What's old is new again
I am old and used keypunch machines and teletype machines back in the day. But computers really took off for me when I got my own video terminal and a command-line prompt. That is how I started with the Apple ][ and IBM PCs, that is how I started with Unix systems.
And now the CLI is hot again. It is the primary way people work with Claude. I just installed cmux, which is a very natural way to work with the Claude CLI — multiple sessions, an integral browser window, vertical tabs.
cmux does make the terminal look kind of like an IDE, and you could configure your IDE to look exactly like cmux. But the core purpose of an IDE is editing files, and increasingly that is NOT what we are doing when we develop software – instead we are having a dialogue with our tool, and monitoring the progress of multiple sub-agents. And cmux is driving exactly at that notion of work.
I did not think I would be installing a new terminal app in 2026, but … Maybe my expertise in editing config.sys and win.ini will come into vogue again.
Also Remember…
Minecraft was such an awesome game in its day (and still is). I “wasted” a lot of time building my Minecraft fortress. I was recently watching a show about a home in the Seattle area, the home of a Minecraft engineer, and he originally designed the home in Minecraft before handing it off to his architect. Oh, and here is a project to generate a Minecraft world based on real-world geography and topology.
Here’s an exhaustive history of the Sound Blaster card ($) (if you have a Stratechery sub, this article is included). I have many memories of struggling with sound card installation, but I still did it because it made the gaming experience so much better.
Ever wonder what happened to SETI@home? They are still cranking through all the data we processed, looking at the top 100 most promising candidates.
Materials and AI
Shifting gears – I am a sucker for material science. The computing industry has been enabled by remarkable advances in materials science. And now, materials science can be accelerated by amazing levels of computing power. Noah’s conversation with Claude around “What are the 10 to 20 most transformative advances in materials science that might be made with the aid of AI in the next few decades?” is a fun read. It gets a little abstract in the end, but I enjoyed the discussion of possible material breakthroughs over the next 10-20 years enabled by AI. There is also an assertion in here about the utility of AI, which seems interesting:
AI is most transformative in fields where the ratio of search space to conceptual confusion is high. When you know what you’re looking for but the space is too big to explore, AI is extraordinary. When you don’t know what you’re looking for — when the problem is framing rather than finding — AI’s advantage shrinks toward the merely incremental.
Electrification
And shifting gears again – The USA could have gone down the path of being a leader in electrification and renewables. We could have funded solar, wind, and geothermal projects, funded nuclear, and pushed our industries to move to electrical power, in the process dramatically reducing our dependence on Mideast oil.
Instead we are shuttering investments in renewables (at a high cost to the taxpayer), cozying up to the Saudis and letting them push our foreign policy, spending $25B to bomb Iran and to try to keep the Strait of Hrmuz open, driving up oil and gas prices, and creating a whole new generation of martyrs fueled by hatred of the US. This is the stupidest industrial and foreign policy that we could have followed.
The US is not safer, nor has our standard of living improved.