Six Clues To Whether AI’s Momentum Builds In 2027
AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: Six Clues To Whether AI’s Momentum Builds In 2027 on ThorstenMeyerAI.com

Age 18–24?Offer from Amazon

Prime made for students and young adults

  • Fast, free delivery for dorm and study essentials
  • Prime Video and Amazon Music included
  • Member-only deals
Try Prime for Young Adults Free trial for eligible 18–24 year olds
As an affiliate, we earn on qualifying purchases.

TL;DR

ThorstenMeyerAI.com published a six-signal watchlist for judging whether a rapid AI-progress scenario is unfolding in 2027. Its current evidence is mixed: lab-reported AI research automation and rising data-center construction point toward faster development, while METR’s public task record remains far short of the article’s threshold and power, robotics and job data suggest slower physical change. The thresholds are the author’s judgments, not official forecasts.

ThorstenMeyerAI.com published a six-signal tracker on Nov. 18 for monitoring whether a scenario of sharply accelerating AI progress is taking shape in 2027. The watchlist compares current evidence with author-set thresholds for AI task performance, automation of AI research, computing infrastructure, electricity, robotics and employment; it is a scenario tracker, not a forecast.

The tracker follows an earlier scenario article arguing that AI systems could increasingly help build successor systems, potentially compressing a year of progress at the previous pace into about two months. The new piece does not make that case anew. It identifies what to measure, who publishes the data, when updates may arrive and what results would strengthen or weaken the scenario. Its author says the thresholds were fixed before the relevant data arrived, but describes them as judgment calls rather than official cut-offs.

On AI task performance, the article cites METR’s last public 80% time-horizon record: 185.9 minutes for an early Claude Mythos Preview model in April 2026. The tracker’s threshold for strengthening the scenario is a measured horizon of 16 hours or more by mid-2027, roughly five times that public record. METR says its current task suite cannot reliably measure results above 16 hours, so a result at that level would require a new or longer evaluation suite.

For AI-assisted research, the tracker reports that Anthropic said Claude “leads” 26% of its AI research and development work as of August 2026, with “leads” meaning the system completes most of a task end to end under human supervision. The source says no measured part of that work runs fully autonomously. It also says OpenAI’s GPT-6.2 Astra reached the company’s “High” threshold for AI self-improvement. These are company assessments, not independent demonstrations that AI systems can autonomously conduct research.

At a glance
reportWhen: Published November 18, 2026; future dat…
The developmentThorstenMeyerAI.com published a tracker on Nov. 18, 2026, setting out six indicators and thresholds for monitoring a rapid-AI-progress scenario in 2027.

The Gap Between AI and Infrastructure

The tracker matters because its indicators test two different parts of the rapid-progress scenario: whether AI can accelerate the research process, and whether physical resources can support that acceleration. Software capability and infrastructure do not move at the same speed. A model may help with research, but training and operating systems at scale also depend on data centers, chips and electricity.

The infrastructure figures cited in the article show both expansion and a possible constraint. US data-center construction spending reached a $96.5 billion annual rate in September 2026, compared with $50.7 billion a year earlier. Meanwhile, the cited forecast expects US power generation to grow 2.7% in 2026 and 1.2% in 2027. Those figures do not by themselves establish whether electricity supply will limit AI development; they show why construction and power need to be tracked alongside model performance.

The other signals guard against reading laboratory progress as immediate economy-wide change. The source says factories installed 603,000 robots in 2025, with 655,000 forecast for 2026, while only 2% of US firms reported AI-related job cuts. These observations point to a distinction between rapidly improving tools and slower adoption across workplaces. No single reading decides the scenario, the tracker says; the direction of several indicators matters.

Shark PowerPro Reveal Plus Cordless Stick Vacuum, Charcoal (IZ376H)

Shark PowerPro Reveal Plus Cordless Stick Vacuum, Charcoal (IZ376H)

Sony WH-1000XM6 Wireless Noise-Cancelling Headphones – Midnight Blue

Sony WH-1000XM6 Wireless Noise-Cancelling Headphones – Midnight Blue

As an affiliate, we earn on qualifying purchases.

How the Six Signals Are Set

The tracker follows a scenario published earlier by Thorsten Meyer, who argues that AI systems could take on more work involved in developing future AI systems. In that account, rapid progress would eventually run into limits in chips, buildings and power, before a slower physical cycle involving robots, factories and laboratories became more important. The Nov. 18 article is designed to test that scenario against future observations rather than present it as an established outcome.

The six signals cover longer task horizons; AI’s share of AI research; formal lab assessments of self-improvement; data-center construction; electricity generation; and industrial robots and employment. The article says publishers do not provide official thresholds for some measures, including compute and labor, so it will track their direction instead. Its dashboard includes both reported observations and the author’s own interpretation of what would count as a gain or loss for the scenario.

