What I learned this week 2026/34

Nautical Miles, Knots, Earth's Circumference

Nautical miles and knots both relate to earth's circumference. If you travel 1 nautical mile on earth, you've moved about one sixtieth of a degree. This is called a minute and is the part in GPS coordinates indicated by the prime symbol (′). So if you go from 12° 34′ 56″ north to 12° 35′ 56″ north, you've traveled about 1 nautical mile. If this took you 1 hour, your speed was 1 knot.

The meter is connected to earth's circumference in a similar way: The earliest definition of the meter was "one ten-millionth of the shortest distance from the North Pole to the equator".

So the same distance from North Pole to equator was defined as 10,000,000 meters and as 5,400 nautical miles (90° times 60). And now – without remembering any number – you can always calculate how many meters are in a nautical mile: 10,000,000 / 5,400 ≈ 1,852m.

Movie-making is more cooperative than you think

There's a cooperative aspect not only within a crew working on a movie, but between everyone creating feature films.

If audiences get in the habit of watching a movie per week, every filmmaker benefits. The surest way for them to form that habit is, if they expect there to be a great movie every week. But a single filmmaker can only make a movie every one to three years. So every filmmaker actually has an incentive for all the other films to also be great.

According to Ben in the Acquired episode on Disney's renaissance this is why George Lucas wanted to make movies better for all filmmakers with what would become ILM and Pixar.

The Pixar genesis is even wilder

Every time I hear about how Pixar got started, the story gets even wilder than what I already knew. Also from the same Acquired episode and a tiny bit from my own research:

Ed Catmull was one of the first hires when George Lucas set up Lucasfilm's Computer Division in 1979. John Lasseter, fired from Disney after pushing a computer-animated feature over his bosses' heads, gets hired by Catmull under the cover title "interface designer" – in fact to bring in animation know-how.

Lucas needs cash after his 1983 divorce but won't touch his core film business, so he sells the computer division.

Steve Jobs is interested but goes quiet while he's dealing with being pushed out of Apple. A deal with GM/EDS and Philips falls apart at the last minute, and in February 1986 Steve buys the group for about $10 million total ($5M to Lucas, $5M into the company) against Lucas's $30M asking price. His original vision: a hardware company selling the $135,000 Pixar Image Computer to hospitals, agencies, and studios.

In 1991 Pixar agrees to the Disney deal: three pictures, Disney funds and owns everything, Pixar gets roughly 12.5% of revenue. Under Katzenberg's notes to make Toy Story "edgier," Woody turns into a jerk, and at a screening in November 1993, Disney hates it and halts production. Lasseter asks for two weeks to fix the script ... and delivers. Production doesn't fully restart until around February 1994, with Jobs personally keeping Pixar afloat in the meantime. The new version wins everyone back.

As Toy Story shapes up, Jobs thinks the movie could become a hit: If they did nothing, Disney would realize they've raised an actual competitor: a studio that's able to create a successful animated movie. The only sensible move for Disney in that situation would be to re-negotiate the deal and bind Pixar even closer to Disney.

To prevent this from happening, Pixar needs leverage. Ideally in the form of its own capital to co-finance films 50/50 and renegotiate with Disney as an equal partner instead of a work-for-hire studio. So he times the IPO for November 29, 1995; exactly one week after Toy Story's release. The plan works and they raise about $140 million, setting up the 50/50 co-production deal signed in 1997.

Everything is a YouTube video

When you upload a video to YouTube, it gives you a really interesting chart: How many people keep watching the video up to which second. It's called the retention graph and YouTube creators study this graph meticulously. The graph tells them when people were watching and when they clicked or swiped away.

YouTube lists a few examples of good retention and retention issues:

For online videos this is easy to measure. But many other areas have their own retention graphs: a landing page, a book, a sales process, the onboarding for an app, this blog post (hi, great that you're still reading), a meeting, a presentation, an email, ...

For most situations it's difficult to measure and you'll never get such an exact graph. But if you could measure it, the second-by-second retention rate would make a great metric to optimize for.

So if you're able to create an online video, that hooks people and keeps them watching, you'll probably succeed in these other areas too.

AI gets used on low-priority tasks

I heard Alex Hormozi say that AI gets used on low-priority work. Put another way: Tokenmaxxing without creating value. And this is something I see happening too. The hard part still is deciding what to do, before deploying an army of agents on tasks that are all nice-to-haves.