The Singularity

I confess that the rise of the AI machines has really piqued my interest. Using them personally I can do all sorts of things that I could not do before, even just entertaining doodles. I asked ChatGPT this week to write me a program in C++ that could identify palindromes in prime numbers. It is absolutely extraordinary in its coding capability; here is a snippet:

I have no skill at all in programming, so can only do what the AI program tells me to do and yet this program worked and it revealed that the 231st palindromic prime is a 3321233, which is both totally useless and remarkable at the same time.

If I am just messing around with it for no particular purpose, what are the people on a mission achieving? It absolutely destroys one of the major barriers to automation, which was a lack of developers. I would think very shortly it will be a matter of just typing up your desires into the machine and it will return a fully operational website. The whole thing massively accelerates automation, and massively accelerates knowledge accumulation.

Now, Raoul is prone to hyperbole as anyone who follows him would know. I agree with this sentiment though. To write the program above would have taken me at least a week. First, I’d have to find a willing developer, tell them what I want, get something back that isn’t quite right, iterate once or twice et voila. Now it’s minutes.

As he says, in five years it will have eaten almost everything. That is not just because more specific machines will be released but because the pace of the improvement of these machines is accelerating and their cost is collapsing. It’s going to make things a lot better at a frightening pace and the best thing you can do is embrace it and include it in what you are doing.

There is almost no argument that I can see that will hold this stuff back. The technological acceleration is massive. Consider then this comment from the Australian Government’s Chief Scientist this week:

“This is an example where the private sector has brought up a technology, it gets adopted really fast, and we haven’t been ready for it, to work out how we manage this,”

“Manage this.” AI is just an algorithm, it’s a maths program and while it’s a complicated and iterative one there isn’t much to “manage”. She went on;

“There’ll be a whole range of approaches, it will evolve over time, we will learn to live with it. That would be my guess.”

Learn to live with it? Surely leverage it! Surely deploy it in a way that massively pushes forward the boundaries of health, education and scientific discovery in Australia.

Sadly, ChatGPT is already banned in NSW and QLD public schools. The theory is that students will use it to do their homework. Yes, they will. What else will they use if for though? What massive discoveries will they make if you actually give them the best tools instead of insisting people flop about in the library with the Dewey Decimal system. It seems mad.

Once again, we have a mathematical algorithm that will eat the government’s lunch. The madness of “managing” or “controlling” will probably last for years but just like when encryption came along the genie doesn’t go back in the bottle.

Google’s chief scientist (whose record is pretty good) had this to say back in 2017:

The raw code for knowledge accumulation, value transfer, contracts, learning and creation is here. The singularity is close. The best thing we can do is embrace it. Unless of course you go to school in NSW or QLD, where you’re blocked so you can learn about it in 10 years when the rest of the world’s children are a decade in front of you.

Ark

It’s ARK time and their annual Big Ideas report is out. As usual a pinch of salt and a deep breath is required but it’s a fascinating read. In particular, again, the AI section. The cost to train AI is just plummeting meaning what we think is brilliant now will surely be a child’s toy in little over a year’s time.

On Bitcoin, ARK had this to say:

A bear case of $258,500 per coin in 2030. I’m not so sure, but let’s suggest it does half as well as the ARK downside case, it would still return 24% annually. My non-investment advice personal view is that it does better than their bear case and I will leave it at that.

One other observation from the report is that if any of it comes true it could be profoundly destructive for a lot of existing businesses and industries and as usual it will likely be profoundly good for Australia. Robots, batteries, computer chips, solar panels and nuclear power stations all need what Australia has in the ground. Most countries should be terrified of what’s coming to eat their industries, while we can just buy shovels and fly to the outback.

Ang Pilipinas

Remittance is big business in the Philippines. Over 6 million Filipinos work overseas and send money home each month and for years they have been getting around a 10% haircut in remittance and exchange fees.

“The Philippines is one of the biggest remitting markets in the world, especially from the United States,” Jack Mallers, CEO of Strike, said to TechCrunch. In 2021, about $12.7 billion in cash remittances was sent from U.S.-based Filipinos to the Philippines”

The beauty of Strike is that it only uses bitcoin behind the scenes. The users only ever engage with fiat currency. So they get paid in USD; transfer locally to a USD account, send on Strike, which remits over Bitcoin’s Lightning Network. At the other end the person receives Pesos immediately.

These transfers normally take 3 days minimum, and now it’s under a minute and 10% of the cost. What’s more, the app can facilitate minimal transfers of <$1 which were previously not feasible because the cost would outweigh the size of the transfers.

At some point, Strike will reach a tipping point because it is so much cheaper than Western Union, all mobile driven and near instant. They will just keep adding countries and it will be a ground up revolution because the poorest countries go first by dint of their dominance of inbound remittance. Don’t expect the service here in Australia any time soon.

Retail

A quick look at the Australian retail spending for December which was ugly. Down 3.9% and consistent across the states. Of course, we can’t read too much into one month, and besides, November does better these days because of the whole Black Friday online horror show.

It’s tricky for the RBA. From here, whatever they do will likely be wrong. I suggest baby steps upward is the sensible approach for interest rates.

Euro-Trash

The Euro-zone banking survey is a widely watched barometer of economic health. You know something is up when even the ECB can’t find a positive spin.

Mortgage demand fell by the highest amount ever recorded and the tightening in corporate lending was the most severe since the European debt crisis of 2011. Yesterday the ECB increased rates again by half a percentage point, they now stand at 3%. Hardly a monstrous rate, but it is a huge difference from the 0% money of only nine months ago.

Madame Lagarde isn’t worried though, growth is so much better than she thought.

She likely will stay the course but come H2 this year, the course might be in a different direction. The truth is, there is no growth in Europe because the population is shrinking and they are about 20 years behind in their adoption of technology so they can’t deliver it through productivity either.

Personally, I love Europe because every time I go back it’s the same. It’s great because it doesn’t grow; it’s cheap because it doesn’t grow and with a birth rate now down at 1.4 births per woman pretty soon we won’t even need to reserve our sun beds first thing in the morning.