Sachin Benny argues that AI modernity is reorganizing our relationship to the past the way industrial capitalism reorganized our relationship to the future, coining 'Archival Time' as a condition where traces of prior events become retrievable without recreating the event. He contrasts this with Bakhtin's 'Carnival Time,' bounded periods where accumulated traces lose authority and things get unmade and remade, drawing on William Sewell's eventfulness framework (path dependency, temporal heterogeneity, contingency) to argue eventful history requires a mix of both archival inheritance and carnivalesque rupture.
Shared by the speakers for the symposium. Email addresses have been removed.
Transcript
YouTube’s automatic captions, unedited, so names and technical terms are often misheard. Each timestamp opens the video at that moment.
0:00 U that’s okay because we put the Zoom link um in Discord and everybody was able to ask access it there access um from the Luma invite. So yeah, I think since it’s just past the hour and I don’t want to cannibalize your time, Satchin, you can go ahead and get started. I’ll do a little more troubleshooting.
0:22 Cool. Uh hey everyone, good morning. Um, welcome to the fourth day of the symposium. Um, [clears throat] yeah, so little intro, I’m Sachin. Uh, I write a newsletter called summer lightning and um, I co-host, uh, two special interest groups in the, uh, protocol institute discord. uh the uh protocols for business uh special interest group with Rafa and the protocol fiction sync with Spencer. Uh in the pro uh protocol for business we’ve been mostly um talking about AI adoption and actually working on AI adoption and protocol fiction is mostly around theory and practice of writing the genre we call uh protocol fiction. Um, cool. And um, uh, I’m going to jump around the tabs a bit uh, instead of completely using the slides. Um, but um, yeah, the title of my talk is archival time and caral time. And I initially got interested in this topic from um like around a year ago when um I think it was EMTT Sheer who had a thread about um LLM models being sort of like uh cultural like archival slices of uh a particular moment in culture. So um LLMs are each LLM model is trained on a large corpus of culture and uh basically the the sort of slice you access is essentially the point at which they release the LM or stopped the training.
2:14 Um and um so he said he argues that essentially that’s why models feel dumber over time. Uh I don’t think this holds up uh that much any more technically because now you have um sort of like more advanced behavior and rag agents and things like that that sort of uh messy this a bit mess this up a bit. Uh but I still think it it’s a useful way to um think of LLMs as these um archival slices that are overfit at a particular moment in time. And at the same time I came across this idea of uh carnival time and um this came from um uh this book by Mikail Bakton called Rebelius and his world and Bakton is a Russian literary critic uh he was writing in the early 20th century and uh in this book he’s essentially literary theory about uh this author uh rebel who wrote um who’s most famous for writing uh Gargantua and Pantagro uh which are these sort of um medieval u comedies um and uh he introduces the idea of carnival time as uh essentially um sort of um bounded periods where normal rules of society don’t apply and uh people are uh sort of allowed to like allowed to both sort of celebrate, [snorts] indulge in exorcist and also invert the normal order of society. Uh and um Bakan makes the claim that a lot of uh medieval culture prior to printing was sort of centered around this uh carnival culture and there were many carnivals throughout uh the year that people participated in. And um the core idea he sort of introduces is like this inversion of uh uh like the normal aristocratic order um the the the the clown becomes the king, the king becomes the fool. And um and u in the books itself there’s a lot of sort of um grotesque uh like body humor uh a lot of like fart jokes and it’s a u he calls it sort of like this bounded period where um [clears throat] things become more looser and uh the normal order of society is inverted and um I’ve written several articles on this and um and recently I sort of got interested in this idea of um Evanfulness which is which we’ll get to uh which is a paper by um William Su uh and um yeah but um we’ll get to that next. So um so yeah so essentially with this talk I sort of want to connect this archival time and conal time idea to um eventfulness. Uh and to the for that I think starting with the definition of an event as like an event is something that generates new information that changes what happens next. Um I’m sort of simplifying some of the definitions I found from like uh Whitehead has this uh idea of [snorts] actual occasion um which feels like very close to uh this idea of an event and uh uh this idea of a trace a trace uh is something from a prior event or state that remains available after the event and can shape later action.
