscribe to his substack thing Notes on the Crisis , I only get the free tier bulletins. But I suspect his opinion was that provided the Federal government keep supplying the mother’s milk (US$ scorepoints) the techbros will suck it all up. FN-4 But as I wrote above, I am sure the chatBot fizzy drinks will lose some fizz as the monopoly forms. I am sure more than a few Insiders may realize the Chinese platforms will be standing on top at the end of the fizz? Like Japan, I wonder when the bureaucrats in these countries will admit that they knew MMT all along? Knew they had MMT, knew they were lying about not knowing it.

(There is no genie that escapes the bottle when the State acknowledges MMT. What’ll happen, I predict (albeit with no great confidence, being very non- genius) that the public will just be holding the government even more to account! To my mind that is a tautology, due to the inherent deep relationship between money and records of account and public accountability. They are literally the same thing, once you know the money is not generated by 𖦪𖥣𖥐𖦙 ꖿ⋒ℼԞ𝕤. On second thoughts, if the public starts holding the government even more to account, that would be a pretty good genie to let loose! Democracy has for a long time been trapped in a conservative whine bottle.)

There is a great (abysmal) parallel here I only recently was aware of, thanks to Ghost , which was the Weimar/Nazi period MEFO bills , which funded German re-armaments, not too far in time from the Weimar hyperinflation, though well afterwards. It was possible the German monetary authorities were (falsely, imho) worried about another inflation event, and this MEFO bill was thus a thinly disguised veil to avoid the accusation of “printing money.” FN-5

The data centre spend will not be inflationary zeroeth order because those guys love hoarding the scorepoints. Thus zeroeth-order will only want to pay the market price for electricity and silicon. Added scorepoints only inflate prices if they are spent.

First-order it should inflate the electricity and water prices, marginally, because of the arms-race dynamics absent a clear monopoly triumph. OpenAI doing a great job with Altman-Elemental-Kaos undermining their chance for monopoly, see Empires of AI by Karen Hao.

The popiness (sp?) and fizz come from the age-old speculative finance dynamics and nowadays also the techbro arms race. They all wish to be the First to Summit the Everest of Superintelligence.

Someone should reach into their 𝕚𝑑🄸🅾𝕥 dreams Inception style and tell them that mountain does not exist.

While there is this fantasy arms-race towards a weapon that will never exist, there is a ponzi-finance dynamic. But we all now know that there is no balance sheet crisis that cannot be eliminated with sufficient fiscal injection.

In String-Spewing-Algorithm finance terms, the based leftist take on this should be that ꕯꕒꕯꗞꖡ small fry investors who risked their houses and failed to early bank their gamble on the “AI” Boom should not have to suffer their own ꕯꕒꕯꗞꖀꖡ … too much. Just a little. In fair due proportion. They should not need to have to sell their house or re-mortgage, they only need a government guarantee of the same degree the techbro’s had to spawn the ponzi period, namely a Job Guarantee with a decent wage so they can make their old mortgage payments.

I am in favour of a Job Guarantee, but not a job guarantee only for people who already have the retirement savings of 100 median families, aka state subsidized bitstring-farming.

The Other Hype & Grift

A few dirtbag physicist/MMT’er talking points (thanks to Prof. Emily Bender for the (un-cited, but see here ) academic backings):

1. There is no such thing as “AI” so stop using the term. The term itself is hyping the grift. Stop using it without quotes or caveats.
2. The human mind is not merely emitting strings like the LLM’s. Or … to assert otherwise is a form of intellectual fascism, or at best techno-utopianism (bad enough). Stop doing that!
3. We do not have to all be Neo-Luddites. But I am. We do not need half the crap we have to live decent lives. More farming robotics I say, but more local & organic and sustainable. (I know “organic” is a dicey word, but you know what I mean. Glass house agriculture is fine if needed to feed the masses, but should be a last resort, due to the nutritional deficits.)
4. Even “neural net” is a bad phrase, I try to avoid it and stick to the acronym ANN/CNN.FN-6 But obviously would prefer to avoid that too. The computer systems are nothing like biological neurons. The computational node architecture is so far abstracted from neurobiology it might as well be fairy dust. It certainly has engendered Peter Pan like fantasy.
5. “Reinforcement learning” is a bad phrase in algorithmic string emission context. There is no learning involved, not of the kind that occurs in the human mind. Weight updates are not “learning” in any way shape or form. It is yet another insidious neoliberal use of language. Every ounce as bad as “debt mountain” and “fiscal cliff.” Science does not even know what human learning is exactly. There is no good definition. We only measure behavioural input/output events and correlations. These are a poor substitute for the actual spiritual phenomenon of learning and spiritual growth (learning to be a better kinder person).FN-6
6. The headlines, “AI Solves/Proves Hundred Year Old Math Theorem” are also fake headlines. Some dopey math nerd had to prompt the algorithm to get it running. The math nerd is where all the intelligence resides. Plus the software engineers who wort the algorithm and fed the algorithm ginormous amounts of mathematical results. To say the machine “thought” of the proof is like saying the lottery ball machine thought Joan Ginther would be a worthy winner.FN-8
7. The human soul is not understood by any science. Science cannot even agree on whether there is such a reality. It certainly never helps that no definition exists.

Like usage for the abused term “hacker”, I quixotically must forever run against the grain. ‘Hacker’ means “playful inventiveness” and has nothing in particular to do with computer coding. The term for a malicious computer security break is “cracker,” — which in turn has nothing to do with the social class people think of as “white trash”.

Like usage for the abused term “hacker”, I quixotically must forever run against the grain. ‘Hacker’ means “playful inventiveness” and has nothing in particular to do with computer coding. The term for a malicious computer security break is “cracker,” — which in turn has nothing to do with the social class people think of as “white trash”. … I forgot the point of this remark … something about how despite materialist scientists broadly dismissing the reality of the human soul, they have no definition of what they are dismissing.FN-9 Sheer folly.

Footnotes: