The Deep Blue Sea on a 35mm print is quite divine
Posts by Corey Atad
like from another universe
this is because i finally watched Damnation
good things are possible
exactly, it's that commitment to what TV can do. similarly, Sopranos was taking the form of TV soap and bringing into a space of self-aware cinematics. the cinematics push and break the form, but the form is there.
yes! that's the perfect term for it. because when you think of what's digressive in The Wire, it's not flashbacks or cinematic tricks like that, but unusual indulgences in writing, which feels like something out of a great novel. talk about breaking the rules of TV.
and crucially, those shows descend from the real roots of the TV tradition, in live performance. in fact, the difference between The Wire and The Sopranos is that The Wire was really firmly rooted in American theatrical realism, while The Sopranos took classic TV and brought Lynchian surrealism.
few movies or TV shows these days trust enough in basic narrative power: the protagonist needed something, and here were the obstacles along the way and how they overcame them (or didn't). everything is born from that, and the audience will infer everything else.
Taylor disappears after the first 35 minutes, DiCaprio and Penn meet exactly once, del Toro and Infiniti only one scene together, three scenes with the Christmas Adventurers. and there's certainly exposition, but it's handled directly, within the plot, while character is built through their choices.
in general, we're suffering a lack of TV and filmmaking that trusts in the development of character through action. One Battle was a masterful example the craft. DiCaprio and Infiniti have one scene together before their reuinion at the end. one scene. it's all you need. action handles the rest.
it's mostly mistaken thinking about the kind of exposition necessary to sustain a longer narrative, and a lack of trust in the development of character through action
Anti-Zionism is rising as a response to what Israel is doing. It will simply not be possible to treat it as a marginal viewpoint that can be shamed or shunned into invisibility. Yes, antisemitism often cloaks itself in anti-Zionism. So don't do the antisemites' work for them. If you keep telling people that if they oppose the Jewish state then they must hate the Jewish people, eventually, they will believe you.
I don’t often find myself agreeing with Ezra Klein, but this is completely correct — and frankly, shows some guts for a mainstream liberal columnist to say. www.nytimes.com/2026/04/12/o...
WIRED sat down with “Democracy Now!” host Amy Goodman to talk politics, AI, and the future of media. A documentary about her rise hits theaters today. www.wired.com/story/the-in...
nope, too late
lotta good things lining up
btw, i also edit video podcasts now
shoutout to all the CR heads! Chris Ryan is on the new episode of @talkeasypod.bsky.social, which i had a lot of fun editing
whoa, these feel like something out of a revisionist western
Brody places Blue Heron in conversation with Nickel Boys, The Mastermind, and The Secret Agent as a memory film, but also he's talked about those movies as best of the decade material!
anyway, Blue Heron starts playing theatres April 17, and you really must go see it if it opens near you
watching your friends do something extraordinary and have it be appreciated is a 10/10 experience, highly recommended
Sophy said she woke up to a rave from Brody, but this is a RAVE
yeah, in a funny way, this is just a more intense and directed version of the use i've found in the otter.ai summaries, which is that i don't actually *need* them, but in the past, if i'd had a secretary doing my transcriptions, they might have also created a little summary. it's helpful.
today is The Deep Blue Sea and Sunset Song
really the only "AI" i use is transcription, and then the summaries it spits out. and even then, i didn't ask for those summaries, but after a certain point i started looking at them and finding them marginally useful.
the key benefit of "AI" is that it improves its output by iteration. this has tons of applications where simple, direct coding wouldn't be able to account for the variability of the input. hence, great for processing photos. but they'll destroy the world with bad output claiming it's a cure-all.
which, btw, fits with how machine learning has worked for many, many years now. amazing at onboard photo processing for phones with tiny, shitty camera sensors and lenses, but reaches clear limits when it's Siri or something general purpose like that.
none of these LLMs are very good is the thing. even on the things they're best at, they're still just mediocre (but fast!). AI only seems to gain deeper utility when it's more localized, highly trained models made to do *extremely* specific tasks.
facts are facts
did you see the French movie from a couple years back?