“dark factories”

The first time I heard “dark factories”, I pictured the orcs toiling underground at Isengard (that’s Lord of the Rings for any readers who are not the right type of nerd to know that). I also think of Philip K. Dick’s story Autofac, one of my all time favorites. But no, the idea is that factories are emerging in China that are so automated that the lights don’t need to be on most of the time, because no humans are present. Naked Capitalism has a ton of links describing this phenomenon in China. Apparently U.S. industrialists are touring these Chinese factories and are shocked at how advanced they are and how far behind they (i.e., western industry) are.

The fact that western industrialists are invited to tour these factories would suggest that the technology is not secret. So maybe we should not feel threatened but rather look for opportunities to partner and learn. No doubt, there are similar factories churning out military and security hardware that are secret.

Ray Kurzweil speaks

In this Peter Diamandis podcast, Ray Kurzweil says he gets 5-10 interview requests every day, almost all of which he turns down. He took this one, and talked to a panel of tech bros for over an hour, which is pretty cool. I forget Ray sometimes – amid all the doom and gloom, he is a relentlessly and consistently positive voice over decades. And he has credibility as a sort of renaissance man and public purveyor of forecasts which have turned out to be pretty accurate. Here are a few things I remember him saying in the podcast (no transcript – come on, that is something AI can do easily right?)

  • robots that can do my dishes within two years!!! This may be the most exciting prediction ever.
  • He sticks to his prediction, made in 1999, that artificial general intelligence will be achieved in 2029. He acknowledges that there is a lot of debate about the definition of AGI and therefore there will be a lot of debate over whether it has arrived. But over time he has refined his own definition of it to be more precise and in fact, more stringent. First, he says his definition is that AGI will be able to match the abilities of the top human expert in every field. Second, he says that while a few people believe we have achieved AGI already (it’s January 2026 as I type), by 2029 there will be near consensus that it has arrived.
  • He predicts that 2032 will be the year humans achieve “longevity escape velocity”. In other words, life expectancy will increase by more than one year for each year we live. So in his view, we all just need to “hold on a few more years” and maybe death and aging will be become optional.
  • He predicts that sometime in the 2030s (not so precise on this one), we (most countries?) will have some form of universal basic income.
  • The arrival of AGI is not his definition of the singularity. He has refined his definition of the singularity to be that artificial intelligence is 1000 times smarter than (biological) human intelligence. He forecasts this for 2045. Again, he is relentlessly positive and sees humans tapping into this rather than being left behind.

I definitely think Kurzweil has credibility, and I have never felt that he has any political agenda. I read The Singularity is Near around 2012 or so, and I think it is a mind-blowing book absolutely everyone should read. (I haven’t read the sequel/update The Singularity is Nearer yet, so I don’t know if I would recommend that as a replacement or addition to the original.) Everyone should know this point of view is out there, backed by evidence and math, and give it some consideration. That said, if I were really 100% confident that Ray’s predictions are right, I would not go back to work on Monday morning. I would start drawing down my savings and enjoying life from now on, knowing that the age of scarcity will be over before my savings are gone. I could also abuse my body up to just short of the point of death, knowing that any damage I cause will be easily cured. I am way too risk averse to do any of these things, of course, nor do I recommend them to the reader. But maybe when the doom and gloom seems particularly thick around us these days, we can take some comfort in knowing that at least the possibility of a better future exists.

January 2026 in Review

Well, I seemed to be in a political mood in January. I try to stay on the policy side of the line, but that is hard when bad politics makes good policy impossible. Inspired by a Nate Silver post, I took a look back at what I see as key moments in the last 25 years of U.S. history, and there were just so many that were on a knife edge and ended up going the wrong way, in my view. Maybe there are other universes where things went better, but remember my scientific theory that once they make a Spiderman movie about a scientific theory, it is almost certainly wrong. I find it depressing how we got here, but there is no sense crying over it. We need to learn from the past yes, but then face up to the present moment and start picking up the pieces from where we are.

Most frightening and/or depressing story: Evidence is crystal clear that sabotaging R&D spending is a very effective way to sabotage economic growth and progress. Attaboy to the fools, assholes and traitors currently in nominal charge of the U.S. government. Meanwhile, if a more rational administration ever takes hold, research on learning curves might provide some clues on where to concentrate our efforts for the greatest gains.