6:14 sort of come it comes from um the stigmology discussions that the formal protocol theory people have been having. Um I attended a bunch of the earlier discussions and um the uh and some of the readings on the earlier papers and the idea well the sort of canonical example everybody uses is like ants and pherommones where um ants leave pheromone traces which other ants can use as like an affordance to like do like themselves. So um so that’s the idea of an event and a trace like events leave traces and um uh yeah then these like these traces can be um affordances for other actors it can be a particular form of memory uh it can be protocols um etc and um and yeah so with this I was thinking that maybe a better way maybe an easy way to understand archival and carinal all the time is to sort of think of it as sort of like interfaces for how we sort of uh deal with uh sort of u culture or uh uh like um yeah interfaces for like our cultural experience of time and um so an archival interface is one where like traces from previous events can travel across events uh it can be retrieved later without recreating the event that produce it. And a carnivalist interface is one where um essentially the trace is reactivated through recurrence. So Bton’s carnival is sort of a good canonical example where uh prior to printing and prior to you know you’re not like recording the um events u literally. So uh to remember the event you have to essentially recreate it or it’s like this embodied memory that you have to uh recreate and maybe a more online example would be I would consider this like yeah like we have this event where uh we have the live streaming we have the rag agent that is trained on all the corpus of uh content that people have submitted and um um the sort of a it’s like archived and even though you can’t recreate the exact experience, you can recreate like something very close when if you if you just watch the videos back. Um but and a good online example for a carnival interface would probably be um Sarah friend had this paper on um called good death uh in the first year of summer protocols and u she uses this example of um um players of a particular video game uh mourning like the death of uh a video game and um yeah I think it’s the game was called Ashon’s Call and uh um and uh she has a good video of like essentially people gathering at the same location within the world of the game u to sort of remember and commemorate it and that’s you could like watch it on again on YouTube but like in some ways you had to be there to sort of it it was um like embodied in some way even though it happened in sort of like a disembodied uh medium. Um yeah and uh uh and I think it’s hard to sort of think of um archival time and carnival time because like we essentially live in a time where we have access to a lot of uh externalized memory in the form of uh books you know like internet lms like we essentially live in a extremely in a in an environment that is extremely rich with external memory. So maybe a good example is like something like far before um printing and uh I got this example from the memory research group that uh KIA runs. Uh I highly recommend attending that. Um and um um this was this idea of like bas a story of um in um in Daphne in the 4th century there was sort of this conflict between uh Christians and pagans. Uh and um the Christians had um a martr named Babilus who they moved into um a temple that was um dedicated to the god Apollo. And um Emperor Julian who uh was pagan like removed uh those relics and essentially tried to restore the pagan order. And um in this book by Mary Karutas, she says that essentially this whole uh uh sort of uh conflict um played out in processions carried out by both sides.
11:37 So each side would uh carry out a procession and then the other side would do it. And um during one of these uh processions uh the temple of Apollo burned down and the the Christians interpreted that fire as like the triumph of Babyl has this great line about how um citizens became figures in a pneummonic sense and meaning was carried by sort of like the repeated arrangement of sight root bodies images and participation. Uh and I think that’s like a beautiful example of sort of um carnival traces where you like repeating the same event sort of creates uh uh creates the memory and uh without it the sort of the the trace sort of uh fades away. Um yeah and I think um uh I think like sort of like pure moments of uh like caragal or archival time are like hard to find. I think every moment has uh mix of both some more than the other.
12:49 And uh I really like this uh quote from um uh Jacob’s room by Virginia Wolf which I read this year. Uh wouldn’t you like my knife mother said Archer sounding at the same moment as the bell her son’s voice mixed life and death inextricably exhilaratingly. And I think I in this line you with with the line sounding at the same moment as the bell she has both she has like clock time. Um she sort of has this exist existential uh social time moment with her son and also sort of like the connotation of the bell as like you know like related to funeral. Uh and in in like like just couple of lines you get sort of collision of um like multiple experiences of uh time. Um okay yeah now sort of switching into the eventfulness u part um so um so um su defines evs a little differently I’ll get to that later but uh his main idea is that um essentially um history unfolds in this sort of uh history unfolds through like events that change um sort of the the underlying structure of the the system or society. And he in the paper itself he’s arguing against other uh types of ways [snorts] of looking at history such as like comparing two different moments in time.
