Personally I was a big fan of Prevent Phishing and Impersonation with Trust Loops, delivered by Karissa McKelvey. The thesis was a fairly radical rethinking of how online identity works: instead of everybody having a single canonical profile/identity that connects to everyone else’s canonical identity, peoples’ identities become functions of specific interpersonal relationships. So if I want to talk to Bob from the gym, instead of getting his single canonical phone number we exchange some kind of single-use connection token (via scanning QR code or whatever) and communicate with each other over that newly-established channel. I would probably call that channel “Bob Gym” which is how peoples’ numbers tend to get stored in my contacts anyway. And I could revoke that channel whenever I want, and Bob can’t turn around and give my contact info to a stalker or whatever. And my conception of who Bob is is separate from who his family thinks he is, or non-gym friends think he is, and so on. Is he the “real” Bob? Who cares. Such questions have no meaning here.
I really like this concept because it maps very well to how I socialize in real life. It also has a glint of antiquity in it, where the concept of a single reified identity didn’t really exist for people at any level of society. Of course it’s difficult to figure out how broadcast-style social media could fit into this scheme but maybe it doesn’t have to.
I remember being blown away by Chuck Moore’s talk on Greenarrays when I saw it in person. I don’t know if I’ll ever use a system like that, but it was really eye opening and inspiring to see what radically different models of computing are possible.
thanks for sharing! gives me TIS-100 flashback
Jean Yang’s Building Observability for 99% Developers is still a talk I regularly revisit. The philosophy inside is what drives the design of the tools I create.
We have spent 3 decades ignoring user research in dev tools. This is finally changing thanks to people that worked on Bundler, Rust, Typescript, VSCode and a few others. But it is still a really wide field. Jean Yang’s talk give a glimpse into it and why it matters.
I really liked Production Prolog, which spiked my interest in real world Prolog. Something that lasts to this day.
I have a soft spot for miniKanren, and other logic languages and systems. Thus, Will Byrd’s 2021 talk Strange Dreams of Stranger Loops is a great hit for me.
Nada Amin’s talk mentioned in the post is probably my favourite, and Amar Shah’s Pointfree or die is a close second.
I really like “The Security of Classic Game Consoles,” https://www.youtube.com/watch?v=s0XmiXs8iRw . I had studied video game history including this stuff, and even wrote a report in middle school about the Nintendo-Tengen battle, titled “The Center of the Universe does not Leave Luck to the Gods,” playing on the Japanese names of both companies. But this talk explained it so well, so briefly, and with actual physical specimens.
This eposide (event?) has some really interesting entries. Especially Languages for 3D Industrial Knitting.