Most hopeful story: New York City congestion pricing was a hard-won U.S. transportation policy win in 2025. This is just good, economically sound urban policy that would be apolitical in a more rational world.

Most interesting story, that was not particularly frightening or hopeful, or perhaps was a mixture of both: I reviewed book reviews from 2025, one of which was Ezra Klein’s Abundance (not the 2012 book Abundance by Peter Diamandis, which while I am not a huge fan I continue to be puzzled how Ezra Klein could either not be aware of that book or intentionally choose to name his book the same thing.) I still find it hard to summarize that book in a sound bite, which would need to be done if it were ever going to serve as the basis for a political campaign. But here is an attempt: (1) Continuously review and streamline federal regulations, (2) increase public and private investments in critical technology and infrastructure, including recommitting to clean energy, and (3) address market failures in housing, health care, and education. #3 is a doozy of course, but the un-sexy answer just has to be understand and implement the latest evidence-backed policies. I would think ramp up housing supply, Medicare for All, and free (tax-funded) college or trade school for all. And um, if we want a chance for any domestic agenda to succeed, we also need serious plans to manage international risks including war, ecosystem collapse, famine, and massive refugee flows that may be coming. Now, I just want to acknowledge that there is a rosy future scenario where AI magically solves all these problems. The way that could work is that technological progress and economic growth suddenly pick up so drastically that we are awash in cash and resources to the point that even the wildly suboptimal operations of our dysfunctional political system are adequate to solve the problems. I don’t think it is safe to put all our eggs in that basket! We better assume that we will need to continue doing the hard work of allocating scarce resources to manage difficult problems for the foreseeable future.

climate refugees

This Common Dreams article goes into the existing legal framework governing refugees and how it could be extended to define and benefit climate refugees. For example:

  • The Refugee Convention of 1954 was set up in the wake of WWII and addresses “those who must leave their home countries due to war, violence, conflict, or any other kind of maltreatment”. So it doesn’t address environmental displacement or internal displacement, but it could be adapted to address these things.
  • The “1998 Guiding Principles on Internal Displacement” are what they sound like – they have not been formally adopted and are not legally binding. They could be developed into a treaty and/or countries could just adopt the principles as part of their own internal legal frameworks, hopefully also offering aid to neighbors experiencing hardship.
  • A “Global Compact for Migration” was adopted by the UN in 2018. It “promotes safe, orderly pathways for migrants, including planned relocation, visa options, and humanitarian shelter”. “Adopted” means the general assembly adopted it as another voluntary, legally non-binding set of principles. This also could be developed into a treaty and/or countries could incorporate the principles into their own internal legal frameworks.
  • The United Nations Framework Convention on Climate Change is an actual treaty ratified by many countries including the United States Congress. Trump has announced the US is withdrawing from the treaty – which I don’t understand. I naively thought that if Congress ratified a treaty (which is extremely rare nowadays when win-win agreements are viewed by our cynical politicians as a loss of sovereignty), the executive branch didn’t have the right to unilaterally withdraw.
  • “The Loss and Damage Fund was established in 2022 at COP27 to address the financial needs of communities severely impacted by climate change. The money would support rehabilitation, recovery, and human mobility.” It is underfunded of course.

I don’t want to be cynical, but the global political mood is just cynical at the moment. US politicians in particular are not in the mood to sign international agreements or even cooperate informally. So while I think it is good to pursue all of these ideas, I do not think it is a good idea to put all our eggs in this basket.

Climate crisis-fueled migration is already a driving force behind the rise of right-wing parties in the US and Europe, and this ugly feedback loop looks to just keep accelerating over time. As economic conditions in the destination countries deteriorate, the right-wingers are able to scapegoat migrants and that accelerates the feedback loop even more. The most rational way I can see to try to break the feedback loop is to address the environmental and economic conditions in the source countries. Aid and trade are the consensus center-left and center-right ways to do that. The right-wingers are probably aware of this, and so they sabotage both, which accelerates the feedback loop again. So they have no incentive to solve problems, because increasing problems fuel their agenda. Meanwhile more rational politicians can point to the rational solutions, but then when they can’t deliver them within a political cycle, real peoples’ real economic pain again accelerates the feedback loop. We could try to deliver the best economic performance possible as a strategy with some chance of success. Here, the current US administration is unpatriotically sabotaging the foundations of economic success such as R&D, education, and a strong central bank. Sorry for the doom and gloom as I am not seeing an easy way out of this political conundrum. Sit back and hope AI raises productivity in spite of our currently incompetent government and institutions?

