Your Reservoir Model Just Changed the Subject. CLARISSA Kept Up.
A reported CLARISSA demonstration turns a black-oil model into a compositional CO2 case. What changed, what was checked, and what engineers still decide.
Print dispatch →Printout № 04 — Dispatch
Unit BW‑26 · 300 baud · ok
Job card — Blauweiss AI
Dispatch archive · verified
AI Reservoir Simulation: Technical Dispatches
Technical essays and research notes from the team building CLARISSA, PetroScript and RIGOR.
Start with the engineering
New to agentic reservoir simulation? Read how AI builds a reservoir simulation model, then explore the CLARISSA technology stack and OPM Flow workflows. For evaluation criteria, see how RIGOR scores an agent.
Chevron, SLB and TotalEnergies Are Advancing Digital Subsurface. But How Do We Validate Autonomous Reservoir Simulation?
As Chevron joins SLB and TotalEnergies in digital subsurface development, Blauweiss examines five tests for trustworthy autonomous reservoir simulation.
Print dispatch →All Wells Shut. Reservoir Still Busy? We Need to Talk.
Before a forecast, CLARISSA checks whether an equilibrium model stays at rest with all wells shut. Why this quiet test matters for reservoir simulation.
Print dispatch →Your Reservoir Model Needs a Memory. Sticky Notes Don’t Count.
CLARISSA records engineering intent and the origins of model assumptions. See why that matters for review, revisions, and the Road to ATCE.
Print dispatch →Your Reservoir Model Came With Baggage. CLARISSA Unpacked It.
CLARISSA reproduced an inherited reservoir model before accepting edits. A Road to ATCE story about dependencies, verification, and engineering confidence.
Print dispatch →Frontier Models Are Getting Better. Reservoir AI Still Needs Engineering Verification.
Frontier AI models can increasingly run reservoir simulation workflows. CLARISSA and RIGOR focus on the harder problem: verifying that the engineering is correct.
Print dispatch →The New Benchmark: CLARISSA + Opus 5.5
Why CLARISSA must prove its value through reservoir-specific verification, reproducibility, and a direct comparison with raw Opus + OPM Flow.
Print dispatch →Can AI Actually Build a Reservoir Simulation Model?
A practical look at AI reservoir simulation, deterministic verification, OPM Flow execution and engineering read-back.
Print dispatch →Reservoir Simulation Is the Proving Ground. Scientific AI Is the Opportunity.
Why reservoir simulation is a powerful proving ground for verifiable scientific AI.
Print dispatch →OpenAI Solved Navier-Stokes. Prandtl Solved It in 1904 — and His Version Made Money.
Why engineering verifiers and useful approximations matter more than headline benchmark wins.
Print dispatch →The Models Didn't Go Rogue. Somebody Turned the Safeties Off.
On containment, reward functions and the engineering control-room problem in AI systems.
Print dispatch →Measuring the Machines: How RIGOR Scores a Reservoir-Engineering Agent
How RIGOR evaluates reservoir-engineering agents with OPM Flow tasks and verifier-grade scoring.
Print dispatch →Your AI Never Signed the NDA
On memory, custody, compliance and local deployment when AI enters the asset team.
Print dispatch →· · · End of Printout · · ·