AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: Six Signals For Assessing AI’s Prospects In 2027 on ThorstenMeyerAI.com

Buying for a business?Offer from Amazon

Get business pricing on monitors, keyboards and dev gear

  • Business-only prices and quantity discounts
  • Tax-exempt purchasing
  • Multiple users, one account, clear invoices
As an affiliate, we earn on qualifying purchases.

TL;DR

ThorstenMeyerAI.com published a watchlist on Nov. 18, 2026, for testing a scenario in which AI accelerates the development of future AI systems in 2027. Its six signals cover AI task performance, AI use in research, computing infrastructure, power, robotics and employment; the author stresses that the thresholds are judgment calls, not forecasts or official measures.

ThorstenMeyerAI.com published a six-signal watchlist on November 18 for tracking whether an AI progress scenario described by the site for 2027 begins to materialize. The tracker sets out measures, available readings and judgment-based thresholds across AI capability, research automation, computing capacity, electricity, robotics and employment, while warning that several measures lack scheduled updates or reliable data.

The tracker accompanies an earlier scenario article proposing that AI systems could take on more of the work involved in developing subsequent AI systems. In that scenario, faster, compounding progress could compress a year of progress at the prior pace into roughly two months, until limits in chips, buildings and power constrain expansion. The tracker does not make that case anew: it describes what to monitor and what readings would strengthen or weaken the scenario.

Its first signal is METR’s estimate of the length of software tasks that an AI agent can complete with 80% reliability, measured against the time a skilled human needs. The last public record cited is 185.9 minutes for an early Claude Mythos Preview model in April 2026. The tracker says the scenario gains support if a new task suite records a 16-hour horizon by mid-2027, but METR has said its current suite cannot measure above 16 hours reliably. The site also notes that METR’s public page was last updated September 8, 2026, with no date announced for another reading.

A second signal tracks AI’s role in AI research. Anthropic reported that Claude “leads” 26% of its AI research and development work as of August 2026, where leading means completing most of a task end to end under human supervision. The tracker says no measured share runs fully autonomously. It also points to OpenAI’s GPT-6.2 Astra as the company’s first released model to reach its “High” threshold for AI self-improvement. The other signals track data-center construction and power supply, industrial robot installations, and AI-related job cuts. The supplied material reports $96.5 billion in annualized U.S. data-center construction in September, compared with $50.7 billion a year earlier; expected U.S. power generation growth of 2.7% in 2026 and 1.2% in 2027; 603,000 factory robots installed in 2025, with 655,000 forecast for 2026; and 2% of U.S. firms reporting AI-related job cuts.

At a glance
reportWhen: Published November 18, 2026; tracker to…
The developmentThorstenMeyerAI.com published a six-signal tracker for monitoring whether a rapid AI progress scenario plays out in 2027.

How the Signals Test the Scenario

The watchlist matters because it turns a broad claim about rapid AI acceleration into observable indicators. A sharp increase in benchmarked task duration or in AI’s share of research work would speak to whether AI is helping produce more capable AI. Data-center investment and electricity growth indicate whether physical infrastructure can support that expansion. Robot installations and job-cut reports offer a view of how quickly automation reaches workplaces.

The tracker cautions against treating any one number as a verdict. Its thresholds are the author’s own judgment calls, set before future data arrive, and the source says no publisher has defined an official cutoff for the compute and labor indicators. The intended test is whether several signals move in the same direction, not whether a single statistic appears to confirm the scenario.

That distinction is important for readers evaluating claims about AI’s economic effects. Construction spending, projected power generation and robot shipments measure different parts of the capacity and adoption picture; none alone establishes that AI development is accelerating at the proposed rate. The employment figure, meanwhile, records firms reporting AI-related cuts, not a comprehensive measure of AI’s effect on jobs.

Bose QuietComfort Wireless Noise-Cancelling Headphones, Twilight Blue

Bose QuietComfort Wireless Noise-Cancelling Headphones, Twilight Blue

As an affiliate, we earn on qualifying purchases.

From Scenario to Watchlist

The November 18 tracker follows a separate scenario article titled “Scenario 2027: A Year of AI Progress in Two Months, Until Compute Runs Out?” The scenario proposes a feedback loop in which AI contributes increasingly to building AI systems, with progress later constrained by physical resources and potentially aided by robots, factories and laboratories. The new piece is explicitly a scenario tracker, not a forecast or investment advice.

