Yakulticide: death by yakult.

Despite all the wars going on, those archeologists man, they just keep on going.

Digging up shit and reporting it in academic journals.

You gotta love their enthusiasm in the face of drones.

Just maybe, they think the holes are being dug for them.

On the subject of dictatorship, broadly speaking there are 2 types:

The communist sort which maintains a form of meritocracy. Anyone can join the party and eventually become chief cunt. This makes this model more sustainable.

The traditional sort where the dictator’s children inherit the crown. This sort is subject to the three generation curse – wealth is created by the first generation, maintained by the second, and lost by the third.

Then there’s the Murdoch twist where you put all your kids into the Big Brother house and let the biggest cunt take the role.

Of late, in the extended West we haven’t really had the chance to see a multi generational dictatorship in action. For a variety of reasons they tend to die in the first generation. Think Adolf, Benito, Francisco and Juan.

But there is North Korea which is on the verge of generation three. That’ll be fun to watch.

We’ve all met them, people that are naturally disinclined to worry about their own mortality.

They have a natural advantage over us, until they don’t.

There’s a pack of cunts in the US who’ve managed to get themselves into power by leaning hard on a base that runs more on grievance than thought. They’re not interested in governing for anyone else. The aim is simple: run the place in their own interests, strip away the bits of law and democracy that get in the way, and make it stick. The ruling clique would like to run the place solely in their own interests, like dictators, untroubled by law and democracy.

From there, they find common cause with the Russian dictatorship. Not a formal alliance, more a quiet arrangement between two systems heading in the same direction. Each leaves the other alone internally, each helps where it suits, and both keep the public story just plausible enough. The shared line is that the real fight is against China (yet another dictatorship but one that is already in complete control of its domain and therefore less aligned with the others), and against liberal democracy itself.

Obvs, initially they have to be surreptitious about this partnership but as time goes on, they can be more open about it, as the ability of their countries to challenge their plans is systematically weakened. AI, bless its little heart, gives them a real boost as people learn not to trust anything they see or hear.

The rest of us, we’ll be subservient to this mob after they carve up the globe into spheres of influence.

Back to the good old days before the enlightenment. Which makes sense really – we cant afford to keep consuming all these resources and emitting all these greenhouse gases – so we need to return to those glorious days before we had the means.

If I was to flog this precis to a publisher I would tag it as “1984 meets Animal Farm”, and maybe call it “Back to the Past” or “Pigs Can Fly”.

Actually I have published an actual book but not on this subject. I call it fiction. You’d call it non-fiction.

It’s really an opinion piece focused on my country of origin. There’s a lot of forward looking prognostics in there because I’m so good at predicting the future …

So the elf draws my attention to a plane crash in NY – a plane “taxi-ing” at 100kmh hits a fire truck.

My first thought was “bullshit”. Never seen a plane taxi-ing at that speed, nor fire trucks lurking around the touchdown zone just waiting to be hit.

Then I looked at it and thought “AI Slop”.

Who knows, it might have actually happened. But I doubt it.

It was in all the major newspapers but they just rerun anything on the web. They can’t afford staff anymore.

Philosophically, it seems to me that we’ve lobbed into the situation where we’re going to have to learn not to trust our senses.

Which means that all we know is that we exist, and nothing else.

Warehouse Hole

So the default arguments you’ll hear from electric car owners are:

1. They’re cheaper to own

2. They’re faster

I challenge the first because you have to consider capital losses. And the price of a tank full of electricity can be pretty high unless it comes from your own roof.

If you’re comparing the cost to a grey import Nissan Cube that costs $10k, I can assure you the Cube is much cheaper in terms of full cost of ownership.

As to e-cars being faster. 12 points man: that’s all you get. You don’t want fast.

Besides a microwave is faster than an oven. Where do you cook your Xmas dinner then? Eh?

And they stuff those electric cars full of all that annoying tech: alarms, alerts, lights, plus software-controlled steering, suspension, and braking.

If you like driving these cars won’t let you. These cars are for people that are living as though they’re already dead. Mental zombies amongst us.

