SuperintelligenceMONITOR
Published by Center for Artificial WisdomLatest newsletter: Issue #1 · draft v1.11

Issue #1 · 2026-10-08 · draft v1.3

AI can help produce research. Can it judge what is worth pursuing?

A catalogue of mathematical results, a failed research task, and a corrected game test reveal the distance between producing an answer and establishing a discovery.

Revised 2026-10-17 · Rubric SI-1.0 · Unsent review draft.

Updated in v1.3 · October 17, 2026 · Opening revision and earlier wording

OpenAI published hundreds of mathematical manuscripts. Two frontier agents failed to match a human research innovation. A model reached a game benchmark’s ceiling, then lost ground when evaluators removed a legal shortcut. Read together, these reports expose a difficult question for AI research: can a system choose a worthwhile problem, produce a useful result, and establish that the result is sound?

We all hope for a happier future with less suffering. Our hopes for a future with superintelligence belong to that same human concern. Reaching and sustaining such a future would mean that we had harnessed wise superintelligence; failing to reach it would mean that we had missed the wisdom needed to get there.

Too much remains unknown about the systems being developed and where they are taking us, in the near term and over generations. The Center for Artificial Wisdom works to reduce that uncertainty through conceptual and empirical inquiry into artificial wisdom and the path to it. Wisdom belongs throughout that inquiry, from the choices shaping development to the conditions that let a desirable future endure.

Superintelligence Monitor, published by CAW, examines the unfolding trajectory. We investigate the causes and conditions behind advances, test expectations against consequences, and preserve the record when our judgments change. CAW’s framework faces those same tests. If it fails to help us understand the evidence, or another framework proves more useful, we will adapt.

Does the development and deployment of AI reinforce or loosen reification, and what does the evidence show about wisdom and an enduring future with less suffering?

This first issue examines five primary reports published October 6–8, 2026. They concern research production, autonomous research, the validity of tests, and the reliability of oversight. The selection is not comprehensive. Sources were checked October 8; older work is identified as historical context. These assessments remain provisional and await human editorial approval. Draft v1.3, revised October 17, states the publication’s underlying purpose more fully; it adds no new evidence and preserves the original information cutoff.

1 / 5 · Lab claim · 2026-10-06

Hundreds of manuscripts leave the central question open: which results hold up?

Clarified in v1.2 · October 17, 2026 · Assessment revision and earlier wording

The release could accelerate mathematical research if its results prove novel, correct, and useful. Its sustained impact remains undetermined.

OpenAI’s October 6 release presents mathematical results produced with an internal model that outsiders cannot yet use. The catalogue groups 719 manuscripts into 372 families; those totals do not count independently verified, distinct open problems. The need for scrutiny is already visible in its October 7 history: a sign error led to three withdrawals and affected dependent results, while 14 other manuscripts received repairs.

A reliable supply of new, reusable results could speed mathematical research. To establish that contribution, researchers must check what each result adds and whether it holds up. The number of manuscripts alone cannot establish a general autonomous scientist or superintelligence.

The work draws on existing mathematics, human-selected problems, verification tools, and computation performed while the model works on an answer. The report does not isolate which recent change made the difference. A specialized production pipeline and a reasoner that improves across fields could generate similar headlines while implying different futures.

Main limitation. The model is unavailable for independent reproduction, and the released work was selected from roughly 4,000 posed problems. That selection prevents a clear general success-rate estimate. A machine-checked formal proof establishes a proposition under stated assumptions; it does not establish the proposition’s novelty or importance, or validate every surrounding informal explanation.

Watch next. Look for expert reviews at the family level, usable proof artifacts, further withdrawals, and results that other researchers build on. A system outsiders can reproduce on problems chosen in advance would provide a stronger test. Reliable new results would strengthen the acceleration claim; persistent errors or failure to transfer would weaken it.

Current assessment, ratings, and review history

2 / 5 · Independent evaluation · Epoch AI · 2026-10-07

Given tools and compute, two agents still fell short of a research innovation

Clarified in v1.2 · October 17, 2026 · Assessment revision and earlier wording

One demanding test exposes a gap in autonomous research. It does not establish a ceiling across research domains.

