Lego is interesting, because on the other hand it is very accessible (to a degree), but at the same time it tends to shift projects from making a thing to working around the limitations of Lego. Kinda reminds me of bash programming, it has similar characteristic of being attractive because it is there, but then actually engineering around the limitations of bash becomes quite the exercise once you start implementing stuff.
Working around the limitations of your materials (including skills, tools, budget, support network, ...) is the story of making anything.
Legos are intended as a kids’ toy, and a few years of experience working within their capabilities and constraints to build increasingly tricky projects is great experience for kids, with lots of skills transferrable to any other kind of creative technical work.
Excellent. I do feel like there's been a bit of a rejection of the polished, streamlined, "Facebook/Reddit" internet of late, and I love reading articles like this.
"I wonder if I can do..." ponderings, followed by a wonderfully absurd, but still mostly functional way of doing it, that involves going down all sorts of fun little rabbit holes to work around nonsensical but entertaining constraints like "Lego gears don't exist in an infinite number of forms."
Though as much as it doesn't need to be https, I still think random personal sites should be https. It makes it harder to inject random bonus conntent in the responses, and I include any form of modification there, not just malicious - ISPs have been known to make changes to add their own little ad banners in the past. Denver International is particularly bad at this, or at least used to be.
I feel a bit sad when I read about this sort of stuff. At one time it was the sort of thing I would do, but nothing seems interesting or motivating enough these days. My first thought is "wow that would take a lot of time" and I abandon the idea.
Maybe start again with something small or something that starts halfway through that you can see to completion. Maybe you don't have to build a sailboat to learn to sail. Maybe I am talking to myself...
The author mentions they were inspired by the Veritasium video on the same topic (without the Lego). [1] I thought I'd link the previous HN discussion. [2]
I'm a big fan of "mechanical solutions to computer problems." I think it's extremely interesting to see how hard it is to encode logic pre-transistor, and makes me feel spoiled today (in a good way).
My favorite example is the writer automaton [1]. It's something that never stops to impress me, just because it's something that wouldn't be that hard nowadays, but must have required effectively PhD levels of understanding to do pre-computer.
And in a couple centuries they're going to find this at the bottom of the sea and wonder what went wrong. (In reference to the "greek" computer that tells tides/lunar cycle and planets.) If plastic will last that long, (I think it does?
But still, It is fun and I like these lego projects! I want to buy a few sets and recreate it.
In the video comments he posted a link [1] to his blog where he explains the design process using simulated annealing. Amazing how close he got to the targeted final ratios using only a sequential gear train.
That was my immediate thought - the poor guy is stuck working with canon Lego gear sizes which makes his job way harder. 3D printing would have solved this, but then produced something not replicable with regular Lego sets.