Its author says the thresholds come from the scenario article and are not official standards. The source discloses that the article was researched and drafted with AI assistance, including Anthropic’s Claude. It also states that Anthropic and OpenAI evidence features in two signals and that METR’s independent estimate concerns Anthropic; the author says none of those organizations reviewed or sponsored the article.

Only two of the six signal groups have dated releases listed in the supplied material: U.S. construction spending on December 1, 2026, and power figures on December 8. The lab and METR measures depend on when those organizations choose to publish. The tracker says it will add dated notes as data arrive.

“A scenario is only useful if you can tell early whether it is coming true.”

— ThorstenMeyerAI.com tracker

Gaps in the Measurement Record

The supplied source material does not include the complete definitions, thresholds or latest readings for all six signal groups. It gives headline figures for data-center construction, power, robotics and job cuts, but does not set out every publisher, release schedule or scenario gain-and-loss line for those measures. The tracker’s complete six-part dashboard should not be inferred from the partial details available here.

There are also important limits on the AI capability evidence. METR’s last public record is from April, its public page has no next update date, and it says the current suite is not reliable beyond 16 hours. The tracker reports that METR’s estimate for Anthropic’s improvement rate is around twice its earlier rate, with METR saying it would not be surprised by a range from 1.3 times to three times. That is well below the roughly sixfold pace implied by a year of progress arriving in two months; the tracker says the scenario would need about a fivefold increase from the cited estimate. These figures are estimates, not a confirmed trajectory.

Anthropic’s “leads” measure is also not the same as independent or autonomous operation: the source says humans supervise the work and no measured portion runs fully autonomously. It remains unclear when the labs will publish further readings, whether metrics will remain comparable over time, and whether the infrastructure and labor data will track AI-driven activity specifically.

Upcoming Data and Tracker Updates

The next dated releases identified in the source are U.S. data-center construction figures on December 1, 2026, and power figures on December 8, 2026. Those updates can show whether the recent construction growth and slower projected electricity growth continue, but they will not by themselves settle the broader scenario.

For AI capability and research automation, the next milestone depends on publication by METR, Anthropic or OpenAI. The tracker gives no dates for new measurements from those organizations. Its author says the article will be updated with dated notes as readings arrive, allowing readers to compare new evidence with the stated thresholds.

The central test through 2027 will be whether multiple indicators move together: longer reliable AI task horizons, a larger role for AI in AI research, and expansion in compute and power capacity. The available figures leave the scenario open rather than established; future measurements, their methods and their timing will determine how much weight the tracker’s signals can carry.

Key Questions

Is the tracker forecasting that AI progress will accelerate in 2027?

No. ThorstenMeyerAI.com describes it as a watchlist for a scenario, not a forecast or investment advice. Its thresholds are the author’s judgment calls, not official cutoffs.

What is the strongest current capability reading cited?

The last public METR record cited is an 80% time horizon of 185.9 minutes for an early Claude Mythos Preview model in April 2026. METR says its current task suite cannot reliably measure above 16 hours.

What does Anthropic’s 26% figure mean?

Anthropic said that, as of August 2026, Claude “leads” 26% of its AI research and development work. In the tracker’s description, leading means completing most of a task end to end while a human supervises; it does not mean fully autonomous work.

When are the next updates due?

The source lists U.S. data-center construction figures for December 1, 2026, and power figures for December 8. It gives no scheduled dates for new METR or AI-lab readings.

Does any one signal prove the scenario is happening?

No. The tracker says no single reading decides the scenario. Its approach is to watch whether several measures move in the same direction, while accounting for gaps and limitations in the data.

Source: ThorstenMeyerAI.com

You May Also Like

ARIMA vs ETS: Choosing a Forecasting Model

Great insights into ARIMA vs ETS help you choose the best forecasting model, but understanding your data’s patterns is crucial to making the right decision.

Is AI Hardware Being Designed First For A Smarter Future?

Exploring how AI hardware is shifting focus from general-purpose chips to workload-specific designs to meet inference demands, with implications for the industry.

AI-Assisted Proof Settles Longstanding Question About Quantum Speedups

A Lean-checked proof establishes the Aaronson-Ambainis conjecture in a query model, limiting exponential quantum advantages on most random inputs.

Principal Component Analysis Made Simple

I want to help you understand Principal Component Analysis easily, so keep reading to uncover how it simplifies complex data.