How To Assess Whether AI Is Poised To Take Off In 2027
AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: How To Assess Whether AI Is Poised To Take 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, setting six indicators to watch for a possible acceleration of AI progress in 2027. Its author says the scenario is not a forecast; current evidence is mixed, and several important readings have no scheduled release dates.

ThorstenMeyerAI.com published a tracker on Nov. 18 for assessing whether AI could make a year’s worth of progress in about two months during 2027, setting six indicators and thresholds for a scenario it explicitly describes as a watchlist, not a forecast. The tracker says current evidence falls short of its central speed threshold, while research automation, data-centre construction, power supply, robotics and employment provide additional tests.

The tracker follows a scenario published separately by Thorsten Meyer: AI systems would take on more of the work of developing subsequent AI systems, potentially accelerating progress until constraints such as chips, buildings and electricity slow it. This question of AI-driven research progress relates to whether LLMs can assess research novelty. The new piece does not repeat the case for that outcome. It specifies what to measure, who publishes the data, where the indicators stood on Nov. 18, and what would strengthen or weaken the scenario.

The largest gap is in AI task performance. The tracker gives METR’s last public record for an 80% task-completion time horizon as 185.9 minutes, for an early Claude Mythos Preview model in April 2026. The scenario’s “gains” threshold is a measured horizon of at least 16 hours by mid-2027—roughly five times that record. METR has cautioned that its current task suite cannot reliably measure results above 16 hours, so reaching the threshold would require a suitable evaluation suite as well as a capable model.

For automation inside AI labs, the tracker cites Anthropic’s report that Claude “leads” 26% of its AI research and development work as of August 2026. “Leads” means the model completes most of a task end to end while a human supervises; the tracker says no measured portion runs fully autonomously. It also reports that OpenAI’s GPT-6.2 Astra is the company’s first released model to reach its “High” threshold for AI self-improvement. These are distinct measures, and neither lab has set a publication schedule for future readings.

At a glance
analysisWhen: Published November 18, 2026; future ind…
The developmentThorstenMeyerAI.com published a dated watchlist of six indicators for evaluating its scenario that AI progress could accelerate sharply in 2027.

The Signals Behind a 2027 Acceleration

The tracker’s value is in making a broad and uncertain scenario testable over time. A sharp increase in AI task horizons alone would not establish that AI is improving itself at a self-sustaining pace. Evidence that models lead more development work, alongside new evaluation results, would offer a separate indication that automation is extending into AI research.

Physical limits matter too. The source reports US data-centre construction running at an annual rate of $96.5 billion in September 2026, up from $50.7 billion a year earlier. But it cites expected US power-generation growth of 2.7% in 2026 and 1.2% in 2027. Those figures point to a potential mismatch between fast-moving construction and slower power growth; they do not, by themselves, establish how much computing capacity will be available or when it can be connected.

Robots and employment provide another check on claims that AI is rapidly changing 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. These measures move on industrial and business timelines, not necessarily at the pace of advances in software. Reading several indicators together may help readers distinguish a concentrated technical acceleration from a broader economic transformation.

Samsung 65-Inch S90H OLED 4K Smart TV Bundle with HW-B630F Soundbar, Subwoofer, and Screen Cleaner (2026)

Samsung 65-Inch S90H OLED 4K Smart TV Bundle with HW-B630F Soundbar, Subwoofer, and Screen Cleaner (2026)

KitchenAid Artisan 5-Quart Tilt-Head Stand Mixer, Almond Cream

KitchenAid Artisan 5-Quart Tilt-Head Stand Mixer, Almond Cream

As an affiliate, we earn on qualifying purchases.

How the Scenario Is Measured

The tracker identifies six signals: AI task time horizons; AI’s share of AI research; data-centre construction; electricity supply; industrial robot installations; and AI-related job cuts. The author says the thresholds are personal judgment calls, not official cut-offs. For computing infrastructure and labour, no publisher defines a threshold, so the tracker proposes looking at the direction of each series. No single reading decides whether the scenario is happening.

