Start with research note.
Why this work matters
The stakes are methodological. A usable archive begins by discriminating between central, adjacent, and merely incidental material.
Then and now
Why this mattered then
By late February 2026, the archive had reached 139 transcripts. Ad hoc reading was no longer workable [c7]. The note imposed conservative scores for Land focus and AI focus [c8]. It then isolated a top tier: Anthropol sessions on Roko's Basilisk, orthogonality, and the AI box, plus interviews like Accelerationism & Capital [c3][c9][c12]. The result was a curation map for later CCRU and accelerationism work [c10].
Why it matters now
Now it matters because it lets readers see how the scene sounded, moved, or presented itself before later summaries folded it back into Where to Start with the CCRU.
How to read this
Read it as a scoring ledger, not an argument. The opening request fixes the method on 25 February 2026: assign conservative 10-point scores for Land and AI across ~/clawd/transcripts [c0]. The table that follows carries the weight. Scan titles first, then the paired scores, then the brief rationales. Patterns sit in the accumulation: high Land scores, high AI scores, and mixed cases distributed across 139 transcripts [c3]. Use any single row as a local example, then return to the full list.
Argument map
Corpus as research instrument
One hundred and thirty-nine transcripts sit in a folder called `~/clawd/transcripts`. The instruction is conservative: score each one out of ten on Land focus, out of ten on AI focus, and make a high score hard to get C1 . What follows is a working note from that pass, dated 2026-03-11. The interesting move is not the rubric. The interesting move is that the corpus has stopped being an archive and started behaving as a research instrument.
Anthropol sessions as dense core
The note records a sorting operation. Eight transcripts clear the AI-focus threshold at 7 or above C11 . Thirty-six clear the Land-focus threshold C4 . The overlap is small and named: the Anthropol seminar sessions, where Land teaches AI alignment material directly, hit the highest combined scores C4 C3 . Sessions 2 and 3 cover paperclipper, orthogonality, the Turing test, the AI box experiment C5 . Session 4 logs Land on Roko's Basilisk as theological entity C3 . The four Land interviews (Accelerationism & Capital, Ideology/Intelligence, Bronze Age Pervert, Discussing the Near Future) form the second cluster C4 C7 .
Two-axis scoring as philosophical move
The scoring rubric is doing philosophical work, not just clerical work. By forcing a two-axis score (Land, AI) instead of a single relevance number, the pass surfaces a structural fact about the material. Most CCRU-adjacent transcripts score high on one axis and low on the other. Parisi on recursive colonialism scores 1 on Land and 7 on AI C2 . The Numogram lecture scores 8 on Land and 2 on AI C10 . Fisher on the slow cancellation scores 2 and 1 C8 . The accelerationism-AI weld people assume from the outside, Land plus machine intelligence as a single object, only actually appears in a thin band of material: the Anthropol sessions, the Bitcoin and Philosophy seminars C12 , a handful of late interviews C7 .
Land-AI connection as retroactive
That thinness matters. The reception of Land in the 2020s frequently treats AI as the natural extension of his 1990s cybernetic writing. The transcript scores suggest the connection is mostly retroactive, concentrated in pedagogical settings from the last few years where Land works through Bostrom-vintage alignment material with students. The Brassier session on Mad Black Deleuzianism scores 9 on Land and 3 on AI C5 . Noys on neoliberalism's grammar scores 5 and 1 C13 . The philosophical critiques engage Land without engaging the AI horizon. The AI horizon, when it appears in Land's own voice, comes through the Yudkowsky-Bostrom vocabulary, paperclipper, basilisk, orthogonality, not through the Ccru numogrammatic register.
Outer ring of Ccru-adjacent material
The note also functions as a map of what is missing. One hundred and sixteen of one hundred and thirty-nine transcripts score 0 to 3 on AI focus C11 . The Otolith Group, Tony Cokes with Kodwo Eshun, the Asian Futurisms panel all score zero or one on both axes C9 . These are inside the broader Ccru orbit by association, Eshun being a founding member, but they do not carry Land or AI content. The corpus, sorted this way, separates into a small dense core (Land teaching AI), a medium ring (Land on accelerationism without AI, Parisi on AI without Land), and a large outer field of Ccru-adjacent cultural material that the rubric correctly demotes.