14:40 Um and uh I I won’t get too much into that because um yeah I think uh yeah like it’s easy to sort of get lost in that but the serial idea of eventfulness is that um um yeah like so history unfolds through uh these events that sort of remake um the structure uh of a culture [snorts] or society. And um uh and maybe an easy easier way to understand is like these three features of eventful history. Uh so the uh the first so it’s part dependency temp temporal hetrogenity and contingency and um the idea of path dependency is like earlier events change later possibilities. So you can’t com you can’t directly compare the French revolution and the Russian revolution and say that like the same uh type of causality was at play because people in the Russian revolution are already influenced by the ideas from the French revolutions and those ideas have seeped through. So there’s like path dependency. Um the other one is uh temporal heterogenity where um like the same um uh the same thing can have different cause and effect. Uh and a great example he uses is before the French revolution u noble status was a prime social asset but by the time of the terrorists uh the same status was like a big liability. So like it it completely flipped poles. And u the third one is a contingency where um regardless of like you have a part dependency temple hetrogenity but there there might be events uh that happen in a short span of time that sort of completely change the trajectory uh of the future. And he says um one example he uses is in the uh Paris revolution of 1848 uh there were these two uh workers who essentially came from same simil like very similar backgrounds had similar social condition social and financial conditions but one ended up supporting the revolution and one ended up opposing the revolution. And he says the reason the it was the national workshops that opposed the revolution.
17:07 And he says the reason was that um the mobile guard was sort of like inside uh uh these um uh barracks and had sort of their demands um met uh and they were pretty happy while the national workshop uh during like a particular 3-month period had their leader jailed and um uh essentially like that sort of created like this uh uh divergence in uh where they ended up. Um yeah, just to clarify sort of like uh how I’m using event and this is historical event. Um uh yeah, just pause for a second.
18:01 Yeah. And um I think um to sort of combine a couple of these ideas um like like eventful periods are essentially uh sort of like vortices of in like historical flow of time where you have path dependency. you have like multiple channels of like uh path dependency and different currents uh coming to the uh same uh moment and you essentially sort of get like this bounded period where the previous uh things don’t like the previous uh things don’t matter as much and uh and you sort of get like this carnivalist kind of atmosphere uh that’s like bounded where so you get an unraveling of previous uh systems uh and like uh institutions and uh sort of previous like the the status quo and um I and I thought this neatly connects to both the idea of archival time and carnival time. So sort of the definition I’m kind of running with now is uh um yeah archival time is the temporal world produced when traces of prior states remain available across situations and become part of the environment. Um and carnival time are sort of like these bounded intervals in which like the accumulated traces uh have like less authority over what can happen next. So you have like more sort of uh possibility and um um yeah essentially sort of uh um like a um bounded uh interval in which things sort of get unmade and uh remade again. Um and and I think important to sort of clarify that uh the reason I did not use the word memory is that carnival does not necessarily mean low memory. It just means memory that lives uh elsewhere. Um I watch I watch a lot of football. So I use the football example. It’s like uh um you know there are like English clubs that are 200 years old and uh the the main the only way those stay alive is sort of like embodied memory of fans that is like recreated every time um there is a game. Um and uh um uh so and that memory is like you like some of these u football clubs are uh 200 or 250 years old which means they have been successful and without that embodied memory you would not have uh the same uh football club. Um yeah and uh yeah and um yeah I think from here on these are all sort of like working theories so I’m not like too uh sure about them and still still sort of working through this out and um yeah I think like eventfulness essentially needs uh both like um some sort of like inheritance uh of like previous traces um and sort of like the bounded carnival uh atmosphere. Um and u because I think sort of like purely living in archival time would essentially sort of lead to um sort of like very unsurprising future. Um and uh pure carnival time would essentially leave no memory and what you sort of get is just um chaos.
22:01 Um yeah and um yeah these were some of the ideas I was working through of like what if um this sort of like bounded carnival event um does not happen. And I sort of call this idea um [clears throat] GSKY uh theme park where um essentially the bounded carninal does not happen. And so you sort of get like this obsession with the idea of um u death and uh different parts of society uh like sort of obsessed with recreating um that and uh um I I sort of used UFC as like a modern example of like interest in uh some sort of like blood sport. Um, and uh, yeah, I guess I’ll sort of pause here for a second since we only have five five more minutes. So, uh, curious if there are any questions, thoughts, etc.
23:18 Yeah, a couple questions in the chat. Um, feel free to just speak up and ask your question live. Um, makes it easier for people just listening in to follow.
