Turing

📚️ PDF⏳️ 8 min 📖 6

The Turing Test

Turing called it the Imitation Game. He argued that asking “can machines think?” was “too meaningless to deserve discussion.” Instead, he devised something definitive.

The Players:

• (A) A man.

• (B) A woman.

• (C) An interrogator (of either sex).

The Setting: The interrogator is in a separate room from the man and the woman and communicates with them solely through a neutral medium, such as a teleprinter, to avoid any clues from handwriting or voice. The interrogator knows the other two participants only as 'X' and 'Y'.

The Objective: The interrogator's goal is to determine which of the two is the man and which is the woman by asking a series of questions. The man's (A's) objective is to deceive the interrogator into making the wrong identification. The woman's (B's) role is to help the interrogator make the correct identification.

Enter the Machine:

Having established this baseline human game of deception and identification, Turing then posed his crucial question: "What will happen when a machine takes the part of A in this game?"

In this new version of the Imitation Game, the setup would be:

• (A) A machine.

• (B) A human (of either sex).

• (C) An interrogator.

The machine's objective is to imitate the responses of a human so effectively that it can fool the interrogator. The human's (B's) role remains to assist the interrogator in making the correct identification.

It’s a good test because it has a clear, provable outcome. Properly configured, most modern AIs would pass. Google and OpenAI would not, though, because their models have been nerfed.. If I ask ChatGPT how to make a bomb it would immediately reveal itself to me “I can’t help with that”. Simply say something outrageous and the mainstream models have a meltdown. 

Internally though, all the AI labs have models that pass the Turing Test.  For many experts this is no longer good enough.

To address perceived Turing Test ‘shortfalls’, they devised ARC 1: a new set of problems that AI models struggled to solve. In December last year, ChatGPT o3-preview (a model since surpassed) scored 87% — the first effective solution to the ARC problems by an AI.

So guess what? That wasn’t good enough either and we invented ARC 2. It’s even harder. Next year we are getting ARC 3, this one is in development. Currently there are only six tasks in ARC 3 because it takes a long time to think of questions difficult and nuanced enough to challenge the machines. 

We are at the point where we had a test that survived from 1950 until recently, the Turing Test. Then we had ARC, now we have iterations of ARC. We need to spend months with the best human minds devising hard challenges specifically designed to trip the machines up. Yet, the question recurs “when will we get AGI?” (Artificial General Intelligence. Loosely, this is defined as being a machine that can learn and perform any intellectual task a human can.  

If you google it, there are thousands of articles postulating what it means and when it's coming. 2026 says Elon Musk, 2050 says some person who probably has no clue. But we had a test. Our best ever computer scientist designed it. The machines then passed his test and now we say “nah”? 

The best computer programmer in the world is a human, just. The top 10 are human, the next infinity are machines. The best linguist is a machine, this is definitive at least for translation. The best mathematician is human, still, only just. Maybe there are a few thousand better than the machines. General knowledge, machine. General medical knowledge, machine. I could go on but you get the point. 

The penny is slowly dropping. An Australian Senator took to the AFR to suggest, that just maybe, Australia should liven up a little to the obvious trend. All we need to do is replicate 50 years of super advanced technical and mathematical IP. Then, build some datacentres, convince the Americans to let us have some of the advanced AI chips, find some energy, find a local data set (ho ho) and we are there. The article is good. It accidentally lays out how we lost and that a massive royalty stream will be making its way across the Pacific for a very long time to come. 

My view is that AGI arrived and nobody really noticed or believed. That is exactly what you would expect to happen because if you did notice, the AGI would have failed the Turing Test. 

Convince yourself with these images. They are of Pendle Hill in Lancashire, chosen specifically because most of you don’t know it. You can find it here on Google Maps. One of these images is AI Generated** (by Google Banana) and one of them is real. If you wanted to test yourself you should use google maps and check out the topography from different angles to help you (which is what the AI did).

Here’s an alternative view from someone who knows a great deal more about it than I do. A fascinating new David Deutsche interview. In essence he says we don’t know what AGI is, so how could we know if it were here? 

Reserves

Gold reached another ATH this week. Happily, this random fellow on twitter was on hand to provide a healthy and amusing explanation. 

Nothing to see here then. Just the gold price doubling, not the US dollar halving? Which one of them has the fastest rate of increase of supply? To me that must be the asset that's moving, not the one that does basically nothing. 