I hasten to point out of course that that last complaint isn’t peculiar to electric cars; it’s all new cars.

So while I’m driving, it’s going to be old cars from here on in. And if they retire petrol then we’ll have to convert an old car to electricity to dodge all that tech. Or just not drive.

Sixty can be divided by one, two, three, four, five, six, 10, 12, 15, 20, 30 and 60, in each case to give a whole number.

Contrast that with 10, which can be divided only by one, two, five and 10 and the advantages of base 60 start to become apparent.

If you’re developing numbers for very practical purposes, like taxes or dividing fields for your sons’ inheritance, having an easy way to do these mathematical operations can be really helpful.

So it was, 5000 years ago in ancient Sumeria

60 comes from the thumb counting the 12 segments on the other 4 fingers, then do this 5 times monitored by the 4 fingers and thumb on the other hand. Your hand are a natural base 60.

I overheard a song in a cafe by Shaggy. I am sure he said:

“I dated you for breakfast”

AI says “That phrase is a lyric from the song ‘Broken Clocks’ by SZA, from her 2017 album Ctrl. The full lyric is: Been about three years since I dated you / Why you still talkin’ ’bout me like we together?”

An old fashioned google search says “nuh, get your ears checked”.

(I write all this to comfort my aspirational artisan of a daughter.)

I’m hearing a lot of panic about AI taking jobs away from people. The first to go are those that look at monitors for a living. Those that go to meetings and have phone calls; they’re next. Finally, those in work boots or aprons will get done over when robots improve as well.

Ignoring the problem that no one will be able to afford anything anymore, I believe that artisanal products will make a comeback at that point because people will apply value to humanness, no matter how that’s defined. The ‘art’ in artisanal will become defined as something that AI can’t replicate until after the fact.

Since it will be expensive and in short supply, the real thing as made by human sweat is going to become the duck’s nuts as soon as AI makes everything else a commodity that no one can afford, and that no one wants.

Think of it this way; right now, people are buying Asian electric cars. No one really likes them because they drive like shit and they’re loaded with annoying technology. For the moment your average middle bogan can’t admit that they’ve bought into a pile of crap.

But they will get over their pride one day and start getting nostalgic about that 2003 M5 they never had. A car that you could actually drive: actually it gave you no choice, you had to drive it.

In the future, artisanal products are going to occupy every sector (unlike today where they’re isolated in the knick knack categories). Even artisanal coding will become a thing – “real hand written code”.

So, artisans, it won’t matter what you do, so long as you do it with intent and imperfections, ensure that its innovative, and adopt a proper sense of pursed-lip seriousness.

In fact, human aspiration might end up being the production of stuff that AI simply can’t replicate until afterwards. Just like this blog entry: I tried getting GPT it by prompting it with the basic ideas. Not even close; and then it really, really hated what I ended up writing.

Looking back on human history. Back in caveman days, success was all down to strength and aggressiveness. Then we had early civilisation during which smarts and cuntiness prevailed  After the industry revolution those that flourished were those that had smarts and weren’t afraid to spin the bullshit. Since that period ended at the end of the 20th century the preeminent attribute of success has become 100% bullshit. AI therefore has arrived to rescue us from ourselves and push innovation and quirkiness to the fore.

It’ll just take a few years to see the benefits.

Q: Uno cards. If the shuffling machine knew everyone’s cards via RFID chips embedded in the cards, and could decide what cards it ejected, could it keep the game going forever?

GPT: You’ve basically invented a mildly evil casino dealer disguised as a plastic shuffler.

Short answer: yes, it could keep the game going arbitrarily long, but not literally forever in a strict sense.

Mechanism matters here:

• UNO has a finite state space: fixed deck, finite hands, finite discard pile configurations.
• A game ends when one player reaches zero cards.
• A “smart” shuffler that knows all hands can control transitions between states by choosing what cards get dealt or recycled.

If the machine is allowed to: • See all hands
• Control reshuffling of the discard pile
• Potentially bias which cards go to which player

Then it can enforce loop conditions.