23:31 Um, so is it right to think that carnolisk interfaces and their traces are essentially stickic whereas archival interfaces and and their traces are not uh I I don’t I don’t think that’s true. I think uh both um I I think I see the difference as like carnolisk interfaces require sort of like repetition of the event to uh retain the trace while uh in the case of archival interfaces um they are externalized and they are available for another actor without sort of um like repetition of the event uh itself.
24:23 uh curious about who decides which traces are impactful enough to be captured in archival time. So the football club whose experience with festival is archived to set the tone direction for the next carnival or can archival time be captured from um different perspectives all in uh one place. Um yeah, I guess this is also sort of a question I’m working through of um I sort of wrote this article about football stadiums uh with this idea in mind of like how you could sort of design a place for archival density versus uh carnival permeability. Um and uh and so like uh I feel like like even when you’re designing um something to be captured in archival time, you need to consider both uh both of these like both archival traces and carnival uh traces. Um um and u yeah.
25:34 Yeah. And I kind of feel like um even in um you know like the the humbult research presentation we had yesterday you could think of that as pure archival time but actually there is like some sort of like carnival permeability of like when we are not uh when Humboldt is not presenting it’s researching and it’s exhibiting some level of agency and there’s like some like unexpectedness going on there. So, um so yeah, I think you you need to sort of like consider um uh both in the case of yeah, like designing something for uh archival time or not.
26:28 Um I was curious. I mean, a lot of the examples you’re giving are kind of like grander scales, and I was curious if you could scope it into to like your personal life or an example of kind of like your personal experience.
26:42 Um, that’s a good question. I uh uh I I do feel like different online spaces have like a different energy to them. In the ca like if for example in the case of protocol institute we have a different type of cadence where the discord is not that active uh and a lot of our discussions are based on somebody shares something reads it gets back and there’s a more sort of um archival interface vibe to it while I’m also in discords where uh people have to turn on like the slow mode because it’s just like like chatter and sort people uh like the discord sort of survives on [clears throat] people being attuned to one another constantly versus um in the case of protocol institute I don’t think it’s as uh constant um so maybe that’s that’s like a example that was top of mind for me
27:42 very cool thank you got one more Uh yeah, Burning Man is like a good example of carnival time as well. Uh I don’t know how it’s like now and I feel like it’s very uh institutionalized in some way.
28:08 So I don’t know. um punctuated dise equilibrium from geology geological time uh I don’t know much but I should uh look into that do superstitions have any meaning in this context um I haven’t really thought about that but I do think like superstition like uh one of the things I had on here was uh this idea of pre-commodification as essentially um so like um in the case of um yeah like some like ideas usually like percolate through people sort of playing around with them and like modifying them. uh but like in the idea the idea of pre-com modification is that because everybody is chatting with the same LLM interface they sort of arrive at roughly the same language to talk about the same thing and um that feels like the like opposite of superstition uh but yeah that’s what came to mind with that question anyway I’ll stop now since we’re at time uh if you post any questions I’ll answer them and Yeah, thank you. Thank you for your time.
29:33 Cool. Thanks, Sagin. That was great. Um, thanks for kicking us off today. And if you’re not subscribed to Summer Lightning on Substack, everybody should do that ASAP. And, uh, next up we’ve got Amber.
This is a polemical visitor's guide to a mysterious geographical site called the Disposition, written by an anonymous author who has observed it for forty years and disputes official narratives about its nature and nomenclature. The text rails against deliberate campaigns of misinformation and mischaracterization by institutional authorities, establishing the author's preferred terminology and framing as the correct interpretation of this contested place.
In this fireside chat, Nadia Asparouhova discusses the concept of anti-memes—ideas that resist spreading or being remembered—drawing on SCP fiction and her book Anti-Mimetic. She extends an epidemiological analogy to online communities, critiquing Venkatesh Rao's 'cozy web' concept by comparing group chats to poorly-ventilated COVID pods that compress and mutate ideas rather than truly insulating people from them, and connects her SOP23 essay's 'Whitehead protocol' (good, attention-freeing automation) versus 'Kafka protocol' (bad, confusion-inducing bureaucracy) to how attention and memory shape which ideas survive.
This issue examines California's wildfire management through the lens of prescribed and wild fires at the wildland-urban interface, tracing how 20th-century fire suppression policies fundamentally contradicted ecological understanding of fire's necessary role in ecosystems. The case study explores combustion as a memory practice and investigates how new tools reshape our epistemological relationship with fire management.