The source discloses that the article was researched and drafted with AI assistance, including Anthropic’s Claude. It also says evidence from Anthropic and OpenAI is central to two signals and that METR’s independent estimate concerns Anthropic. The author says none of the three organizations reviewed or sponsored the tracker. Those disclosures matter because several key readings depend on information published by the companies whose systems are being evaluated.

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

— Thorsten Meyer, ThorstenMeyerAI.com

The Missing Readings and Caveats

Several of the fastest-moving indicators have no announced publication dates. METR’s public results page was last updated on Sept. 8, 2026, and the source says the organization has not scheduled its next public time-horizon measurement. Anthropic and OpenAI have also not stated when they will publish further readings relevant to the tracker. The source says the next scheduled dates among its indicators are Dec. 1 for construction data and Dec. 8 for power figures.

The time-horizon comparison is also limited by measurement uncertainty. METR’s existing suite is not reliable above 16 hours, and the tracker cites an example in which assumptions about attempts to cheat changed a model’s estimated 50% horizon from around 11.3 hours to beyond 270 hours. METR did not consider either figure a robust measurement. The source also reports that METR estimates Anthropic’s model-improvement rate at around twice its earlier rate, while saying it “would not be surprised” by estimates from 1.3 to 3 times; that range is far below the roughly sixfold pace associated with a year of progress in two months.

Other uncertainties concern how company-defined labels such as “leads” and “High” compare across labs and over time, and whether construction, electricity or robot installations will translate into greater AI capacity. The source material supplied for this article gives only partial details for some of the six signals. Neither the thresholds nor the scenario’s outcome are established facts; future data may be delayed, revised or measured differently.

The Next Scheduled Data Points

The tracker identifies Dec. 1, 2026, for US data-center construction and Dec. 8 for power figures as its next announced data points. The author says the article will be updated with dated notes as new readings become available. Those releases can indicate whether infrastructure growth is continuing, but they will not settle the broader scenario on their own.

Readers will also need to watch for new METR task-suite results and further AI research-automation or self-improvement assessments from Anthropic and OpenAI. No publication timetable is given for those measures. The central test laid out by the tracker is whether several signals move together: much longer, reliably measured AI task horizons; a rising role for AI in AI research; and continued growth in the physical capacity needed to develop and run systems.

For now, the tracker’s evidence does not establish that a year of AI progress will arrive in two months. It records some signs of rapid investment and lab-level automation, alongside a large gap between the latest public task-horizon record and the author’s threshold, plus slower-moving power, robotics and employment measures. The 2027 scenario remains something to monitor, not a confirmed forecast.

Key Questions

What is the six-signal tracker?

It is a watchlist published by ThorstenMeyerAI.com on Nov. 18, 2026, for monitoring a proposed rapid-AI-progress scenario in 2027. It lists measures, data sources and author-set thresholds; it is not a forecast.

What result would strengthen the tracker’s AI capability signal?

The article says the scenario gains support if an AI agent reaches an 80% task time horizon of at least 16 hours by mid-2027. METR has cautioned that its current task suite cannot reliably measure above that level, so an appropriate evaluation method would be needed.

Does Anthropic say its AI systems work autonomously on AI research?

No. According to the source, Anthropic said Claude “leads” 26% of its AI research and development work as of August 2026, with human supervision, and that no measured part runs fully autonomously.

When are the next scheduled data releases?

The tracker lists Dec. 1, 2026, for construction data and Dec. 8 for power figures. It gives no scheduled dates for the next METR, Anthropic or OpenAI readings.

Is the 2027 rapid-progress scenario confirmed?

No. The source describes it as a scenario to monitor, and its thresholds are the author’s judgment calls. The current evidence is mixed, and future measurements may change the assessment.

Source: ThorstenMeyerAI.com

You May Also Like

Solving 20 Erdős Problems with 20 Codex Accounts Running in Parallel

Researchers used 20 AI accounts to solve 20 open Erdős problems, marking a significant advancement in mathematical problem-solving with AI.

Accessibility issue triage board for small websites

A new accessibility issue triage board for small websites is being tested as a simple workflow to help small business owners prioritize fixes after audits.

GPT-5.6 Used A Prompt To Close A 30-Year Gap In Convex Optimization

GPT-5.6 used a novel prompt to close a three-decade gap in convex optimization, marking a breakthrough in mathematical problem-solving.

The Switch: You Never Owned the AI You Depend On

Recent events reveal that AI reliance is vulnerable to instant shutdowns by governments and companies, exposing dependency risks without ownership rights.