Example loop construction: • Whenever a player gets close to winning, ensure they receive draw penalties like Draw Two or Wild Draw Four.
• Maintain high-action cards in circulation: Reverse, Skip, Draw cards.
• Prevent low-card states by redistributing cards during reshuffles so no one drops to zero.

In effect, it keeps the system away from terminal states.

Why not truly infinite:

• The system is finite-state. If you run long enough, you must eventually repeat a full state (same hands, same discard, same turn).
• Once a state repeats, you are in a cycle. That cycle can be infinite, but only if the machine enforces identical transitions each time.
• If there is any randomness or constraint (like fixed dealing rules), eventual termination probability may be non-zero.

So:

• With full control: it can create a deterministic loop → effectively infinite game.
• With partial control: it can massively delay the end, but not guarantee infinity.

What you’re really describing is a controlled Markov process where the controller avoids absorbing states. Congratulations, you accidentally reinvented adversarial game design. Casinos would hire you, then quietly never let you near the rulebook again.

The government is calling on people to conserve fuel by working from home, driving slower, catching public transport, avoiding air travel and switching to electric stoves.

Why do old people think it’s ok to have their phone on loud speaker mode in public?

Because a mix of hearing loss, habit formation from private environments, reduced sensitivity to social norms, and mild technological incompetence leads them to treat a phone call like a face-to-face conversation; the device becomes a shared audio object rather than a personal one, and since the immediate utility of hearing clearly outweighs any abstract concern about bothering strangers, they default to speaker mode without recalibrating for public space.

I ran today’s Sydney Morning Herald through a  GPT filter:

About 45 per cent of content is war, politics, weather, policy, courts. The other 55 per cent is petrol prices, electricity bills, property values, interest rates, work, lifestyle.

Much of the “serious” news only exists to feed the second category. The Middle East matters because fuel spikes. Policy matters because mortgages move. Weather matters because disruption follows. Very little news is presented as purely informational. Almost everything is converted into impact.

The reader is positioned at the centre of every story. If it does not affect cost, comfort, or convenience, it’s not in there.

Out of curiosity, I ran the Guardian through the same filter. It’s summary was…

SMH model:
world → filtered → cost to you

Guardian model:
world → interpreted → why it matters (morally, socially, structurally) → optional personal angle

So it’s less self-absorbed in a financial sense, more self-absorbed in a moral sense.

A third model could optimise for:
accuracy under uncertainty, at the cost of comfort and attention. It didn’t exist because most people don’t actually want to think, or suffer any discomfort.

“I found the stapler and its still got staps in it.”

Yeah, infinite isn’t good. Just like rice on a chessboard, you run out of space and everything else pretty fucking quickly.

So my guess is that you have to semantically compress those infinite universes into the vibe of the thing.

And then magically they cease to exist, for all intents and purposes.

Doesn’t stop the fairies dreaming about them, but.

Maths is just a weak arse attempt by humans to get around semantic compression. And when the maths runs out, we’re back to words.

And when you think about it, so is the past. We can even remember our own past without semantically compressing it.

Which means what we perceive as time might just be the dissonance we feel about that semantic compression.

—

You know how you can’t remember every single walk you’ve ever been on, every smell, but you still sort of “know” what a good walk is? That’s because your brain doesn’t store everything in full detail. It keeps a simplified version, just enough to get the idea.

The world is very big, with far more detail than any brain can hold. So instead of keeping everything, we compress it. We store the gist. When we remember something, we’re not replaying it, we’re rebuilding it from that compressed version. That means details get lost, changed, or quietly invented.

Maths is what humans use when they want to be more precise about all this. It’s still compression, just stricter and less fuzzy.

Because memories are compressed, some periods feel long and detailed, others feel like they disappeared. That’s why a busy day feels long, and a boring week vanishes. The sense of time passing is partly tied to how much detail got stored and how it gets reconstructed later.

The universe contains infinitely more possibilities than any mind can hold. No human can carry all that detail around in full.