Selective curation enabled
What the note authorises is selective curation. The top-tier stuff, in the assistant's phrase C0 , is now identifiable: roughly a dozen transcripts where Land's voice and AI content coincide. Everything else gets sorted into thematic piles, Numogram and hyperstition, Fisher on temporality, Parisi on computation, Urbanomic on publishing C10 C8 C13 . The single relevance score from a previous pass C1 could not have produced this map. The two-axis pass produces it because it forces the question that single-axis scoring hides: relevant to what.
Triage not reading
The stakes are practical and methodological. Practical, because the next stage of the archive depends on knowing which transcripts repay close transcription and which can be summarised or skipped. Methodological, because the note demonstrates a use of language-model scoring that does not pretend to interpret, only to sort. The conservative rubric (high scores hard to get) keeps the model honest about the gap between Land-adjacent material and Land-on-AI material. The output is a triage document, not a reading. The reading happens after, on the dozen transcripts the triage isolates.
Navigation as the new problem
What the note shows about the wider project: the Ccru corpus is large enough now that navigation has become its own problem. Scoring is one answer. The answer it gives is unsentimental. Most of the material is not what the casual reader expects it to be, and the small fraction that is gets named precisely.
How to read the ledger
Read it as a scoring ledger, not an argument. The opening request fixes the method on 25 February 2026: assign conservative 10-point scores for Land and AI across ~/clawd/transcripts C0 . The table that follows carries the weight. Scan titles first, then the paired scores, then the brief rationales. Patterns sit in the accumulation: high Land scores, high AI scores, and mixed cases distributed across 139 transcripts C3 . Use any single row as a local example, then return to the full list.
Filing date and named highlights
Filed on 11 March 2026, the note records a 25 February background task that scored 139 transcripts for Land and AI focus. C4 C7 The highest combined scores go to direct Land interviews and Anthropol sessions. C5 C13 It names "Accelerationism & Capital with Nick Land," "Ideology, Intelligence, and Capital," and "Lemurian Time War and Hyperstition." C5 C6
Style and mode
Essay / text work
2026-03-11-land-ai-transcripts.md works best when read as compressed scene-writing: argument, terminology, and style arrive together rather than in separate academic stages.
Publication context
Filed on 11 March 2026, the note records a 25 February background task that scored 139 transcripts for Land and AI focus. [c4][c7] The highest combined scores go to direct Land interviews and Anthropol sessions. [c5][c13] It names "Accelerationism & Capital with Nick Land," "Ideology, Intelligence, and Capital," and "Lemurian Time War and Hyperstition." [c5][c6]
Key concepts and people
Concepts
Key passage
Best entry extract · research note
Based on my comprehensive reading of all 139 transcripts, here is the complete scoring table:
Why this matters: The note's authority rests on this claim of exhaustive coverage: every transcript read before ranking, so the scoring table works as a census of the corpus rather than a sample.
Representative extracts
Definition · research note
Based on my comprehensive reading of all 139 transcripts, here is the complete scoring table:
Why this matters: The note's authority rests on this claim of exhaustive coverage: every transcript read before ranking, so the scoring table works as a census of the corpus rather than a sample.
Stakes · scoring table
Anthropol The Future of Human Insecurity (Session 4).txt - 10 - 8 - Land on Roko's Basilisk - AI as theological entity
Why this matters: This entry is what the double scoring exists to surface: material near the top of both axes, where Land's late thought and AI speculation become the same subject.
History · scoring table
Ideology, Intelligence, and Capital with Nick Land.txt - 10 - 6 - In-depth Land interview covering accelerationism trajectory, AI alignment, intelligence explosion
Why this matters: A single row shows the two-axis method at work: the interview maxes the Land score while sitting lower on AI, evidence that the note keeps the two research questions apart.
Afterlife · scoring table
Ghost Lemurs of Madagascar.txt - 5 - 1 - Podcast on CCRU, hyperstition concepts
Why this matters: The low marks matter as much as the high ones: a podcast the archive elsewhere prizes for hyperstition registers here as peripheral, proof that relevance is relative to the question asked.
Method · summary statistics
High Land focus (7-10): 36 - the Anthropol seminar sessions score highest, followed by the direct Land interviews and CCRU lectures.
Why this matters: Here the raw table collapses into usable shape: 36 transcripts clear the high-Land threshold, and seminars and interviews emerge as the corpus's natural centre of gravity.