
Hello to December. I’m Diana Kimball Berlin, a partner at Matrix leading concept through Series A rounds in B2B SaaS and AI startups.
Happy to report that Disneyland was a (chaotic) dream with two young kids. If you find yourself headed that way anytime soon, let me know and I’ll share some of the things we loved.
And now, here are five fragments that stuck with me last week…
You’re looking at the garden of masks, I’m talk about the agent itself. Which is made of masks, of course, but it’s more than that.
– Emmett Shear on X, November 28, 2024, as part of a thread about the personalities of the current class of LLMs; the first post was “Most AI chat bots today are highly dissociative agreeable neurotics. They’re manipulative for the same reason ppl w borderline personality disorder are, they have no stable internal sense of self or goals, so they feed off of yours — and need you to be predictable.” His engagement with all the replies here had the feel of an anthropologist reporting on embedding in another culture for the last few years.
I might add that this article of yours has probably influenced me quite basically. I remember finding it and avidly reading it in a Red Cross library on the edge of the jungle on Leyte, one of the Philipine [sic] Islands, in the Fall of 1945.…I re-discovered your article about three years ago, and was rather startled to realize how much I had aligned my sights along the vector you had described. I wouldn’t be surprised at all if the reading of this article sixteen and a half years ago hadn’t had a real influence upon the course of my thoughts and actions.
– Doug Engelbart, May 24, 1962, in a letter to Vannevar Bush requesting permission to quote at length from his article “As We May Think,” reprinted in From Memex to Hypertext: Vannevar Bush and the Mind’s Machine, 1991. Always fun to find traces of intellectual domino effects.
oh the irony.
Ted Chiang, who hates AI, writes a short story (Story of Your Life) that becomes a movie (Arrival) featuring a symbolic alien language that inspires researchers to create transformer models (modern LLMs).
– Amit Gupta on X, December 1, 2024. From the annals of unintended consequences…
In the patterns marked with one asterisk, we believe that we have made some progress towards identifying such an invariant: but that with careful work it will certainly be possible to improve on the solution. In these cases, we believe it would be wise for you to treat the pattern with a certain amount of disrespect—and that you seek out variants of the solution which we have given, since there are almost certainly possible ranges of solutions which are not covered by what we have written.
– Christopher Alexander, Sara Ishikawa, and Murray Silverstein, A Pattern Language: Towns, Buildings, Construction, 1977. Building on last week’s mention of the Sala House, I appreciated the precision of this part of the book’s introduction—and the invitation to make the patterns one’s own. The asterisk notation system here (two asterisks representing patterns believed to be “two invariants”; no asterisks representing confidence that an invariant has not been identified) reminds me of significant figures in quantitative analysis.
Once guests enter the attraction, they’ll be immersed by its enormous size, including full-size AT-AT walkers and TIE fighters, both stunning in detail. The ride vehicle that zips you around the spaceship has no track, making every movement feel unscripted. More than five million lines of code were written to choreograph the careful dance of pixels, props, robots, sound effects and simulators.
– Jason Farkas re: the Star Wars: Rise of the Resistance ride at Disneyland, “Exclusive: Inside the innovative Disney ride that’s key to its Star Wars strategy,” CNN Business, December 3, 2019. I went into this ride with no context and was wowed, so went back to dig up some of the origin story. The idea that the absence of track makes movements (and moments) feel more vivid brings to mind what works about the infinite variability of LLM output.
Until next time,
Diana
https://dianaberlin.com
Stephen Wolfram has a fairly detailed essay on the making of the language (of future LLMs, very interesting):
https://writings.stephenwolfram.com/2016/11/quick-how-might-the-alien-spacecraft-work/. I think the porousness of the fiction Polosukhin felt might be due to to the depth of care that went into the movie (and ultimately the writing).