Yes, what we have access to today is definitely headed for commodity. I doubt the current levels of investment at the leading edge can be sustained indefinitely, but given the rate of progress I'm hesitant to predict very far into the future.

Of course, SGI eventually collapsed, in 2008, after a decade of difficulty, but much of the talent and technology on which NVidia was built migrated from SGI.

The specific capabilities (e.g graphics) are entirely beside the point. Employers were prepared to pay a significant premium over direct salary for great tools in many domains. The toolmakers, SGI were also losing hundreds of millions of dollars annually. e.g. www.theregister.com/1999/10/21/s...

Yes, a real problem. I wired the house for Ethernet and have WiFi access points on each floor.

The Weaire-Phelan Structure is the best known solution to the Kelvin problem of tiling space by equal volume cells of minimum surface area. I started out by using more artisanal visualisation techniques, but it soon became clear this would be massive time committment.

We entered the 1980s writing in assembly, entered the 1990s writing in C and C++.

A good reason to use pen and paper.

Key scientific tools: Hypothesis formation, experimental design, controlling for confounding factors, parsimony of explanation, replicability, critical evaluation of evidence ... even iterative model refinement originated in Software Engineering.

Bleats better captures the tone.

As the author of a few books (and a programmer), the trade-offs are in favour of AI. There's no money in 99% of book authorship. My writing is likely to have more reach and societal benefit as a result of being consumed, remixed and regurgitated by AI than being read my 10000 book-buyers.

Indeed!

Sounds like the trial succeeded but the drug failed.

The climate thanks you for your restraint.

The city of Drammen, near where I live in Norway, has had a 14 MW heat pump for district heating since 2011. New buildings are required to use it. The heat is taken from the fjord.

en.wikipedia.org/wiki/Drammen...

Claude Code, mostly.

In fairness, I was asking it to do some stuff that involved git, it’s just that is more imaginative than me in how it should achieve the goal I gave it. Still learning!

In mitigation Claude also knows how to fish the ‘permanently deleted’ files out of the innards of Git, which I certainly don’t.