CM-001DEFINITION NOT LOCKED
Will an AI coding agent resolve 90% of SWE-bench Pro V2 private tasks by October 2027?
Open quality gate: Freeze a permanent private-split identifier, task denominator and official archive before probability issuance.
CORTIST / PUBLIC SCORECARDShow the misses.
Show the misses.
Show the evidence.
A public accounting of locked forecasts, resolved outcomes and scoring. Nothing is called a success until an actual independent outcome is documented.
SCORING / VERIFIED OUTCOMES · The public scorecard contains zero independently resolved or scored forecasts. Editorial drafts and internal working models do not earn scores. See the prospective protocol. Editorial responsibility ↗
PUBLIC STATUS / 27 OCT 2026
Forecast results.
Reference scoring: binary Brier score, lower is better. No outcomes exist yet.
Published numeric drafts8
Unscored. Original manifest verifies 2 founding SHA-256 records.
Editorially locked milestones2
CM-003 20% and CM-004 15%; neither resolved nor independently reviewed.
Resolved and scored0
Requires prior lock and independent evidence
Mean Brier scoreNot available
Never divide by zero or fabricate outcomes
PUBLIC LEDGER / DATA STATUS
What's on record?
| Forecast | Draft value | Deadline | Status | Snapshot |
|---|---|---|---|---|
| F-001 / agi | 38% | 2030-12-31 UTC | DRAFT / NOT SCORED | v0.1 JSON |
| F-002 / Superintelligence | 24% | 2040-12-31 UTC | DRAFT / NOT SCORED | v0.1 JSON |
| F-003 / AI & Robotics | 35% | 2035-12-31 UTC | DRAFT / NOT SCORED | v0.1 JSON |
| F-004 / AI & Cybersecurity | 45% | 2030-12-31 UTC | DRAFT / NOT SCORED | v0.1 JSON |
| F-005 / AI Chips & Compute | 55% | 2030-12-31 UTC | DRAFT / NOT SCORED | v0.1 JSON |
| F-006 / AI & Science | 40% | 2035-12-31 UTC | DRAFT / NOT SCORED | v0.1 JSON |
| F-007 / AI & Economy | 35% | 2035-12-31 UTC | DRAFT / NOT SCORED | v0.1 JSON |
| F-008 / AI & Governance | 60% | 2032-12-31 UTC | DRAFT / NOT SCORED | v0.1 JSON |
The original checksum checker verifies the two founding F-001/F-002 records only. Six later F-003–F-008 drafts are in the 13-topic version history but not in the original SHA-256 manifest. Site-controlled hashes do not independently prove time of publication.
Verifying the two original founding snapshots...
Verifying record checksums in supported browsers...
NEW / FIVE SOURCE-BACKED MILESTONE REVIEWS
Our dated primary-source audit explains benchmark version ambiguity, legal decision criteria, operator timelines and missing power normalization. None of the five candidates is independently locked or ready to score.
Explore the five detailed evidence dossiersFORECAST READINESS / 27 OCTOBER 2026
Five milestones. Five open quality gates.
Before a future probability can be scored, Cortist must freeze the outcome definition, relevant benchmark version, verifiable evidence and independent timestamp. The five original candidate definitions remain historical. CM-003 and CM-004 now have separately registered 20% and 15% original forecasts, prospectively locked by publisher approval at live release; CM-001, CM-002 and CM-005 remain unapproved candidates.
CM-002DEFINITION NOT LOCKED
Will a robot complete at least 50% of five-object LADDER tasks by October 2027?
Open quality gate: Freeze exact five-object subset, trial counts, evaluation protocol and result archival before locking.
CM-003ORIGINAL CANDIDATE; NEW V1 LOCKED
Will the European Commission impose an Article 101 GPAI fine by October 2027?
Open quality gate: Independently check baseline for earlier decisions and decide how to handle adoption versus publication timing.
Official original v1: 20%, publisher-approved and prospectively registered. Read locked rules.
CM-004ORIGINAL CANDIDATE; NEW V1 LOCKED
Will paid, driverless Waymo rides launch in Munich by October 2027?
Open quality gate: Freeze definition of Munich service area and evidence standard for paid publicly available service.
Official original v1: 15%, publisher-approved and prospectively registered. Read locked rules.
CM-005DEFINITION NOT LOCKED
Will verified AI inference performance per watt double by October 2027?
Open quality gate: Critical: freeze 2026 baseline, comparable power normalization, model/tokenizer, quality tier, latency and reviewer-approved rules.
How new evidence enters the review
Cortist News links to its published Topic through Evidence Pulse 2.0. A keyword match can suggest further review, but cannot confirm an outcome, rank evidence, or change a published percentage. Revisions require separately justified, date-stamped forecast versions.
HOW FUTURE SCORES WORK
One rule, no favorable exceptions.
Inclusion
Only predictions prospectively locked with archived rules and date, associated with a published binary adjudication and independent evidence, enter scoring. Initial locked probability is the primary score; later updates can be presented separately.
Exclusions and baselines
Drafts, disputed and unresolved outcomes stay visible but do not receive a manufactured zero or one. We will publish exclusions, per-question Brier scores, a naive 50% reference (0.25 per binary outcome) and sample-size limits.