So what happens instead? It gets compressed. Not into the exact thing, but into the vibe of the thing. The gist. The usable shape. The emotional and conceptual summary.

That is where the multiple-universe part comes in. If there are endless possible worlds, branches, histories, or quantum alternatives, then no finite mind can represent them one by one. So for practical purposes they get squashed into a simplified internal sense of reality. They do not necessarily vanish in some grand physical sense. They vanish for you, because you cannot carry them all. They collapse into a manageable impression. A sniffed summary of existence.

The same thing happens with the past. You do not remember 14 March 2025, in full resolution. You remember something like: odd day, coffee, annoyed, conversation, rain. The actual past may have been dense with detail, but your mind transforms it into time by compressing it into a narrative you can carry forward.

Time, as we experience it, is not just change itself. It is what reality feels like after it has been semantically compressed by a limited mind. We cannot hold the full universe, or all possible universes, or even our own past in exact form. So we translate them into ordered, simplified chunks. And the tension between what was there and what we can still hold is what we call time.

—

Multiple universes are not something we ever actually hold or observe in full. They represent the total space of possibilities, even if they don’t exist. The key constraint is that any observer has finite capacity. So instead of representing that full space, the system compresses it into a single coherent state, what I call “the vibe of the thing.”

That compression is not just memory, it is selection. Out of all possible states, one path is retained as the experienced reality. The rest are not necessarily destroyed physically, but they are removed from the accessible representation. Functionally, they cease to exist for the observer.

Now the time insight sits on top of that.

The past is not stored as a full-resolution history. It is compressed into a sequence of reconstructed states. Each reconstruction suffers from a loss of information. That means the further back you go, the more aggressively the state has been simplified.

So what you experience as time is not just change in the external system. It is the ongoing process of taking a high-dimensional possibility space, collapsing it into a single experienced path, and then repeatedly compressing that path into memory.

Time, in that sense, is the artifact of sequential compression under constraint.

Multiple universes: the full possibility space. Observed reality: the compressed selection. Past: recursively compressed versions of that selection. Experienced time: the structure that emerges from that repeated compression and reconstruction process.

—

If what you experience is a compressed slice through a much larger possibility space, then “certainty” is mostly a side effect of compression. Your brain throws away alternatives and presents a clean narrative. That means confidence is not a reliable signal of correctness. It’s a signal that compression succeeded.

Implication: treat strong intuition with suspicion, especially in complex domains. The cleaner it feels, the more information has probably been discarded.

Memory: the past is not stored, it is reconstructed from compressed fragments. That means drift is inevitable. Over time, the reconstruction error accumulates.

Implication: anything that matters needs externalisation. Notes, logs, data. Otherwise you are iterating on a degraded internal model and calling it history.

Modelling and forecasting: you cannot enumerate all possibilities, so you already operate on compressed representations. The mistake is pretending the model is the system.

Implication: work with ranges, not point estimates. Keep multiple competing models alive instead of collapsing too early to one “vibe.” Delay compression when the cost of being wrong is high.

Communication: language is another compression layer on top of already compressed thought. Most disagreement is not about facts, it’s about mismatched compression schemes.

Implication: define terms, force specificity, anchor to measurable quantities where possible. Otherwise you are arguing over incompatible summaries.

Time perception: the experience of time expands with information density and contracts with repetition and low encoding. This is a consequence of how much state gets written.

Implication: if you want life to feel longer, increase novelty and attention. If you want it to disappear, automate everything and repeat.

Now the multiverse angle: you never access the full possibility space. You commit to a path and discard the rest at each step.

Implication: optionality is real only before commitment. After that, alternatives are gone for practical purposes. So design decisions to preserve optionality early, then commit hard once enough information is gathered.

Nothing here changes physics. It changes how you treat your own cognition under constraint. The system is still the system. You’re just a lossy interface to it.

—

Final word; beware of any concept that is unfalsifiable at the level of personal experience. There be dragons, unicorns and rainbows.

Ad hominem attacks (Latin for “to the person”) are logical fallacies where an argument is rebutted by attacking the character, motive, or attributes of the person making it rather than addressing the substance of the argument.