Publication timing is uneven. The source lists Dec. 1, 2026, for construction figures and Dec. 8 for power figures as the next scheduled releases. METR says its public page is no longer actively updated; its last change was dated Sept. 8, 2026. It has said new funding will support tracking recursive self-improvement, but has not announced a schedule. Anthropic and OpenAI have also not stated when they will publish further readings on their respective measures.

The article identifies itself as opinion and says it was researched and drafted with AI assistance, including Anthropic’s Claude. It discloses that evidence from Anthropic and OpenAI is central to two signals, and that METR’s estimate concerns Anthropic. According to the disclosure, none of those organizations reviewed or sponsored the piece. The thresholds and scenario interpretation are the author’s, not findings endorsed by the organizations supplying the data.

“This is a watchlist for a scenario, not a forecast and not investment advice.”

— Thorsten Meyer, ThorstenMeyerAI.com

Where the Evidence Falls Short

The tracker does not establish that AI progress will accelerate in 2027. Its task-horizon benchmark has a measurement ceiling near the proposed threshold, and its public results are not current enough to show whether models have since crossed it. METR’s estimate that Anthropic’s model improvements are proceeding at around twice an earlier rate is not equivalent to the roughly sixfold pace implied by a year of progress in two months; the source says METR would not be surprised by estimates ranging from 1.3 times to three times.

The AI-lab automation figures also have limits: “leading” supervised work is not full autonomy, and the tracker reports no measured portion of Anthropic’s research and development operating fully autonomously. The construction, power, robot and employment figures describe different activities and time periods; the source does not show that they directly measure AI-driven productivity or compute available to frontier labs. Several signals have no announced release schedule, so the tracker may not receive timely updates.

The supplied source material ends partway through its discussion of the second signal. It does not provide all six signal readings, thresholds or supporting details. The figures and descriptions above reflect the material provided, and a complete assessment of the full dashboard cannot be confirmed from it.

The Next Scheduled Data Releases

The tracker names Dec. 1, 2026, for US data-centre construction data and Dec. 8 for US power figures as its next dated checkpoints. Those releases can update the infrastructure picture, but will not settle whether AI systems are taking on enough AI-development work to drive a rapid feedback loop.

For that, readers must wait for further results from METR and new disclosures from Anthropic or OpenAI. The tracker says those organizations have not set dates for the relevant updates. Meyer says the page will be revised with dated notes as data arrive. The scenario gains support if multiple indicators move toward their stated thresholds; a single result, especially one constrained by measurement limits, would not resolve the question.

Key Questions

Is the article predicting an AI takeoff in 2027?

No. Its author calls it a watchlist for a scenario, not a forecast. The thresholds are the author’s judgments about what would strengthen or weaken that scenario.

What is the tracker’s main test for faster AI progress?

It focuses on METR’s 80% task-completion time horizon. The last public record cited is 185.9 minutes; the tracker says a result of at least 16 hours by mid-2027 would strengthen the scenario. METR cautions that its current suite cannot reliably measure beyond 16 hours.

Does Anthropic report that AI runs its research independently?

No. The tracker says Anthropic reported that Claude leads 26% of its AI research and development work, where “leads” involves completing most of a task while a human supervises. It says no measured portion is fully autonomous.

When will the next indicators be updated?

The tracker lists construction figures for Dec. 1, 2026, and power figures for Dec. 8. It gives no scheduled date for new METR, Anthropic or OpenAI readings.

Source: ThorstenMeyerAI.com

You May Also Like

How Product Catalogs Can Inspire Reels And TikToks

A proposed workflow would turn ecommerce catalog images and product details into faceless short-form videos, but its sales impact remains untested.

How New Experts Are Shaping Google’s AI & Economic Future

Google appoints leading economists and researchers to its AI & Economy program, aiming to study AI’s impact on productivity, labor, and scientific discovery.

Why Grok 4.6 Is A Milestone In AI For Long-Form And Persistent Agent Performance

SpaceXAI announces Grok 4.6, a model with a 500K context window for long-running agents, coding, and knowledge work, marking a significant development in AI capabilities.

The AI Boom Is A Big Reason China’s Trade Surplus Tops $1 Trillion Again

China recorded a $1.043 trillion goods surplus in the first 10 months of 2026 as exports of chips and computers surged, customs data show.