🔍 Read the full analysis: 50 Signs Of Superintelligence: A Question Of AI’s Readiness For AGI on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
A review of Dr. Alan D. Thompson’s 50-indicator tracker says two signs are achieved, 33 partially achieved and 15 not achieved. An October 26, 2026, review of 18 prominent claims classified seven as verified, five as vendor claims, three as based on a single source and three as disputed or contradicted.
A review published October 26 examined 18 claims in researcher Alan D. Thompson’s 50-sign artificial superintelligence tracker, finding that seven were verified, while others relied on company statements, single sources or disputed evidence. Thompson’s checklist records two indicators as achieved, 33 as partially achieved and 15 as not achieved, but its broad definition of “partially” does not mean systems are close to meeting those goals.
The review, published by ThorstenMeyerAI.com, traced selected claims from Thompson’s 2026 video and public tracker to papers, researchers’ accounts, laboratory posts and regulatory documents. It classified five claims as vendor claims, three as supported by a single source, and three as disputed or contradicted, in addition to the seven it called verified. The supplied review does not identify the remaining category for all 18 claims in its summary.
Some of the stronger evidence concerned mathematical work. The review describes contributions discussed by mathematicians Terence Tao, Don Knuth and Tim Gowers, while retaining caveats about human guidance, the problem’s formulation and the kind of task involved. In one example, Tao characterized GPT-5.2’s work on an Erdős problem as “more or less autonomously,” but the account also says the system received feedback and that the problem had been misformulated.
The review draws a different line around claims of AI improving its own development. It cites Anthropic’s report that its best internal model scored 62.8% on a test, compared with an estimated 85% threshold for fully replacing its researchers. Anthropic also said its research and development was faster with AI, “but not yet by a factor of 2.” These figures describe the company’s own assessment, not an independently established measure of industry-wide progress.
Whose superintelligence?
One word, very different bars
Dr Alan D. Thompson's ASI starts where Google DeepMind's scale puts "Expert AGI". The same model can be "early ASI" on one scale and a strong but narrow AGI candidate on another.
- Median human50th percentileThompson: AGIa machine at the level of an average (median) humanDeepMind: Competent AGIat least the 50th percentile of skilled adults
- Expert90th percentileThompson: ASIexpert-human level in practically any fieldDeepMind: Expert AGIat least the 90th percentile of skilled adults
- Virtuoso99th percentile—DeepMind: Virtuoso AGIat least the 99th percentile of skilled adults
- Beyond every human100%—DeepMind: Superhuman (Level 5)outperforms 100% of humans
- Beyond organisations2026 report—DeepMind: ASImore capable than large organisations of humans
Thompson’s tracker status as of 26 October 2026. DeepMind rows: “Levels of AGI” (2023) and, for the last row, “From AGI to ASI” (June 2026). OpenAI’s Charter: AGI = “highly autonomous systems that outperform humans at most economically valuable work”.
Sources: LifeArchitect.ai ASI checklist (26 Oct 2026), AGI countdown, Morris et al., Levels of AGI (2023), Genewein et al., From AGI to ASI (2026), OpenAI Charter.
Powered by Thorsten Meyer AI
Why the Checklist’s Thresholds Matter
The tracker brings together indicators ranging from mathematical discoveries and automated research to robots performing household tasks and AI systems running valuable companies. That breadth can make it useful as a watchlist, but the review argues that its status labels need careful reading: “partially achieved” can include funded programmes, research infrastructure, scientific missions and other enabling work, rather than a system that has demonstrated the stated outcome.
Definitions also change what a declaration of “superintelligence” means. Thompson defines artificial superintelligence as performance at the level of an expert human in practically any field. Google DeepMind’s 2023 framework calls expert-level performance “Expert” AGI; its 2026 report describes superintelligence as exceeding the cognitive capabilities of large human organisations. The same system may fit one researcher’s label without meeting another organisation’s threshold.
For readers, the distinction matters because claims about imminent AGI can influence business decisions, public policy and expectations about employment. The reviewed evidence points to impressive results in particular, checkable tasks, but it does not establish that today’s systems can reliably perform across fields, act autonomously at scale or replace human research teams.