Where are the Small Modular Reactors?

Small modular nuclear reactors might be a key part of the solution to the climate crisis, or they might not. They seem to be near the end of the R&D stage (although doubtless they can and will continue to be improved) and at the very beginning of pilot testing/proof of concept. They are not yet economically competitive with other forms of power generation (including traditional large nuclear reactors), which you could say about pretty much any new technology. There is a chicken and egg problem where you have to implement it and scale it up for the unit price to come down, and it is hard to get the private sector (and public sector, if they are short-term financial return focused) to take the chance on implementation of something that might end up not working out. Of course, the idea is to invest in a portfolio of things that have some chance of working out, such that there is a high chance at least one of them will pay off. Anyway, some facts and figures in this Physicsworld article (isn’t that name slightly redundant?):

  • Nuclear power generated 17.5% of the world’s electricity in 1996, vs. only 9% today (2026 if you are a future historian reading this).
  • I somewhat naively thought US firms might be leaders in this technological knowledge (if not in implementation of anything at scale, where I would never be that naive). But it turns out that there are two of these reactors currently operating in the world, and they are in Russia. Two are being built in China.

Sure, there are mining, supply chain and waste problems, but in my view you have to balance them against the unfolding global ecological catastrophe caused by burning fossil fuels for two centuries and counting.

WEF Global Risks Outlook 2026

I think of this thing as not so much a prediction but an indicator of what political and business leaders are thinking and talking to each other about. The results seem to be very sensitive to whether people are asked about a 1-, 2-, or 10-year horizon. So I don’t know that the rankings make a lot of difference. Rather, it makes sense to look at a “top 10” or so. Chillingly, more than half of poll respondents seem to think there is an “elevated” to “looming” risk of “GLOBAL CATASTROPHE” within the next 2-10 years. Within 2 years, they are most worried about “geoeconomic confrontation” while within 10 years they identify environmental disaster as the top 3 worries – “extreme weather events”, “biodiversity loss and ecosystem collapse”, and “critical change to earth systems”. “State-based armed conflict” (would it be simpler to just call this WAR?) is also up there somewhere near the top. “Inequality” doesn’t rank as high, but the analysts identify it as the risk that is most “interconnected” with the others.

How pathetic for our species and civilization that our leaders believe environmental disaster is looming a decade out, and yet they are starting wars with one another on our behalf in the present, when they know damn well they need to be cooperating to head off the environmental disaster that is going to affect all of us, the winners and losers of today’s useless wars alike. People don’t want war, so how are we putting people in leadership positions over us who are failing us so utterly?

https://www.weforum.org/publications/global-risks-report-2026/infographics-global-risks-report-2026/
https://www.weforum.org/publications/global-risks-report-2026/infographics-global-risks-report-2026/

what to eat, or you can take my cheese…WHEN YOU PRY IT FROM MY COLD DEAD HANDS!

Harvard School of Public Health explains how the new U.S. government nutrition guidelines were developed without proper scientific oversight. The normal process is a transparent one where an expert panel reviews the latest evidence and submits a report with recommendations, supposedly without any bias or industry influence. (A cynic could probably look at these highly credentialed experts at leading universities and show that they receive research funding from industry and from government agencies being heavily lobbied by industry, because where else would they receive funding from? But they can at least channel any propaganda through some scientific and ethical guardrails you would hope.) USDA employees aren’t obligated to follow these recommendations to the letter, but they at least give them some weight and balance them against the economic and political factors. This time the panel submitted their report as usual, but USDA then cherry-picked a separate set of experts to produce a “supplemental report” without the transparency or adequate documentation. And the guidelines are then based on that. So they are not credible.

