The theft of progress

📚️ PDF⏳️ 8 min 📖 7

Still important

This was rather unhelpful on a Saturday morning. It precipitated one of the biggest liquidation cascades seen across this sector. The numbers are variously exotic but it is certain that billions of dollars of leveraged positions were wiped out. For scale, it was 10x the size of the FTX wipe-out. 

Shortly thereafter, prices recovered and we are now pretty much where we were 10 days ago. That is only true of course for those that don’t use leverage. The vast majority of the damage was leveraged positions being taken out; some of the smaller tokens dropped over 50% on Saturday morning. Irrespective of your collateral that would be sufficient to take almost any leveraged trade to zero. 

None of this is new of course. Back in 2019 we wrote this in an investor update which is still true despite the industry being 10x as large. 

In a new investment sector like this, liquidity is key. New assets launch all the time and return spectacular results but there is very little volume, making buying and selling at scale extremely difficult and risky. We operate with strict liquidity conditions such that we disqualify coins where liquidity is low relative to the major index component, bitcoin. This can mean we miss strong rallies in smaller coins but as you can see from the overall performance for the year versus the base index, it works. The top 30 single day gains all went to assets other than bitcoin in the last 12 months, for example Bitcoin Cash, gained 54% in a single day on December 15, 2018 but is down 44% for the year.

These sorts of market moves are likely to get bigger as time goes on and the industry grows in size. The fact that crypto markets trade all weekend is a huge opportunity for well financed traders to exploit thin liquidity. “Record liquidations" isn’t a headline anyone wants but it’s true because the industry is growing so fast, so that headline will keep repeating. With growth comes all the attendant drama of get-rich quick schemes. 

Hold the best assets for the longest time and leave it at that. 

Joule

It’s odd that this opinion is controversial. Nobody wants to admit (yet) just how much energy is going to be required to continue the growth in AI. The country with the largest energy capacity is going to win. It's pretty much as simple as that but building that capacity is not as simple as printing the money to do it. 

As Elon points out here. That is the invention at the heart of Bitcoin “Proof of Work”, to mine bitcoin you have to prove that you have used the energy through computation. 

Proximity to energy is going to be the primary driver of national wealth. Arguably it always has been - Saudi Arabia, Norway, United States - dare I say it, Australia. That reality is much better understood in some parts of the world than others. 

In September the World Nuclear Association welcomed its newest member, Microsoft. Microsoft made a deal last year to reopen Three Mile Island, the US nuclear plant in Pennsylvania. A reopening that is now a year ahead of schedule. Not just Three Mile Island, the Palisades plant in Michigan which was turned off in 2022 is now operational again too. They are all going to feed data centres. 

Australia is heading down a different path but at least we might be able to sell a bit of uranium to countries that understand the Kardashev scale.

A lot of things are going to manifest in energy decisions in the next decade. Precisely for Elon’s reason “you can’t fake energy” the decisions that government’s make will be very important because it will take decades to correct them if they are wrong. The basic strategy ought to be “we will have as much of it as possible”. From there work back to how we make sure it's clean and doesn’t kill us. At present our settings appear to be “let’s just have less of it so it doesn’t kill us”.

This is Australia’s energy dashboard on Thursday morning. Wholesale energy price, negative everywhere. Super-negative in our most ‘net-zero like’ state South Australia. The reason this doesn’t reflect in your bills is because all the old school generation is still running in case it gets cloudy or the wind stops. Lucky you! you get to pay twice. 

This is before Australia’s massive renewables build-out. We are going to get enormously more amounts of power during the day from both wind and solar, and do what with it? 

If you want a wealthy country that can afford the sorts of advanced technologies and comforts people want you need a lot of it. What you don’t want to do is produce lots of it and then for whatever reasons decide you don’t want to use it, or can’t use it because it’s intermittent. 

As far as money itself goes, the perfect money was and always will be the joule. Currently only one thing in the world currently comes close. Thank you, Elon Musk, for the reminder. 

Luxembourg

Luxembourg has become the first European sovereign wealth fund to invest in Bitcoin ETFs. I was relating this to someone earlier in the week and they pointed out that Norway did so a few years ago. Norway however is not in the European Union (they have had referendums on it in 1972 and again in 1994. People voted no both times). So, Luxembourg takes the crown. 

Their sovereign fund name translates from French as the “Intergenerational Sovereign Wealth Fund”. It reported last month that it holds its $811 million portfolio as 57% bonds, 40% in equities, and 3% cash. So a 1% allocation to Bitcoin ETFs amounts to roughly $8 million.

It’s a nothingburger in the scheme of things. Perhaps more surprising is that if you were running an ‘intergenerational fund’ would you really have 60% in bonds and cash?  All told though, we have for many years discussed nation states buying bitcoin, now they do. Even in Europe. 

