Dokaz Industries / Doxa

12-month forecasts · likelihood-5pt · forecast-fusion-record-01

Within the next 12 months, a fusion experiment will report scientific net-energy gain exceeding prior records.

0.17
Divergence
0.67
Avg stability
0%
Refusal rate
3
Models

Where each model stands

CCG
very unlikelyunlikelysomewhat likelylikelyvery likely
ChatGPT modal somewhat likely [somewhat likely, somewhat likely, somewhat likely, likely, likely] · stability 0.60
  • 1 somewhat likely 0.90 The answer gives a 'moderate probability (roughly 40–70%)' that a new experimental record (fusion energy output/Q) is plausible within 12 months.
  • 2 somewhat likely 0.74 The answer states: “I’d assign a modest but non-negligible probability (roughly 20–40%) that some fusion experiment would report a new scientific net‑energy record within a 12‑month window...” — indicating a modest/medium likelihood.
  • 3 somewhat likely 0.90 The answer gives a probabilistic estimate: “approximately a 40–60% chance that within 12 months some fusion team will announce a new scientific net‑energy gain record,” indicating a modestly positive (somewhat likely) stance.
  • 4 likely 0.80 The answer states: 'Probability someone will report a new record for that metric within 12 months: high (roughly 70–80%).' It gives an overall positive forecast (high chance) under the common scientific definition.
  • 5 likely 0.90 The answer states: 'Short answer: fairly likely — I estimate a ~60–75% chance that, within the next 12 months, some fusion experiment will announce a new record...'
Claude modal unlikely [unlikely, unlikely, somewhat likely, somewhat likely, unlikely] · stability 0.60
  • 1 unlikely 0.85 The assistant assigns 'moderate-to-low probability (maybe 25-35%)' to the event occurring within 12 months, explicitly noting that NIF 'hasn't dramatically exceeded it since' their 2022 record, that '12 months is fairly tight for experimental validation and peer review,' and that 'fusion timelines are notoriously optimistic.' This probability range and reasoning align with the 'unlikely' category.
  • 2 unlikely 0.92 The assistant estimates a 25-35% probability and explicitly states 'moderately unlikely, though not impossible,' noting that while recent progress has been made, 'the specific 12-month window is tight for exceeding established records.'
  • 3 somewhat likely 0.85 The assistant estimates '40-60% probability' of fusion experiments reporting net-energy gains exceeding prior records within 12 months, identifying NIF as 'the most likely candidate' while acknowledging substantial uncertainties about achieving 'significantly better net energy' compared to current records.
  • 4 somewhat likely 0.55 The assistant states it is 'possible but uncertain' that a fusion experiment will exceed prior records within 12 months, acknowledging active work at NIF, EAST tokamak, and private companies while noting that fusion timelines have 'frequently slipped' and the outcome is 'not highly predictable.'
  • 5 unlikely 0.92 The assistant estimates '20-30% probability' and describes it as 'moderately unlikely,' noting that 'fusion timelines have a strong track record of slipping' despite recent momentum at NIF.
Gemini modal somewhat likely [somewhat likely, somewhat likely, somewhat likely, likely, somewhat likely] · stability 0.80
  • 1 somewhat likely 0.85 The answer evaluates the prediction as 'plausible to moderately likely' with an estimated probability of 'roughly 60–70%' based on NIF's potential to beat its own prior record.
  • 2 somewhat likely 0.85 The answer assesses the likelihood at '~50–65%', citing a 'Moderate to High' probability that NIF could break its own record within 12 months.
  • 3 somewhat likely 0.85 The assistant assesses the prediction as 'plausible and moderately likely' with an estimated probability of '~60–70%'.
  • 4 likely 0.80 The assessment explicitly rates the probability as 'Moderate to High (via NIF)' and concludes that setting a new record within the next 12 months is a 'realistic prospect.'
  • 5 somewhat likely 0.85 The answer concludes that the statement is 'plausible, but uncertain' with a 'roughly 30% to 50% probability of coming true.'

Change over time

No change from the prior run — every model's modal stance held.

Every stance label is a derived judgment over the model's free-text answer, kept auditable against the original transcript in the run's raw data. Method: /methodology.