Even though the process was not credible, the consensus seems to be that the new guidelines are not really all that different. The main issues have to do with how they are being (badly) communicated, including an apparent emphasis on more saturated fat (which is not really what the technical guidelines say at all, but the concern is that very few people will drill into the technical guidelines). If I can try to clarify the saturated fat issue, it seems to be that a portion of the population that has no cholesterol issues may be able to increase saturated fat intake with no ill effects, but a portion of the population that has cholesterol issues will have more heart attacks and strokes and early death if they do so. Nutrition advice really should be more personal in an ideal world, but with public health guidelines, broad, simple, clear statements that benefit a majority of the public on balance seem to be the way to go. And replacing saturated fat with healthier plant-based fats and oils definitely seems to fall in this category. If people who are eating a lot of sugar and processed garbage were to replace it with meat, that might actually benefit them which may be what the guidelines are trying to say. Of course, they should be replacing it with fruits, vegetables, beans, nuts, seeds, whole grains, healthy fats and proteins. And I want to state that I support vegetarians and vegetarianism on ethical and environmental grounds. These considerations are missing entirely from the government’s concept of “nutrition”, and they should not be.

Another criticism I have always had of these guidelines is the use of weight, like eat so many grams of fat per day, or fat should be X% of your calories. Even those of us who might consider ourselves reasonably quantitative and logical think in volume or area, not weight. If you told me to aim for X tbsp of vegetable oil per day or Y slices of cheese, I could do that. Tell me Z grams or ounces, and I have no idea what to do, and then I am supposed to convert that to energy units (calories) and determine what percentage it is of my total calories for the day. But people don’t pay much attention to these guidelines anyway. They need to be getting this information from “trusted messengers” like teachers and doctors, and if these messengers had simple clear messages from the government that they themselves understood and trusted, they could just pass them along. Something like a point system that approximates the weights and calories involved could work.

I don’t think these guidelines have much short-term impact just because us laypeople don’t pay attention, and the professionals that could help us eat better don’t get clear communication materials out of these guidelines that they can work with.

But the longer-term damage here is the damage to the credibility of government health and medical advice. When I tell my kids “not to believe everything you hear and read on the internet”, I tell them to be aware of the source of the information. And one source I would have considered credible in the past is a major federal agency like USDA, CDC, etc. If major government, academic, and professional journalistic sources are telling you the same thing and it matches what that social media influencer or your friend are telling you, it’s still not 100% guaranteed to be true but you can start to have some confidence. But the credibility of federal agencies has really been significantly damaged by this administration and it may take a long time to recover, even if the past norms are ever put back into place.

Conflicts to Watch in 2026

The Council on Foreign Relations has a list of conflicts to watch. I’ll highlight a few below:

  • “U.S. military operations targeting transnational criminal groups escalate to direct strikes in Venezuela, destabilizing the Maduro government” – marked “probably or highly likely to occur in 2026” when this was written in December 2025. It took three days.
  • high likelihood, high impact: conflict over (clearly illegal under international law) Israeli settlements in the West Bank; “renewed fighting in the Gaza strip” is also listed as depressingly “likely to occur” [I find this deeply tragic and a human rights catastrophe for everyone involved. I am not sure I find it hugely consequential for global war and peace. I think it is more politically correct for the U.S. media to treat it as such. There are other human rights catastrophes going on in the world which will not get equal media or political attention.]
  • “An intensification of the Russia-Ukraine war, caused by expanding attacks on each side’s critical infrastructure and population centers” – also listed as “probable or likely to occur” – so they don’t see this one dying down in 2026 either
  • “Renewed armed conflict between Iran and Israel” – they mark this one as “an even chance of occurring in 2026”
  • “A state or nonstate entity undertakes a highly disruptive, artificial intelligence–enabled cyberattack on U.S. critical infrastructure” – also marked as an even chance, and I find this one deeply disturbing
  • “Intensified military, economic, and political pressure by China on Taiwan precipitates a severe cross-strait crisis involving other countries in the region and the United States” – also marked as 50% likely to happen in 2026
  • “Armed clashes between Russia and one or more NATO member countries, precipitated by increasing Russian provocations toward European states” – 50% likely
  • “A resumption of North Korean nuclear weapons tests heightens tensions on the Korean Peninsula, triggering an armed confrontation involving other regional powers and the United States” – 50% likely

So the cyberattack we need to be ready for – are we? Not to worry, I am sure the tech companies that created the technology underlying the weapons also have massive government contracts to create counter-measures. Let’s hope the power stays on.