Epoch AI asked Claude Fable 5 and GPT-5.6 Sol to develop a post-training method comparable to a recent human innovation. Neither matched it. Each run could use up to 3,000 GPU-hours and a 10-billion-token inference budget; those were limits, not reported expenditure. Newer successors already knew the task, so their results offered a weaker test of fresh research.

An AI system that improves research methods could help accelerate the development of later systems. Epoch tested the connected work of choosing an idea, implementing it, and running experiments. Success at an isolated coding task leaves much of that process unexamined.

The agents had compute and tools. Choosing productive ideas and checking improvements within the task’s scope remained difficult. The surrounding software and instructions may also have constrained them. Fresh problems, different setups, and comparable human resources would help separate a reasoning gap from a limitation of the test.

Main limitation. The study centers on one research problem and one main evaluation per model. Human judgments helped determine what counted as in scope, and the resource budgets may not be directly comparable. The failures therefore constrain claims about these runs without showing that every system would fail across research.

Watch next. Look for repeated runs on several tasks chosen before testing, with credible human baselines and independent checks of novelty and results. Reproducible success across fresh research problems would change this assessment more than a higher coding score.

Current assessment, ratings, and review history

3 / 5 · Independent evaluation · Epoch AI · 2026-10-07

A legal shortcut took a game test to its ceiling. Removing it changed the result.

Clarified in v1.2 · October 17, 2026 · Assessment revision and earlier wording

The model made real progress at the game. Neither its perfect scores nor its revised results establish general scientific discovery.

GPT-6 Astra reached perfect scores in more than half of its original EBR-bench attempts by using a legal card combination that bypassed time pressure. Epoch then banned the card. The model’s mean score fell from 19.8 to 16 of 21 objectives; its best run under the revised rules reached 20.

Finding a legal shortcut is itself a skill. It also reveals that a test may reward a different ability from the one readers assume it measures. Strong performance after the rule change supports progress within this game. Whether that progress carries into open-ended research requires separate evidence.

The card combination made the original ceiling reachable. Removing it supplied a useful comparison: the model remained strong, but its average score fell. That comparison helps explain the game result without establishing why the model improved or whether its approach works elsewhere.

Main limitation. A top human used the same legal tactic. The episode therefore does not, by itself, demonstrate misconduct or confusion by the model. Results under the old and revised rules need separate comparisons. Our limited-impact judgment concerns this game result’s significance for the specified SI claim, not the model’s overall capability.

Watch next. Look for repeated performance under the revised rules and tests in other environments. We have not measured public attention to the result, so the evidence does not support a resolved false-alarm judgment.

Current assessment, ratings, and review history

4 / 5 · Lab-affiliated researcher report · 2026-10-08

An AI-assisted sky map still needed a scientist to catch an error

Clarified in v1.2 · October 17, 2026 · Assessment revision and earlier wording

The project shows a useful route to assisted scientific production. Autonomous discovery and downstream impact remain unproven.

Astrophysicist and Anthropic researcher Brice Ménard describes using Claude Science to combine ultraviolet surveys into an educational sky map. About a third of the map is statistically predicted rather than newly observed, and the map distinguishes those regions. Ménard directed the work and caught an artifact that two agent reviews missed.

Scientific progress can come from making neglected work practical. If AI reduces the labor of assembling research resources, more projects may become feasible. This case points to that possibility while leaving the size and reliability of the saving unmeasured.

Public telescope archives supplied the observations. Model-assisted coding and coordination helped assemble them, while expert judgment helped check the output. A comparison with the same work performed through an alternative workflow would help establish whether the main contribution was reduced labor or an improvement in scientific reasoning.

Main limitation. There is no independent replication or matched comparison of labor. Predicted regions are not new measurements, and the case does not establish validated new astronomical knowledge. The developer affiliation also matters when weighing the account.

Watch next. Look for independent checks of the map’s predictions and uncertainty, reuse by researchers, and replicated savings in time or cost across projects. Those results would support a specific research-productivity claim.

Current assessment, ratings, and review history

5 / 5 · Independent evaluator security test · METR · 2026-10-06

An oversight tool could alter the evidence reviewers saw

Clarified in v1.2 · October 17, 2026 · Assessment revision and earlier wording

METR demonstrated a weakness in an evaluation tool. It did not observe evaluated agents using it to conceal misconduct.