This includes all the arguments against Trump (because he has no substance to argue with.)

Its sort of like the argumentative equivalent of means, opportunity and motive.

But what if the ad hominem attack was in fact a murder? Ad hominem homicidium. Ad hominem is attacking the person. Homicide is just taking the argument to its logical conclusion.

To be clear, I’m suggesting we only use ad hominem attacks to counter modern “popularist” movements.

“Dad’s not very punctual”.

“What do you mean? Yes he is.” 

“No his punctuation is terrible.”

Queensland arts minister ignored recommendation that new theatre be named after poet Oodgeroo Noonuccal.

I’m in two minds on this one.

“I hate it when they make a really bad TV series off the back of a great movie. That’s called copyright. No … copywrong.”

Some people have what might politely be called a relaxed relationship with thinking.

You know the type. Loud. Cheerful. Firm opinions on everything. Reasoning appears optional. They are not malicious. In fact they are usually quite nice. They simply operate on a conversational model where the mouth begins work several seconds before the brain clocks on for the shift.

Psychology has various clinical ways to describe this. Lower working memory. Lower cognitive reflection. High extraversion. Reduced metacognition. In plain English this means the first thought that arrives is immediately broadcast to the room without any tiresome internal quality control.

Some people find this unbearable. Their brain keeps noticing the logical gaps, the abandoned reasoning, the confident conclusions assembled from approximately two facts and a childhood memory. And loud.

Personally I don’t mind them. Often I find them funny and entertaining.

Generally they are good-natured and I don’t believe that every conversation needs to be a peer-reviewed journal article. And these people are surprisingly useful in the wild. They introduce people, start conversations, volunteer for things, and create momentum while the careful thinkers are still sitting quietly in the corner calculating probabilities.

Gambling (verb); an emotional-cognitive state in which attention anchors on the possibility of an outcome rather than the probability of an outcome.

Actual message received…

“The difficulty for us is to have a vision for what is possible when we don’t know what the elements of the vision are or could be.”

Translation: “we haven’t got a fucking clue but someone is paying us by the hour to figure it out.”

“The illegal trade is considered a form of biopiracy, damaging local ecosystems and breaking international biodiversity treaties.”

Some Chinese fella was caught smuggling 2000 ants out of Kenya. Yep, ants are collected.

That’s called biopiracy, apparently.

I’m wondering if there’s a biospiracy going on.

The buggers get you to do their work for them; build your family tree inside their app.

Of course, if you do so there’s no motivation for them to spend any money on your DNA test. Six generations of your family tree is all they need to give you your results as a percentage of regional DNA clusters.

Say you had one German ring-in over 6 generations, then you’d have 4% South German in your results, for example.

After six generations they don’t need any more because the impact of further generations becomes diminishingly small. Six generations is, coincidentally exactly what most databases provide.

Since they have aggregated every database of births, deaths and marriages in the world, it would be a trivial task for them to automate the generation of your family tree.

But they go through the charade of involving you in a manual process so you feel like you’ve got your money’s worth.

But do they actually do DNA tests?

A woodcutter and a girl lost their dog in the forest. Later the dog returned, leading a frightened sheep back to the camp the woodcutter and the girl had built, saving the animal’s life.

The woodcutter’s daughter eventually became Australian of the Year for her campaign to prevent harm to lost farm animals. She got twenty tattoos, each of a different farm animal.

At first her campaign consisted of a handwritten sign nailed to a gum tree that said: “Please stop losing sheep.”

Someone took a photo of the sign. The photo appeared on social media. Within two days the girl was invited onto a breakfast television program where three presenters nodded gravely while she explained that sheep sometimes feel confused.

A week later the campaign had a logo, a website, and a committee. None of them owned sheep. They held meetings in cafes and discussed awareness.

Soon the movement expanded. Posters appeared saying “Every Sheep Matters.” Someone suggested a national day of reflection for misplaced livestock. The girl agreed it sounded important.

