🔍 Read the full analysis: A Look At 50 Signs That Could Put AI Closer To AGI Or ASI on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
A review of researcher Alan D. Thompson’s 50-item ASI tracker reports two indicators achieved, 33 partially achieved and 15 not achieved as of October 26, 2026. The review says several mathematical and scientific advances are supported by outside evidence, while other claims rely on companies or single sources; the tracker’s broad “partial” category does not mean superintelligence is near.
A review published October 26 of researcher Alan D. Thompson’s 50-sign artificial superintelligence tracker found that some prominent AI achievements are independently supported, while many other entries remain partial, company-reported or disputed. Thompson’s tracker lists two indicators as achieved, 33 as partially achieved and 15 as not achieved; the review cautions that its broad “partial” category is not evidence that superintelligence is imminent.
The review examined 18 claims from Thompson’s 2026 video and checklist, tracing them to papers, lab statements, researchers and official records. It classified seven as verified, five as vendor claims, three as resting on a single source and three as disputed or contradicted. The review says the better-supported results cluster in areas with checkable evidence, including mathematical proofs, a physics calculation and regulatory filings.
Examples include AI work on open mathematical problems. The review cites mathematicians Terence Tao, Don Knuth and Tim Gowers as describing genuine contributions, while also documenting limits: some work involved coaching, human-written proofs, or problems whose formulation and prior literature needed scrutiny. The review says an OpenAI model’s work on a unit-distance conjecture was verified by outside mathematicians; other claims, including results attributed to GPT-6 Astra, were still vendor claims in the material examined.
The tracker’s status categories have a wide scope. Thompson wrote in an August correction note that “partially achieved” can include funded programmes, research infrastructure, formal scientific missions and enabling work. The review points to examples where progress toward a distant goal is counted as partial rather than achievement of the goal itself. It says the tracker should be read as one researcher’s checklist, not a standardized measurement of AI capability.
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 ASI label varies
The review highlights a basic problem in claims that AI has reached AGI or ASI: the labels depend on the definition. Thompson defines ASI as performing at expert-human level in practically any field. Under Google DeepMind’s 2023 framework, that is closer to “Expert AGI,” a level below “Superhuman.” DeepMind’s 2026 report describes superintelligence more expansively, as a system more capable than large organisations of humans.
That gap affects how readers interpret a claim such as “early ASI.” A model that helps produce a strong mathematical result may show meaningful capability in one domain without demonstrating broad, autonomous performance across work and daily life. OpenAI’s charter also places emphasis on autonomy and performance across economically valuable work. Different yardsticks can make the same system sound far more or less advanced, so claims need both evidence and a stated benchmark.
The distinction matters beyond terminology. Governments, companies and the public use capability claims to weigh safety, economic effects and investment. A checklist that treats preliminary infrastructure as partial progress can be useful for tracking research activity, but it cannot by itself show that a system can perform the end goal. The review’s case-by-case approach gives readers a way to separate demonstrated results from forecasts and promotional claims.
Products Worth Considering
Apple Mac mini Desktop Computer with M5 Pro Chip (2026)
HP Sprocket Photo Booth Machine Instant Color Photo Printer with 10.1 Touchscreen, LED Ring Light, 250+ Frames, 3×4 ZINK Paper, QR Code Shar
FLEXISPOT 32-Inch Electric Standing Desk Converter with Keyboard Tray
As an affiliate, we earn on qualifying purchases.
How the tracker defines progress
Thompson, an Australian AI consultant who publishes The Memo newsletter, maintains a public 50-indicator ASI checklist. Its items range from recursive code improvement to much more speculative milestones. His October video, “ASI: A better world – 2026,” presents a montage of developments from January through October. The review reports that Thompson identifies Claude Mythos Preview and GPT-6 Astra as “early ASI” systems; that characterization is his view, not an independently established classification.
Thompson’s definitions also differ from other commonly cited frameworks. He describes AGI as a machine performing at the level of an average human and ASI as one performing at expert level across practically any field. DeepMind’s 2023 levels run from “Competent” through “Expert” and “Virtuoso” to “Superhuman,” while its 2026 report sets a higher bar for superintelligence. OpenAI’s charter defines AGI around highly autonomous systems that outperform humans at most economically valuable work. There is no single agreed threshold in the source material.
The source review also notes that Thompson maintains a separate “conservative countdown to AGI,” which stood at 99%. According to the account, the remaining step is a robot completing multi-step fine assembly, such as putting together furniture, at average-human reliability. This is a distinct measure from the 50-item ASI tracker and should not be treated as confirmation that AGI is 99% complete.
““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 on his tracker
What the evidence cannot settle
The review’s findings cover 18 selected claims, not every entry in the 50-sign tracker, and its verdicts reflect the sources available to its author. The supplied material does not provide the full item-by-item evidence for all 18 claims, so readers cannot independently assess every classification from this account alone. Some results are described as vendor claims because public certificates or company figures have not been independently reproduced.
It also remains unclear whether current systems can carry out broad, reliable work without close human direction. The review says there are no household robots deployed at scale, no verifiably billion-dollar AI-run company, and no visible break in the U.S. unemployment rate attributable to AI. Those observations do not rule out future effects, but they do not support claims that such milestones have already occurred. The material also does not establish a shared scientific test for when AGI or ASI has arrived.
Evidence to watch next
Further progress will be easier to evaluate when researchers and labs publish reproducible results, independent checks and clear benchmark definitions. For mathematical claims, that includes full proofs and enough detail to distinguish autonomous discovery from work shaped by human prompts or verification. For robotics, a relevant test would be sustained multi-step performance in ordinary tasks, rather than a demonstration limited to a controlled setting.
Readers can also watch for updates to Thompson’s tracker and for follow-up assessments from researchers named in the review. The next important development is not simply another label or status change, but evidence showing what a system did, how much human assistance it received and whether others can repeat the result. Until then, the tracker remains a set of indicators and interpretations rather than a definitive measure of how close AI is to superintelligence.
Key Questions
How many indicators does Thompson’s tracker mark as achieved?
As of October 26, 2026, it lists two achieved, 33 partially achieved and 15 not achieved.
Does “partially achieved” mean an indicator is nearly complete?
Not necessarily. Thompson says the category can include funded programmes, research infrastructure and other enabling work, so it does not mean the final capability has been demonstrated.
Did the review find that AI has reached ASI?
No. The review reports a mixture of verified results, vendor claims, single-source claims and disputed claims. Thompson’s description of some models as “early ASI” is his assessment, and definitions of ASI differ.
What AI achievements did the review find best supported?
It says the strongest evidence was in areas that can be checked, including mathematical work, a validated physics calculation and regulatory filings. It also describes caveats around human coaching, proof-writing and prior work.
What major milestones does the source say have not happened?
The review says there are not yet household robots at scale or a verifiably billion-dollar AI-run company, and it reports no visible break in the U.S. unemployment rate. These points do not establish what may happen later.
Source: ThorstenMeyerAI.com