METR reported a proof of concept in which a vulnerability in its transcript viewer could change what reviewers saw. The underlying database logs remained intact. METR patched the flaw and said it had not observed evaluated agents exploiting it.

An evaluator’s conclusion depends on a trustworthy account of what a system did. A weakness in the tools that display that account can undermine both capability assessment and oversight, even when the underlying records survive.

The viewer’s handling of untrusted material created the opening. Keeping authoritative logs separate from their display and securing the review tools address that mechanism. Broader tests of agent behavior are still needed.

Main limitation. A staged exploit establishes a possible failure mode. It does not establish spontaneous concealment by a deployed agent, a general tendency to deceive, or imminent superintelligence.

Watch next. Look for independent security reviews, preserved audit trails, and similar weaknesses in other evaluation tools. Evidence of an agent initiating such an exploit would support a stronger, separate claim.

Current assessment, ratings, and review history

Our synthesis · editorial interpretation

What would establish a discovery?

Updated in v1.3 · October 17, 2026 · Synthesis revision and earlier wording

These reports show why a future with more capable AI depends on more than producing answers. A mathematical manuscript needs scrutiny before another researcher can rely on it. A sky map needs a scientist who can distinguish a prediction from an observation and catch an error. An evaluation needs a trustworthy account of what the system did. Those conditions help determine whether a technical achievement becomes something people can use and govern.

The evidence here supports gains in research assistance. It leaves open whether systems can reliably choose worthwhile problems, establish novel results, and validate them without substantial human direction. It also leaves open the broader trajectory toward wise superintelligence. We will follow fresh tests, reproducible results, human assistance, and the consequences of the choices made along the way. CAW’s account of wisdom must remain open to revision as that record develops.

SIM connects the future we hope for with the evidence and understanding of the causes and conditions for bringing it to fruition.

Clarified in v1.2 · October 17, 2026 · Historical summary revision and earlier wording

Historical context: AlexNet’s result depended on a combination of resources; A translation architecture became a foundation for later language models. These selected cases illustrate enabling conditions; they do not establish forecasting skill.

What we expect next

Two draft forecasts make our follow-up questions checkable. Their windows set deadlines for review; they do not predict an arrival date for superintelligence.

Draft forecast · October 9–November 30, 2026

More mathematical results will receive formal artifacts

We expect continued formalization, with moderate qualitative confidence. This is a verification milestone, not a forecast of superintelligence.

Review deadline: 2026-11-30 · not yet resolved

Draft forecast · October 9–December 31, 2026

General research autonomy will remain unestablished in the open evidence

We expect public evidence to remain insufficient for the broad claim below, with low qualitative confidence. A negative finding may reflect limited evaluation access rather than a capability limit.

Review deadline: 2026-12-31 · not yet resolved

Publication record

Initial draft prepared 2026-10-08. Draft v1.3 · 2026-10-17. States the underlying logic connecting hopes for a happier future with less suffering, wise superintelligence, CAW’s inquiry, and SIM’s causal and accountable record. This updates the publication’s purpose; evidence, ratings, claim identities, information cutoff, and all forecast fields are unchanged. Continuing assessments have their own histories.

Current draft · Preserved draft v1.2 · Preserved draft v1.1 · Initial draft · Revision log with original wording

Research, source checking, drafting, and site preparation used AI assistance. This is an editorial draft; it has not received human editorial approval. No claim is made that SI Monitor independently reproduced the experiments or verified the mathematical proofs.

SI Monitor by email

Understand what changed—and why it matters.

Follow the evidence shaping our future with superintelligence. Our planned weekly newsletter explains the findings and choices that could shape a happier future with less suffering. The website’s daily briefing reviews sources between issues. The publication is in public preview. Issue #1 is still an unsent review draft. A launch date and regular send day have not been set; material corrections receive a notice when needed.

Brevo handles signup and email delivery. You can unsubscribe using the link in each newsletter. Read our subscription privacy information before submitting.

The publisher reported a successful signup and email confirmation test on October 15. Campaign delivery and unsubscribe checks remain incomplete.

Continue to signup

You’ll open Brevo’s hosted form. Enter your email address there and follow the confirmation instructions. Reading this page does not load the signup service.

If confirmation does not arrive, check your spam folder or contact us for subscription help.

Preview the newsletter first