Popular right now
Breville Bambino Espresso Machine, Brushed Stainless Steel
De'Longhi Dinamica Plus Fully Automatic Espresso Machine
KitchenAid Artisan 5-Qt Tilt-Head Stand Mixer, Cardamom
As an affiliate, we earn on qualifying purchases.
How the 50 Indicators Are Defined
Thompson, an Australian AI consultant who publishes The Memo newsletter, maintains a public checklist of 50 indicators, from recursive code improvement to much more speculative milestones. His October video, “ASI: A better world – 2026,” presents developments from January through October. The tracker’s October 26 status grid lists two achieved indicators, 33 partial and 15 not achieved.
In an August correction note, Thompson said partial status includes “documented material progress toward an indicator,” including funded programmes and enabling work. The review gives as an example two exoplanet candidates reportedly found with Claude Code that were placed under an indicator about a planet beyond Earth being colonized. That classification reflects the tracker’s broad method; it is not evidence that colonization has occurred.
Other organisations use different benchmarks. Google DeepMind’s 2023 scale ranges from competent performance to expert, virtuoso and superhuman levels. OpenAI’s charter defines AGI around highly autonomous systems that outperform humans at most economically valuable work. Autonomy, breadth and the comparison group are therefore central to deciding whether a particular achievement supports an AGI or ASI claim.
““documented material progress toward an indicator, including funded programs, research infrastructure, formal scientific missions, and enabling work””
— Alan D. Thompson, in an August correction note about the tracker
What the Evidence Cannot Establish
The review checks a selected set of 18 claims, not all 50 indicators, and its conclusions depend on the sources available for each example. Some results are independently supported, while others remain company-reported, single-sourced or disputed. The supplied material does not provide enough detail to assess every claim in the tracker or reproduce each assessment independently.
There is also no shared definition of ASI in the material. Thompson’s expert-level standard is less demanding than DeepMind’s description of systems surpassing large human organisations, while OpenAI places emphasis on autonomy and economically valuable work. The label “early ASI” for named models is therefore Thompson’s characterization, not a settled finding shared across those frameworks.
The review says household robots are not operating at scale, it found no verifiable AI-run company worth a billion dollars, and it saw no visible break in the U.S. unemployment rate. Those observations do not by themselves settle what future systems can do, nor establish that AI had no effects on jobs or businesses. The supplied source gives no measurement method or time series for those claims.
Evidence Needed for the Remaining Signs
The next test for the tracker is whether partial indicators turn into demonstrated outcomes that independent researchers can inspect. For mathematical claims, that means clear problem statements, proofs or formal certificates, and transparent accounts of human input. For claims about AI-assisted research, it means reproducible measures that show how much work systems complete and how much time or effort people still contribute.
Robotics provides another concrete milestone. Thompson’s separate conservative AGI countdown is reported at 99%, with its final percentage tied to a robot completing multi-step fine assembly, such as building flat-pack furniture, at average-human reliability. That is a proposed benchmark within his framework, not an independent consensus measure. The review does not give a date for when that performance might be reached.
For now, the tracker remains one researcher’s framework for monitoring developments. Further updates will depend on new results, their primary documentation and whether independent experts can confirm that they meet the stated indicators. No timetable for achieving the remaining signs is established in the material reviewed.
Key Questions
How many of the 50 indicators does the tracker mark as achieved?
Thompson’s October 26 tracker lists two achieved, 33 partially achieved and 15 not achieved. Its partial category includes enabling work and funded programmes, not just near-completion of an indicator.
Does the review say current AI systems are superintelligent?
No. It reports that Thompson calls certain models “early ASI,” but explains that this depends on his definition. Other frameworks set different thresholds, and the label is not presented as a shared scientific conclusion.
What did the review find about mathematical achievements?
It found several claims involving mathematical work supported by accounts from mathematicians or external researchers. Those accounts include qualifications about feedback, human coaching, problem formulation and who completed the proof.
What does Anthropic say about AI improving its research work?
The review cites Anthropic’s report that its best internal model scored 62.8% on a test, below the company’s estimated 85% level for fully replacing its researchers. Anthropic said its R&D was faster with AI, “but not yet by a factor of 2.”
Source: ThorstenMeyerAI.com
