Quantum

📚️ PDF⏳️ 5 min 📖 6

Google Quantum

Views were mixed on the Google Quantum announcement this week. The good part is that their new Willow chip performed a calculation that would take a classical computer until the end of the universe to attempt. The bad part is that the calculation itself is not a practical one. It was specifically chosen by Google as the benchmark simply because their machine can do it. So, while no classical computer could ever do the calculation, there would be no reason for it to try either.

Google’s chart shows the equivalent number of years a classical computer would take to compete with the different quantum chips. The Y axis is a log scale, so clearly Willow is nuts in terms of potential power. 

In any event. The ‘breakthrough’, if it is one, dredged up the fear that if we do get a quantum computer it will immediately destroy bitcoin because encryption will be finished. I must say firstly, if there were a quantum computer at sufficient scale this would be true. What’s more, encryption is probably the first thing one would set a quantum computer to work on. Why? because if a quantum computer does a calculation you couldn’t do before, how do you check it is correct? At present. With encryption we can go from private key (the answer) to public key (the question) easily. We cannot go in the other direction though. A quantum computer could in theory and if it did, we would know the answer was correct.

So, if Google really had a quantum computer, they would use encryption (specifically decryption) to prove it. To my mind then, they don’t. 

All the same it prompted discussion in bitcoin about quantum proof algorithms. Jonas Nick of Blockstream published this proposal (ably assisted by Australia’s own Anthony Towns). It shows how the introduction of a quantum proof algorithm called the Winternitz One-Time Signature scheme might work. (Heavy duty maths paper on it here.). It requires protocol changes, but it is workable if implemented. 

To summarise. Google does not have a practical working quantum computer. They are a very long way from having one that can crack encryption (Willow has 105 Qbits, to crack bitcoin would need 13 million Qbits). Quantum proof algorithms exist but bitcoin does not use such a thing today. It could in the future and the proposal shows how that might work. Progress with quantum chips is quite slow though. That’s because the more Qubits you add the more interference you get. 

Considering the rate of addition of Qbits, doubling every four years you don’t hit 13 million until the end of this century. It will probably be quicker, but this is not an urgent problem for bitcoin. 

Back to the announcement for a minute. I do not want to diminish the effort here. What Google did in running a quantum computer for a decent period of time without the error rate spiking off to infinity was very special indeed. If it did something, it indicated that the speed of progress will now increase. 

The ideal world is quantum computers with quantum proof algorithms. The sooner the better. 

The AFR summary on all this was rather good too. 

A week when decades happen

Another one. Feels like we have had a few this year. 

I’m referring to OpenAI’s 10 days of Shipmass. New AI products every day for 10 days. Last Friday brought the $200/month pro subscription. Monday brought Sora. The video production AI. 

Right now the videos are short, not that good, you couldn’t make a film for example. Importantly though, The cost of producing these small clips is next to zero. That is likely to drive a profound change in global cinema. My favourite examples to date are here. 

Film is a powerful but expensive medium. Hollywood, at least to my mind, was the marketing department for American power. All the movies I watched as a kid showed America in an amazing light. Perhaps deservedly so. Hollywood provided the means for the US to project both culture and their perceived vision of capitalism through cinema. 

Bollywood is the same. To me it projects Indian growth and vitality in a very specific way.

The medium of cinema was really the preserve of an elite group for a very long time. It was fairly easy to coordinate messaging across the population about, shall we say, “what is good and what this country stands for”. We have seen that for a decade now with Hollywood turning out films that match the political zeitgeist. 

Save for Dune (which I did enjoy) there have been so few good movies in the last decade. Everything is a prequel or sequel because of the funding required. You need to nearly guarantee success if you want a film to be made. If you want studio funding you better be absolutely on message or it's going nowhere.  Within the next two years we are likely to get open source, feature length movies directed by random nobodies and produced by AI. So we will get actual art again, instead of politically transcribed messaging. 

Far from AI killing off art, which is a discussion I see everywhere, I think the barriers to entry will just disappear. Make and distribute your film as you wish for near zero cost. 

Think Mad Max (budget $350,000- grossed $100m) or Blair Witch Project (budget $60,000 - grossed $250m). This time. Cost $1,000 - grossed $1bn). 

Good original films, from underfunded talented people. It will happen. 

The full list of OpenAI products released over “Shipmass” are here. The rate of change is accelerating. 

Microsoft

Microsoft shareholders were underwhelmed by the prospect of adding bitcoin to their corporate treasury. 0.55% of shareholders said yes.

I agree with them. If I want to buy bitcoin, I buy bitcoin. If I want to buy Microsoft, I buy Microsoft. 

Ultimately if you are a corporate treasurer and recommend adding bitcoin, your chances of ‘pursuing new opportunities’ are going to go up a lot. At its present $2 trillion scale bitcoin is not really large enough to accommodate US corporations. Perhaps at $10 trillion it will lose some of its volatility. Really though, if you are Microsoft, use the money to do what you are actually good at. Maybe something like replacing Microsoft Excel?

If anything good came from this piece of theatre it’s that we now have a benchmark against which to measure openness to bitcoin in corporate treasuries. It’s likely this will make an annual appearance on the Microsoft ballot and we can track the slow demise of the United States Dollar. Bitcoin is $100k but there is quite obviously a very long way to go here. 

WSJ

 While there is a long way to go. There were signs of progress in the Wall Street Journal. 

I rather like the idea that if you can’t explain it simply, you don’t understand it.

Q: Why do you like bitcoin? 

A: There aren’t many of them, and I can personally prove it. 

Euro-Trash

A new EU ban! Unfortunately for weary European manufacturers this one, which banned inputs grown on deforested land, has been paused until next year. Everyone is very upset because it was extremely costly to implement the now unwanted compliance procedures. 

You can imagine the frustration. The law was always a bit daft because it is incredibly difficult to police and everyone will almost certainly fall foul of it somehow. Particularly if your company or sector happens to be out of favour at the time the EU comes looking. So you spend money, you do your best because that is all you can do and then they turn around and say “nah”. 

The best way to experience EU regulation is to browse the internet with a European IP address (use a VPN). It's literally a bombardment of pop ups and messages about privacy and accepting cookies and whatever else. It is so hard to do business. 

Once again, all this week's brilliant AI releases from OpenAI and Google are not available in Europe. It’s just too hard to comply with whatever nonsense they have dreamt up this week. 

I mean, it's a joke. The CEO of OpenAI is getting bombarded with abuse because regulations prevent him from releasing product in Europe. As the comment underneath shows, it is also a reality that the Taliban do get OpenAI’s products before European customers.

2025 promises a lot, but perhaps not west of the Bosphorus. Once again, this list is a joke, except it really isn’t. 

Further information

Our November 2024 report to investors can be found here.

If you are considering an investment in the Managed Fund, you can apply using our online application form: