TIL about caffeinate a built-in macOS command-line tool that prevents your Mac from sleeping.

Here are some common usages

• caffeinate -d - prevents the display from sleeping, but allows other parts of the system to sleep.
• caffeinate -s - prevents sleep when there’s an active user session (helpful in remote work or screen sharing).
• caffeinate -dt 600 - prevents display sleep for 10 minutes.
• caffeinate -i script.sh - keeps system awake only while the script runs, then exits.

via Scott Knight

In Calculators & Writing, Chris Coyer points to something that’s been lingering in the back of my mind when using AI for certain tasks. My partner actually brought this up to me a while ago, and at the time I thought they were wrong – but I’ve come around. The quote below gets to the heart of it:

… with writing, humans are thinking, feeling, and communicating. Computers do not do these things.

Using AI to write means you’re robbed of the thinking and feeling it takes to write, which is (most?) of the value. Communicating, I suppose, you’re doing either way, you’re just communicating something you didn’t think about or feel if your writing is generated, which is fucked.

Despite remarkable progress in AI, some tasks still demand our effortful engagement. Learning, for example, requires one “to make eye contact with the idea” because the value lies in the effort, not just the outcome.

Related:

• Don’t delegate understanding.
• Understanding requires effortful engagement.

The New York Times has a wonderful review of the effects of congestion pricing since it began: Here Is Everything That Has Changed Since Congestion Pricing Started in New York.

It’s been a policy miracle. Here’s hoping the success continues before the current administration does something reckless to jeopardize it.

Related: Congestion Pricing Tracker

ASI existential risk: reconsidering alignment as a goal.

A lot of smart people are worried that future AI systems could become dangerously misaligned with human values. The concern is that as these systems gain power, their goals might drift from ours–leaving humans vulnerable. Researchers in the AI safety world are working to prevent this by aligning systems with our values. At the same time, there’s another group of researchers, equally brilliant, who find the idea of a rogue AI implausible and mostly dismiss the risks.

Michael Nielsen’s essay presents a nuanced thesis: artificial superintelligence (ASI) may pose an existential threat, but not in the way current AI safety research tends to frame it. He argues that alignment research may inadvertently accelerate the path to such powerful AGI; and that while those efforts might help mitigate certain dangers, they also obscure or neglect other existential risks.

At the heart of his argument is this striking line: “Deep understanding of reality is intrinsically dual use.” Nielsen explores how this dynamic plays out as science and technology advance over time. The essay is accessible, thought-provoking, and by far the most insightful piece I’ve read on the potential dangers of frontier AI systems. Highly recommended.

When thinking about current tariffs, the main talking points have mostly focused on manufacturing jobs. But lately, I’ve come across a set of arguments that highlight a different concern: the U.S. lacks sufficient industrial capacity.

This becomes particularly relevant when considering the possibility of conflict. In such a scenario, the ability to produce essential goods rapidly is critical. If a country is dependent on a rival for key supplies, that reliance could become a serious strategic vulnerability.

Here’s some exerpts that gesture at this issue.

Dynomight’s notes on Landmark ruling on the WTO national security exception

… it doesn’t seem crazy to argue that you need to maintain some industrial base for the sake of national security. Recent history unfortunately shows that brutal land wars between rich countries still happen and still require enormous quantities of matériel. According to some sources, Russia is using around 10k shells per day in Ukraine, while after several years of ramping up production, the EU hopes to produce 5.5k shells per day in 2025 and the US 2.5k. In 1995, the US could make 22k shells per day.

In other words: in 2025, the US makes about 1/10 of what it did in 1995.

Samuel Hammond on his NYT essay, Tariffs Won’t Fix Our Trade Imbalance. This Will.. The whole essay is insightful.

Republicans and Democrats now recognize the need to rebuild America’s industrial base, if only for national security reasons. The United States cannot deter Beijing if we are dependent on Chinese supply chains for our military drones. But reindustrialization means more than protecting existing factories behind a tariff wall. It requires building new industries and pushing them to aggressively compete with rivals, scale up production and export their products. Countries don’t get rich by creating expensive substitutes to cheaper imports; they get rich by making things the world wants to buy.

…

Instead of cutbacks and tariffs, what America needs is a full-spectrum industrial strategy: an Industrial Finance Corporation to invest in tradable sectors, an expansion of export credit guarantees and the creation of special economic zones to cluster suppliers and lower input costs.

The Ezra Klein Show episode on Why Trump Could Lose His Trade War With China

… what you see when you’re there is the product of 30 years of being in the fitness gym. It works like this: A new industry comes along. Let’s call it solar panels. Every major city in China decides they need a solar panel factory. The local government subsidizes it — maybe domestically born, maybe in partnership with a foreign one.

And you end up, in a very short period of time with — I’m making the number up, but 75 solar panel companies. They then compete like crazy against each other in the fitness gym, and five of them survive. Those five are so fit that they can then go global at a price and level of innovation that is very hard for a foreign competitor to deal with — which is why China today basically controls the global solar panel market.

But what you also don’t see is that process of winnowing down from a hundred of those solar panel companies to the five produced a massive explosion of supply chains domestically to feed that industry.

The same thing went on with cars. The same thing goes on with robots. So where you end up five years later is with an interlocking set of supply chains that now, if you’re a young Chinese and say: I just got this idea. I want to produce a shirt that has a pink polka dot button that can sing the Chinese national anthem backward — someone will have it for you tomorrow.