Her tattoos multiplied. Cow. Goat. Pig. Alpaca. Another sheep, in case the first sheep felt under-represented. When asked why she had so many tattoos she explained that each one symbolised a commitment to raising awareness.

The dog, meanwhile, continued doing the only useful thing in the entire story. Every few weeks it wandered off and came back with another animal that had got itself lost in the bush. Sometimes a sheep. Once a goat. Once a chicken that had clearly made several bad life decisions.

Then the woodcutter installed a camera. The footage showed the dog quietly opening a paddock gate late at night, watching the sheep wander off into the bush, and then waiting about twenty minutes before trotting out to heroically “find” the missing animal.

The pattern repeated several times. Release. Wait. Recapture. Return to camp. Accept praise. Eat better food.

When confronted with the evidence the girl stared at the screen for a long time.

The girl sighed. “Technically it’s still raising awareness.”

The dog, sitting beside the table, wagged its tail slowly and looked extremely pleased with the entire system.

People infer meaning from text and images quite differently.

While LLMs are equally fluent in generating both text and images, they tend to get away with more “bullshit” in text because humans are far less accurate and precise in their scrutiny of text.

How does this work?

With text, people are naturally inclined to infer meaning, fill in gaps, and overlook inconsistencies. Human working memory can only maintain a limited number of elements simultaneously. As a sentence unfolds, the brain compresses earlier details into a simplified representation of meaning. Exact wording, numerical values, and logical constraints are discarded once the gist has been extracted. That compression is what allows readers to follow long passages (e.g. a book) efficiently, but it reduces the likelihood that errors will be detected.

In contrast, in images, errors like misaligned elements, unrealistic proportions, or incorrect labels are immediately noticeable. The visual system evolved to rapidly detect opportunities or threats. The brain processes large portions of a visual scene in parallel rather than sequentially. Errors are immediately flagged.

Text processing relies heavily on top-down inference. Readers reconstruct meaning by combining the words with prior knowledge, expectations, and within the constraints of the inferred narrative. Experiments show that readers frequently fail to detect missing words, duplicated words, or semantic contradictions if the sentence structure remains plausible.

An example is the sentence “Paris in the the spring”, where many readers miss the duplication because the brain predicts the phrase rather than verifying each word.

Another example is “the company’s revenue grew 30 percent from $100 million to $120 million”. It’s not 30 percent, it’s 20 percent, but rarely would a person notice.

A final semantic example is the sentence above: “In contrast, in images, errors like misaligned elements, unrealistic proportions, or incorrect labels are immediately noticeable”. Noticeable means it could be noticed, whereas the point being made is that it is actually “noticed”, and is not just “noticeable”.

Also, unless you were really paying attention you probably didn’t even recall that sentence from four paragraphs prior. And some of you, as you read the last paragraph, just thought “which sentence?”

The purveyors of large language models exploit this tolerance. Their models generate text sequences with strong statistical coherence at the sentence level, which produces the subjective experience of competent text. If the narrative holds together, readers rarely audit each element. The result is a high rate of unnoticed hallucination: semantic meaning can be inferred even if, logically, it isn’t present.

In contrast LLMs are fairly useless at creating accurate and precise images. The irony is that they are just as useless at creating accurate and precise text: we just don’t notice.

Politicians, business leaders, and just about every other human that wants more than their fair share: they also exploit this tolerance in their fellow fuzzy thinkers.

And get this, it’s even worse in speech. Listeners must construct meaning in real time while retaining only fragments of the stream in working memory. Prosody, cadence, and confidence substitute for logical structure. If the narrative sounds coherent, the brain accepts it.

Read this blog entry carefully – it explains how LLMs work.

A vending machine that dispenses little air canisters.

Each one would be bottled in different parts of the world;

Rio

Daintree

Mt Everest

London

Tehran

Yeah mate, I love emergencies.

That in response to “are you willing and able…”

In life, there are times when you just have to lean into Jetstar.

Really, who needs consensus anyway?

For the record, I wasn’t seated in an emergency row and there was somebody in my seat anyway.

So I took my chances with 12D.