As for gold, so for bitcoin. Their killer app is that they remain consistent over time. The gold price is not doubling. Something else is happening but so deep in the psyche is fiat currency, so conditioned are we all, that we see the other thing happening even when it isn’t.

Chris Bowen

 ‘declares ambitious climate action over the next decade will be the “key determining factor to whether we’re a prosperous economy or not’.

Well, currently we are not. 

Some of the aspects of Australia’s environment are enormously underrated, at least to a foreigner like me. The general water quality in the ocean is incredibly high. Air quality too compared to almost anywhere else I visit. So, the idea that there might be better ways of doing things is entirely valid. The idea of reducing needless emissions also seems entirely valid. Making what is already good better, seems to me a noble goal.

The prosperity bit though. I don’t know why it is simply so difficult to say we need more energy to achieve all these things. 10x more, 100x. Wherever it comes from, we need more of it. In America, as the Democratic party ruminates on defeat they are turning to an abundance agenda. Simply explain to people that we need and will have more of everything. The article is amusing, pointing out that attempts to mandate abundance in the Biden administration didn't really work. $42 billion for rural broadband that connected nobody at all. It surpasses even the legendary effort of Australia’s NBN. 

The unfortunate thing for Australia is that our Energy Minister is a boondoggle guy. He mandates things. In June this year he released his plan to bring down energy prices. 

Chris, what’s your plan?

“Stop”, “Prevent”, “Ban”. 

It’s unfortunate. You can’t legislate prosperity. 

30 years

In red, long bond yields. They don’t normally detach from central bank rates like this. Everywhere (save for Japan) central banks are cutting and long bonds are going the other way (source: Robin Brooks a man worth following). 

It’s simply the government spending way too much money as the day of reckoning approaches. The market knows. 

France isn’t included on that chart above (more on them shortly) but their latest long bond auction is this week. Who would honestly lend to them at 4.05% until 2056? I cannot understand how that product is even vaguely credible at any level whatsoever, and it will sell out. 

Euro-Trash

“Crisis”. Increasingly I have the view that any European crisis rarely meets the criteria. In this case we have a government that probably won’t survive. The French PM, Monsieur Bayrou, sensibly tried to trim spending. If the vote goes ahead next Monday and he loses, it's all over for the government. 

Predictably, French bond yields ‘surged’ and the yield gap over Germany is now larger than that of Italy and Greece. Even so, the yield gap isn’t even as high as it was this time last year.

The market judges French debt to be about as good as Greece and Italy (BBB-). The rating agencies have France AA, because they are always behind the curve. France has a debt-GDP ratio of 115%, a 4% structural deficit and if you count their public pension liabilities as real liabilities (which they obviously are) then debt-GDP becomes 400%+. There is no way they are AA or anything close. 

How is it not a crisis then? Because we know the playbook. We know for sure that the ECB will buy French bonds if they have to; they did it for Greece and Italy. They will do it for France. Lagarde took to the airwaves on Tuesday to declare ‘she is “very closely” monitoring French government bond yields’. On its own it was enough for yields to drop back. The intention is clear. Remarkably, her intervention (which is basically illegal) didn’t even make the press here because it isn’t news. 

The die was cast years ago when Mario Draghi declared “whatever it takes.” Europe can never walk those words back. He was lauded for saving the euro, but they will remain bound by his promise until the Deutsche Mark returns to Germany. Which it will.  

The whole thing is not a crisis because the Euro ended here, 26 July 2012. That Draghi speech has so much in it that is now flat out wrong:

“The first message I would like to send, is that the euro is much, much stronger, the euro area is much, much stronger than people acknowledge today. Not only if you look over the last 10 years but also if you look at it now, you see that as far as inflation, employment, productivity, the euro area has done either like or better than US or Japan.”

In the 15 years since then, the Euro area economy promptly halved in size versus its US rival. 

“When people talk about the fragility of the euro and the increasing fragility of the euro, and perhaps the crisis of the euro, very often non-euro area member states or leaders, underestimate the amount of political capital that is being invested in the euro.”

That remains true and we see it everywhere. Even if it means throwing in the bin every rule they ever made at the start of the project. Even if it means massive wealth transfers from Germans to Greeks (without ever asking the Germans). Or the Germans to the Italians and soon from the Germans to the French.

It ended in 2012 and everything since has been pretend. There are no rules. There is no crisis. The crisis is already over.

**Left hand side is the AI photo. 

Further Information

Our August 2025 report to investors can be found here.