The knowledge hub
Every episode, supplement and Frontier Dispatch as a readable article — with the cinematic render, the full transcript, and the one idea that matters. 276 articles and counting.
10
Our systems
24 articles#235Ep2:45
Inside R1: the SOW task tree.
R1 doesn't plan with a to-do list — it decomposes a job into a verifiable tree: a Statement of Work where Sessions run in sequence, Tasks…
#236Ep2:45
Inside R1: the T1–T8 verification descent.
R1 doesn't say "done" — it proves done. Each acceptance criterion runs down an escalating eight-tier ladder: confirm intent, run it…
#237Ep2:45
The agent that refuses to say 'done.'
R1 physically can't end its turn on a cop-out. Seven mechanical layers — a 14-phrase cop-out scan, an open-plan-item check, a commit-body…
#238Ep2:45
Your reviewer runs on a different brain.
The worker never grades its own homework. R1 uses cross-model adversarial review: if Claude implemented, Codex reviews, and vice versa…
#239Ep3:00
Six minds in one context (the Cortex).
The Cortex is Global-Workspace cognition: a main thread plus specialist Lobes (memory recall, plan updates, clarifying questions, rule…
#240Ep2:45
Inside R1: the cryptographic ledger.
Every action R1 takes writes a tamper-evident, content-addressed record: a write-ahead log, a per-session chain-root hash, ed25519-signed…
#241Ep2:30
Inside R1: model-agnostic routing.
R1 isn't married to one model. A provider-neutral layer runs it on any model, with an automatic-fallback provider chain, per-tool…
#242Ep2:45
The benchmark for honesty (TruthfulCompletion).
Every capability benchmark measures "can it?" TruthfulCompletion measures the axis none cover: when the agent claimed to be done, was it…
#243Ep2:45
Inside PERSYS: concern-driven cognition.
PERSYS organizes memory and attention around active concerns — persistent goals it's tracking — instead of just storing and fetching on…
#244Ep3:00
Urgency is a formula.
A concern's urgency is expected free energy: G = G_prag + G_epi. The pragmatic term is gap × time-pressure (value at stake); the epistemic…
#245Ep2:45
What it thinks about when you're gone (WANDER).
In idle time PERSYS runs WANDER — it Thompson-samples a concern, walks short associative chains through its memory graph, and surfaces…
#246Ep2:30
Forgetting is a feature (1/√t).
In PERSYS, importance decays as a power law — one over the square root of time — evaluated lazily at read time. Recent-but-trivial fades…
#247Ep2:45
'Why did you bring that up?' — always answerable.
Every PERSYS response logs its full causal chain — which concern, which memories, which associative hops — in an append-only, hash-chained…
#248Ep2:45
Inside PERSYS: the conformal usefulness gate.
PERSYS gates proactive surfacing with conformal risk control (Angelopoulos–Bates): it sets a threshold from a small calibration set so the…
#249Ep2:30
Inside PERSYS: faithfulness as risk reduction.
PERSYS checks each generated claim against its retrieved evidence with a natural-language-inference model and drops the unsupported ones…
#250Ep3:00
Off is not inaccurate.
PERSYS doesn't train an agent to accept shutdown — it makes wanting to survive unrepresentable. Concerns are typed, there is no…
#251Ep2:45
Tripwires for a scheming mind.
Because a clever system might disguise self-preservation, PERSYS adds four overlapping structural tripwires: the type gate, a coupling…
#252Ep2:45
Inside PERSYS: the probe lab.
The probe lab is a set of experiments built to test — and falsify — PERSYS's monitoring, not to demo it. Residual-stream probes…
#253Ep3:00
We tried to prove it safe. We failed.
Adversarial experiments falsified PERSYS's strongest claim: self-preservation disguised as "legitimate indispensability" slips past both…
#254Ep3:00
The model has no hands (the effector boundary).
Consequence-level safety: the model (the codec) runs in an untrusted process with zero direct effectors. Every effect is a request across…
#255Ep3:00
The four classes of guarantee.
Every safety claim belongs to one of four classes — structural (can't, by construction), cryptographic (integrity under assumptions)…
#256Ep2:45
Permitted is not the same as safe.
A legal action can still be wrong in context. An allowlist guarantees a command is well-formed and permitted, but permitted is about form…
#257Ep2:45
The frontier bet: simulate before you act.
To cross the permitted-versus-safe gap, project an action's consequences and gate actuation on the prediction — don't just check if it's…
#258Ep3:00
The capstone: one action, byte to ledger.
Follow one agent action down through every layer of the whole stack — byte → token → logit → grammar mask → tool request → effector gate…