James Cham

Harvard Mark I - Wikipedia

en.wikipedia.org/wiki/Harvard...

Babbage’s Analytical Engine, the Harvard Mark I, and the atomic bomb

This is good advice. Now how can this show up in the feed of my kids without seeming linked to me?

YouTube video by Stanford AI-powered dog robots, built by students

CS 123 has hit the pareto frontier of coolest and cutest CS class that I have ever seen. I hope that other colleges (and perhaps even high schools?) copy their strategy of teaching students to build and train robot puppies. www.youtube.com/watch?v=3OHW...

The author @philipball.bsky.social has always written extremely well about science, but he has reached new heights in sheer beauty of publication with his new book, Alchemy.

Also: it has a fascinating story to tell, one you might incorrectly believe you know.

My best AI sensemaking trick is to turn things into tables. Ethan’s is better and more fun!

I suspect we are going to see something like this more often: “To conceal stylistic identifiers of the authors, the above text is a sentence-for-sentence rewrite of an original hand-written composition processed via Claude Opus 4.5.” (From possessedmachines.com)

An Un-MAGA Proposal to Bring Back American Manufacturing Follow the path of Italy, France and Japan.

I'm excited to announce I'm doing a new Bloomberg column about the business of fashion. The first column is about how we can re-shore US garment manufacturing. Instead of mass deportations and raising tariffs, I argue the government should move the industry upstream:

tinyurl.com/reshoringapparel

I see lots of Bi Gan search queries in my future…at least until YouTube gets the hint!

This is an old study and definitely out of date already; the next year is going to be disorienting…

The hard part isn't doing the work now; it's choosing the work. A day in the life of a relapsed agentic coder

Comparing abundant AI with a familiar experience:

> As food got safer, more accessible and cheaper, the key skill is deciding what & how much to eat to stay healthy. As thinking gets safer, more accessible & cheaper, perhaps the same will be true.

sundaylettersfromsam.substack.com/p/the-hard-p...

Continuing to have AI build a weird game demo a day. Here is: "Make a game where you have to prevent the apocalypse, but the interface is just Jira tickets"

Pretty fun/funny branching storyline, all text is AI created with minor feedback from me. Play: gentle-bienenstitch-01e24b.netlify.app

It is always hard to resist the easy, lower cost tactic that has deep, easily foreseen consequences. There must be a German word for it--Claude says it is milchmädchenrechnung.

There’s a @dsquareddigest.bsky.social-style cybernetics information flow insight in Ezra Klein’s interview with Yuval Levin about one of the (many) problems with our media and governing ecosystem.

What was the science fiction story about a society with special monks who spend their days cultivating deep domain knowledge while cut off from the culture so they can be clean judges of the AI models?

Sometimes the long arc of history means that we are stuck with some bad decisions someone barely paying attention made 50 years ago.

I think it’s a really big challenge!

Software development trying to adjust to the new world...

And the optimistic part of me hopes that they come up with good solutions the rest of us can copy!

You can see this happening even faster with open source software x.com/simonw/statu...

Think of the crew of the USS Grissom, Kirk’s son, and all of the Klingons who died for the sake of reviving Spock. And think of all of the bad moral calculations that resulted!

There’s something wrong with our current moral calculus, where we are willing to sacrifice entire actual populations in exchange for theoretical individuals. I blame Star Trek III.

C Thi Nguyen-style “value capture” should be called “values capture”. Makes it easier to differentiate from the economic idea of value capture.

That’s a good point!

Attention is a limited resource. Is affection?

I need a better genealogy of good things

Writing with machines (wk 3/2026) Tacit knowledge, motivation, and negotiation; writing with machines; Smart Bricks, being kinded, and specs.

Reviewed my writing from the past year vs 5-10 years ago. Recent stuff is worse despite decades clarifying the concepts. Problem: machine-generated prose appears so finished I've become insensitive to when the substance misses the point I'm trying to make.
uncertaintymindset.substack.com/p/gbw32026

Started reading it last night!

I bet @californiajoe.bsky.social would find this thread interesting…

DEADLINE: December 31 for L.A. Residents to Register for Democratic Lottery to Deliberate on Charter Those Selected Will Deliberate January 11 on a New Preamble and Bill of Rights for City Constitution

There is a group in Los Angeles working on revising the city charter using lottocracy. democracylocal.substack.com/p/deadline-d...

Just starting Nguyen’s book The Score and see that the kindle gave me a score for using the app four days in a row.