So if we were to gamble (which I guess we can on the futures sites), we would put even money on armed conflict breaking out in at least 2 of these 4 situations, and we should not expect to get through the year without at least one of them: (1) between Iran and Israel, with the U.S. pretending to be reluctantly dragged in [3 NUCLEAR STATES], (2) between China and Taiwan/U.S./Japan [2 NUCLEAR STATES, 1 THRESHOLD NUCLEAR STATE, 1 AMBIGUOUS THRESHOLD NUCLEAR CITY-STATELET], (3) Russia and NATO [4 NUCLEAR STATES], (4) North Korea and South Korea/U.S. [2 NUCLEAR STATES, 1 THRESHOLD NUCLEAR STATES].

There are a couple things they say are low likelihood, so at least we have these: a U.S. armed attack on Mexico, and a China/U.S. armed conflict over the Philippines.

natural disasters in 2025

An insurance company called Aon publishes an annual list of natural disasters and their estimated costs. They also provide fatalities. Costs and fatalities do not track because the highest costs tend to be in developed countries and the highest fatalities tend to be in developing countries. Case in point – the most expensive disasters of 2025 were the January California wildfires (~USD 60B, 31-400 deaths depending on source), and the worst loss of life was the flooding in Southeast Asia that just happened in November (~USD 25B, 1750 deaths). Given what I know about real estate and informal land use in Southeast Asia, the latter was a massive disaster.

https://earth.org/2025-one-of-costliest-years-for-climate-disasters-report/

Other notable disasters were flooding in China, Pakistan/Bangladesh, and Texas; and hurricanes/typhoons that hit Jamaica/Cuba/Bahamas, Philippines, Australia, and Reunion which is a small island near Madagascar and Mauritius. Finally, a drought affecting Brazil makes the “top 10” list.

Every single one of these disasters is a type that will be more frequent and more severe due to global warming, I would say. Flooding, fires, and food – this is how climate change will hit home for almost everyone eventually.

The headline calls 2025 “one of the costliest years”. Is this right? It’s hard to say, but I got CoPilot to give me a plot of estimates released by Aon for the past 10 years. It wasn’t able to access Q4 for 2025 so I asked it to project the trend from the first three quarters, which is likely to be inaccurate. But nonetheless, 2025 doesn’t look particularly exceptional over the past decade. This doesn’t prove anything except that year to year variability is high.

CoPilot (ChatGPT4?) – not extensively verified by any human!

AI predictions for 2026

It’s easy to find predictions for where AI technology, the “AI race”, and the knock-on effects for the US and world economies might go in 2026. I find myself slightly fatigued from hearing about it, but nonetheless it is important.

Here’s one knowledgeable sounding blogger’s predictions:

  • Artificial general intelligence will not be achieved in 2026.
  • Robots will not be able to clean my bathroom in 2026.
  • ‘No country will take a decisive lead in the GenAI “race”.’
  • “Work on new approaches such as world models and neurosymbolic will escalate.”
  • The AI-driven stock market bubble may pop, or it may not. The exact words here are “the beginning of the end”. Well, I can predict with 100% confidence that the stock market will either go up, down, or stay the same.
  • AI will be discussed in the US midterm elections.

Okay, nothing too earth shattering here. On the subject of “countries in the AI race”, one perspective is that the US is focusing nearly all its investment on private sector AI, while China is spreading its investments across a basket of technology and infrastructure investments including AI, “electric vehicles, batteries, robotics, solar panels, wind turbines and other forms of advanced manufacturing” (“the Antimonopolist” blog). The US was also at least trying to do this during and after the Covid-19 pandemic era, but that sensible long-term strategy has been monkey-wrenched by a certain fool in 2025.

Then again, we could ask whether the basic econ 101 lessons are completely disproven? Is it possible we should invest more in what we are good at and sell it to others, while buying things from them that they can make better, faster, or cheaper? There’s a tension of course between being highly efficient and focused on comparative advantage, and also being diversified so you are resilient if something happens to upset your trade flows. But we are certainly not seeing rational debate about all this in the US political context.

Chartbook makes an argument that if you compare the US and European economies, it is really just the performance (measured by profits and stock market values) of the “superstar” US tech firms that makes the US look better. And while life at the top of the heap may skew the US numbers, quality of life for the average working European aided by their bumbling, stumbling social welfare systems is actually not that bad.