As an aside, the Norway comment did pique my interest (I’d forgotten they weren’t in the EU too). I mapped out Norwegian growth since the launch of the Euro versus the EU as a whole. Their economic growth is 25% larger than that of the EU despite the tough times they have had recently. Perhaps it would have been higher if they were a member but it does seem rather unlikely.

The theft of progress

Excitement then. The new MacBook is out with its M5 chip. Apple is promising all sorts of AI excellence from the device. I will not be buying one following the same promises on the iPhone 16 which delivered nothing at all. To Apple’s credit, my now seven year old MacBook Pro is still excellent (apart from the now noisy fan). 

The price of the base model in Australia will be A$2,499, exactly the same price as the base M4 model when it came out last year. So in theory, inflation in this category is zero. Sadly for us that is not how inflation is calculated. The new machine is better and so adjustments (known as a hedonic adjustment) are made to strip back its capabilities to those of its predecessor. The process is more complicated and more general than that because it is not done device by device but it certainly is done. 

In this case I gave GPT-5 the specs of the new machine and the old machine and asked it to estimate the size of the hedonic adjustment. 

13% is conservative. Probably not because lots of other things about the machine are superior like the battery and the screen for example. Both are not considered here but let’s be even more conservative and say it is 10% better. 

For the purposes of CPI then, computers are now 10% cheaper year over year. To you, the end user, they’re the same price. All buyers of this Apple product are indeed pretty happy that it is the same price as the predecessor for something so objectively better. You might say in that case that we absolutely do need the hedonic adjustment. From a pure stats perspective in fact, we do.

Come the next CPI announcement, the government will say “Inflation back in target range of 2-3%”. That will include all the hedonic adjustments across the economy. They are prevalent in tech and less so elsewhere but even so the progress that Apple makes and technologists around the world make is simply consumed by the government. 

We are conditioned that prices should always rise. Prices should always fall. Technological progress artificially masks the true extent of inflation in all modern economies. This is going to get even more acute with AI. As it gets integrated into existing technology, the hedonic adjustments are going to be huge because technically things are going to get so much better. Take x-rays, currently your friendly orthopaedic surgeon muses over the image and if the break is complex will consult with a colleague. Next year (now in fact),  Dr Bone can simply load it into a machine that has seen every x-ray in history and “assist with his opinion”. I’m using inverted commas because I would trust the machine's opinion more than the doctor’s.  

Explanations for why inflation is good from economists and politicians are generally laughable. The worst one I have seen is from the Bank of England 

“the 2% target helps everyone plan for the future. But if inflation is too low, or negative, then some people may put off spending because they expect prices to fall. Although lower prices sounds like a good thing, if everybody reduced their spending then companies could fail and people might lose their jobs.”

If prices fall, everyone will stop buying things, lose their jobs and die. 2% it is then. In fact, inflation in the UK is 4%, so it's twice as easy for everyone to plan over there.  

The technical adjustment is a huge and very little understood rip-off.

Euro-Trash

Interesting plan laid out in full here on the European Commission website earlier in the year which has now gone live. 

Strategic actions include 

• measures to attract global scientific talent and highly skilled professionals to 'Choose Europe'

• €600 million from Horizon Europe to enhance access to computational power for science, which will secure access to AI gigafactories for EU researchers and startups

• plans for doubling Horizon Europe's annual investments in AI to over €3 billion, including doubling funding for AI in science

• support for scientists to identify strategic data gaps and gather, curate, and integrate the datasets needed for AI in science

Of course the issue for Europe is that they do not have the “computational power for science”. They do not have the energy capacity for it either. In America they are raising 500x what Europe is about to spend and doing so from the private sector. There seems to be no acknowledgement in Europe that their AI objectives are completely at odds with their net zero goals. AI is going to use an unimaginable amount of energy that they don’t have. 

What is more, Europe has zero IP in this area; all the meaningful models are American or Chinese. Certainly the ones capable of making scientific breakthroughs are behind closed doors in Palo Alto. 

Frustratingly for Europe they did have a chance in all of this and wrecked it on day one. So much so in fact that Marc Andreessen never hesitates to remind the world of it. 

Why does he post this man's face? In perhaps one of the most memorable euro-tweets ever. Commissioner Breton announced that the EU had AI Regulation, nobody else did. He mocked America and China which was very odd because despite having regulations the EU had almost no AI companies. They had no IP in the area, and they still don’t. 

It was a monumental moment of stupidity almost two years ago. Since which date, Europe’s relative position in AI has gone backwards several orders of magnitude. 

Further Information

Our September 2025 report to investors can be found here.