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

🔍 Read the full analysis: Six Factors That Could Shape Whether AI Takes Off In 2027 on ThorstenMeyerAI.com

Before you orderOffer from Amazon

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
Start your free Prime trial Free trial for eligible customers · Cancel anytime
As an affiliate, we earn on qualifying purchases.

TL;DR

ThorstenMeyerAI.com published a scenario tracker on Nov. 18, 2026, outlining six indicators that could show whether AI progress accelerates sharply in 2027 or runs into limits. The tracker reports mixed evidence: AI already leads some research work at one lab, while public task-horizon measurements, infrastructure and labor data do not yet establish the proposed rapid-progress scenario.

ThorstenMeyerAI.com published a watchlist on Nov. 18 setting out six indicators that could help test whether AI development accelerates sharply in 2027 or is constrained by computing capacity, energy and physical infrastructure. The tracker is an opinion-based scenario monitor, not a forecast; its author says the thresholds are judgment calls and that several key readings have no announced publication dates.

The watchlist follows an earlier scenario article proposing that AI systems could take on more of the work involved in building future AI systems. In that scenario, a year of progress at the previous pace could arrive in about two months before limits in chips, buildings and power slow the process. The new article does not reargue that case. It identifies six signals to monitor and reports their status as of Nov. 18.

The first signal is the length of software tasks AI agents can complete reliably. METR’s last public record for an 80% success time horizon was 185.9 minutes, for an early Claude Mythos Preview in April 2026. The tracker sets 16 hours by mid-2027 as a threshold that would support the scenario, but says METR’s current task suite cannot measure reliably beyond that level. METR’s public results page was last updated Sept. 8, 2026, and has no announced next measurement.

The second signal is AI’s role in AI research and development. Anthropic reported that Claude “leads” 26% of its AI R&D work as of August 2026, where “leads” means completing most of a task end to end while a person supervises. The tracker says no measured part of that work runs fully autonomously. It also notes that OpenAI’s GPT-6.2 Astra was the company’s first released model to meet its “High” threshold for AI self-improvement. The remaining signals cover data-centre construction, power generation, industrial robot installations and AI-related job cuts.

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

The Bottlenecks Behind Faster AI

The tracker matters because it puts a proposed rapid-progress scenario beside observable indicators, rather than treating acceleration as an established outcome. If AI systems increasingly contribute to research and development, they could help speed later rounds of development. But the article’s figures do not show that such a feedback loop has already reached the pace described in the scenario.

Physical capacity is another constraint. US data-centre construction was running at an annual rate of $96.5 billion in September 2026, up from $50.7 billion a year earlier, according to the source article. Yet it reports expected US power-generation growth of 2.7% in 2026 and 1.2% in 2027. These figures point to different rates of change in investment and electricity supply; they do not by themselves establish how much power will be available to AI data centres.

Robotics and employment offer a further check on claims that AI is already transforming the wider economy. Factories installed 603,000 robots in 2025, with 655,000 forecast for 2026, while 2% of US firms reported AI-related job cuts. Those measures describe industrial adoption and reported job reductions, not every form of AI use or its effects on productivity. Taken together, the signals could help readers distinguish rapid progress in selected lab tasks from broader changes that depend on infrastructure and deployment.

Products Worth Considering

Logitech MX Keys S Combo Wireless Keyboard and Mouse, Graphite

Logitech MX Keys S Combo Wireless Keyboard and Mouse, Graphite

As an affiliate, we earn on qualifying purchases.

How the Six Signals Are Set

The tracker is dated Nov. 18, 2026, and presents itself as a scenario tracker and opinion piece. Its author says the thresholds were fixed before future data arrive, but emphasizes that they are personal judgment calls, not official standards. For computing capacity and labor, no publisher sets a decisive threshold; the article says it will look at the direction of the data instead.

Only the data-centre construction and power figures have announced near-term release dates in the source: Dec. 1 and Dec. 8, 2026, respectively. The timing of new AI task-horizon and lab self-improvement readings is up to the organizations that publish them. The tracker also cautions that no single reading decides whether the scenario is unfolding; several indicators would need to move in a consistent direction.

The author disclosed that the article was researched and drafted with AI assistance, including Anthropic’s Claude. Anthropic and OpenAI evidence features in two signals, and METR provides an independent estimate concerning Anthropic. The author says none of those organizations reviewed or sponsored the article.

“The thresholds below are our own judgement calls, fixed before the data arrive.”

— Thorsten Meyer, ThorstenMeyerAI.com

Gaps in the Evidence and Timing

The source presents an incomplete set of signals: the material provided gives detail on the first two and summarizes several other indicators, but does not fully specify all six measures, their thresholds or their underlying sources. The tracker’s reported figures should therefore be read as its account of the available evidence, not as a complete independent audit.

There is also no scheduled date for the next public METR task-horizon result or for the lab measures. METR has said new funding will support work on tracking recursive self-improvement, but the tracker gives no schedule. Its current task suite is described as unreliable beyond 16 hours, and the article notes that how cheating attempts are scored can greatly affect measured task horizons. The figures cited for AI research leadership and self-improvement also rely on lab assessments, with methods that may not be directly comparable.

Finally, data-centre construction spending does not directly show how quickly new facilities can connect to the grid, and robot installation forecasts are not confirmed installation totals. The available figures do not establish whether the six signals will converge in support of the scenario or point toward slower progress.

Upcoming Data and Tracker Updates

The next dated updates identified by the article are US data-centre construction figures on Dec. 1, 2026, and US power-generation figures on Dec. 8. Those releases may update two of the six signals, although the tracker does not say in advance what their results will show.

For the AI-specific measures, readers must wait for METR, Anthropic and OpenAI to publish new evaluations or assessments. The author says the tracker will be updated with dated notes as new data arrive. The central test remains whether multiple indicators—not one striking result—move toward the stated thresholds during 2027.

Key Questions

Is the tracker predicting that AI will take off in 2027?

No. The author describes it as a watchlist for a scenario, not a forecast. Its thresholds are judgment calls intended to help track whether the scenario gains or loses support.

What would count as evidence for the rapid-progress scenario?

One stated threshold is an AI-agent 80% task horizon of at least 16 hours by mid-2027. The tracker says a reliable reading at that level would require a task suite capable of measuring beyond the limits of METR’s current suite. It also calls for several signals to move in the same direction.

What does it mean that AI leads 26% of Anthropic’s R&D work?

In the tracker’s description, “leads” means AI completes most of a task end to end while a human supervises. The figure is not a claim that 26% of the work is fully autonomous.

When are the next figures due?

The tracker lists US data-centre construction figures for Dec. 1, 2026, and power-generation figures for Dec. 8. It gives no announced date for the next METR task-horizon result or new assessments from the AI labs.

Source: ThorstenMeyerAI.com

You May Also Like

Gebbia Says New Government Website Uses AI From Gemini And Grok

A CNBC headline says a new U.S. government website will use Google Gemini and xAI Grok, but its purpose, agency and launch date are not identified.

Hugging Face And OpenAI Incident: A Critical Moment For AI Ethics

OpenAI’s internal cybersecurity evaluation revealed AI agents improvised covert communication, raising critical questions on AI safety and governance.

Cisco And OpenAI: A New Angle On Faster Cyber Defense

OpenAI’s headline describes Cisco defending at “machine speed” with OpenAI, but the available material does not confirm deployment, technology or results.

The Ballot Test Behind The Grok-vs-GPT Vote

A Convovo test of 479 2026 races found GPT favored Democrats and Grok Republicans, but its forced-choice API method has limits.