Recently an interviewer asked me how I got to be such a good forecaster, and I replied by saying something humble. In retrospect it was a bad answer because I should have instead used the opportunity to give actual advice on how to forecast AI well. Here's a stream-of-consciousness attempt to do that:
The heuristic that things which sound weird and sci-fi are less likely to happen in reality, is bad. I suspect that's what really going on is that things which sound weird and sci-fi put you at risk of being judged a weirdo if you talk about them which is not the same thing as are unlikely to happen. Repeat to yourself the mantra that some weird sci-fi things really do happen, and others don't, and you have to take them on a case by case basis.
Trend extrapolation is your friend. Your best friend. Don't let anyone tell you otherwise. I've actually only rarely see someone extrapolate a trend too credulously; more often, people have a trend staring them in the face and extrapolate it a tiny bit into the future and then are too timid to keep extrapolating it. In general I think it's reasonable to extrapolate a trend as far into the future as it extends into the past. Now obviously, trend extrapolation is just the beginning of the forecasting process, it's not the end. For each trend you should ask yourself whether it makes sense for it to continue like that and if not why not etc. You'll usually end up with some sort of sophisticated view about how the trend will probably continue but bend downwards a bit and then ultimately plateau around X level.
Explicit models are also your friend. Things like Bio Anchors, the AI Futures Model, https://t.co/kScy7UC1In, etc. The process of making your own complicated model like this, and engaging with the models made by others, is... well, I think it's pretty edifying. I'm not sure why but I could speculate. Maybe something about teaching you to be appropriately humble (e.g. when the model output is sensitive to a parameter you have no clue about) and also teaching you to more quickly identify the considerations that matter most, and ignore the rest?
For short term predictions, especially about geopolitical events, the sorts of things that people are gambling about on Polymarket, the heuristic "nothing ever happens" is pretty good. Things do in fact happen, of course, but betting markets and online discourse tends to be biased a bit towards things being more likely to happen than they really are, and so you can get an easy win by just correcting a bit downwards from the 'wisdom' of the crowds.
Scenario forecasts are also your friend. They help you ask yourself the right questions, and they help you notice when some of the things you thought contradict each other.
One of the more uncomfortable observations in our AI Value Capture piece is internal: our token spend at SemiAnalysis now runs at roughly 30% of employee compensation, with employees pulling just under 5 billion tokens per month on average, over 5x more than Meta, and our top contributors clearing 100 billion. We wrote about it openly because every research firm, hedge fund, and law firm we know is heading toward a similar number, just on a delay. (1/4)🧵
I have to think of this cartoon by @tomgauld often when I read online comments regarding articles. Pretty sure RHPAC is the most common type of online comment…
I have argued for years that #China cannot make the world dedollarize. This piece by @AgatheDemarais lays out why, with the numbers to back it.
#Beijing has built the plumbing. It cannot manufacture demand.
The renminbi settles ~30% of China’s own trade, but under 5% of global trade, and almost only when a Chinese firm sits on one side. Offshore #yuan deposits are $234bn. #Dollar assets outside the #US are $15tn. That is the whole story in two numbers.
CIPS, China’s system for moving yuan across borders, sounds formidable until you look closely: of 1,791 “members,” just 194 connect directly, and nearly all are Chinese. ~80% of CIPS transactions still route through #SWIFT, the messaging network the West controls. That is redundancy, not independence. Not the sanctions shield Beijing markets.
The e-CNY is no different. $6bn a day against $150bn+ for Alipay and WeChat Pay. Cities now pay officials in digital yuan to force adoption. Cross-border, mBridge clears about four payments a day.
Here is the irony. Demand for digital money is real. It just isn’t going to China. #Stablecoins went from zero to $317bn, 99% pegged to the dollar, most held in the emerging markets Beijing wants to court.
You can build the rails. You cannot force the world to ride them. Absent a massive shock, the takers aren’t coming.
https://t.co/BBwJNmPgA7
The average life expectancy of a new Russian recruit—from arrival at a training ground to death in a combat zone—lies somewhere between 10 days and three weeks. Once sent onto the battlefield, they survive an average of 20 to 35 minutes. @peterfrankopan https://t.co/W3UhBerdH0
I don't know if I should be relieved that I can add value here or terrified about what the heck these agents are building at my direction and if it will ever work.
Managing AI agents is scarily and eerily similar to managing people. I just led Claude and Codex through a Q&A session where I kept asking questions and pointing out issues until they both agreed on the actual root cause of a problem. So eerily like managing a human team.
I just kept pointing out logic errors and asking about "why did this work but not that" and "did you compare your results to any others". And I am not a coding expert. I didn't even look at the code. I just asked outcome based questions.
Om's death hit me hard. Reading everyone's memories and tributes helped me feel connected to them and soothed my grief. And I wanted to provide context or meaning from his life in a way that would make @om proud and maybe offer comfort to others.
https://t.co/LwbqjQMWfO
@JamesSurowiecki I agree. The police acted shamefully on what was an obviously politically motivated complaint with weak sourcing. They should be held to account for this.