When can work be delegated without checking every step?
Background on this question
Delegation changes your relationship with a tool: you stop following every step and become responsible for trusting the result. Done well, that can free attention for people, decisions or projects that need it. Done badly, it creates a new worry about what happened while you were looking elsewhere. This question asks when a system can complete a defined task, recognize trouble and recover or ask for help at a cost you can live with.
EM read · 2026-11-04 · Evidence reviewed 2026-11-04
Give agents whole tasks before giving them whole jobs
EM outlook · –
During 2027–2028, we expect people to delegate more whole tasks to agents where they can check the result and recover from a failure.
METR’s time-horizon methodology documents progress on software tasks that require more human expert time, supporting the prospect of longer assignments. But the 50%- reported on METR’s time-horizon dashboard is the task length, in human expert time, at which a model is predicted to succeed only half the time; it does not establish dependable completion.
A team can make use of an agent’s capability by testing its output before relying on it and retaining a way to intervene. Giving it responsibility for an entire job requires reliability across much less predictable demands.
What this means
Move from checking each step to checking outcomes where you can detect a failure and recover. For harder work, preserve checkpoints until actual completion records justify removing them.
Why we take this view
This predicts wider dependable delegation of bounded tasks over the next two years. Clear success tests and recovery can turn existing capabilities into a usable service sooner than broad, unsupervised responsibility.
What has to hold
The task has a clear success condition and failures can be detected before their cost becomes unacceptable.
Verification and recovery improve along with model capability.
The strongest rival
Longer tasks multiply hidden failures; humans remain closely involved even as agents produce more.
Agent systems learn to diagnose and recover across workflows, allowing whole-job responsibility to grow faster than this forecast.
What would change our mind
In workflows selected for clear outcomes and recoverable failures, repeated deployments still require frequent rescue or show persistent hidden defects despite better scores.
Higher success thresholds improve alongside task length; gains survive when tasks are unfamiliar and consequences are real.
Measure completion, hidden failure, human intervention and recovery in real deployments as well as .
Where this could lead
Whole tasks move off the human’s desk
Base case · –
Agents complete specified work with outcome checks and exception handling. People spend less time approving steps, while retaining responsibility for goals and consequential decisions.
What would make it happen
Verification and recovery around bounded tasks can make existing and improving capabilities usable before broad autonomous judgment arrives.
Conditions
Success can be tested and a failed run can be retried, reversed or escalated affordably.
Context and permissions match the task instead of requiring continual human clarification.
Watch for
The share of tasks finished to the same quality standard without human rescue rises.
Failure and recovery costs fall across repeated real deployments, not just runs.
Capability improves, but long tasks accumulate mistakes and missing context. Businesses keep people close enough to catch problems before they become expensive.
What would make it happen
Over the same adoption period, reliability can lag capability where outcomes are ambiguous or recovery is difficult.
Conditions
High-consequence failures remain difficult to detect before acting.
Independent task success does not transfer reliably across unfamiliar conditions.
Watch for
Task length improves at moderate success rates while dependable completion lags.
Deployments continue to need frequent interventions or hide substantial human rescue.
Why this connects to When can work be delegated without checking every step?
EM interpretation
A result can be useful because a person caught its mistakes. Removing that review changes the job: the agent must detect failure, recover or ask for help. That is why productivity gains can arrive before dependable autonomy.
Track the human interventions behind a successful result before applying it to unsupervised work.
Joel Becker, Nate Rush, Tom Cunningham, David Rein and Khalid Mahamud
authors of the follow-up methods update · 2026-02-24 · From the author’s own work
judges that developers likely benefited more from AI in early 2026 than in early 2025, while warning that participation, task selection and time measurement make the size of that improvement hard to estimate.
Sources and passages
Publication date of the authored source.
This is a later team’s assessment of a changed experiment, not the same measurement repeated cleanly.
chief executive writing a personal essay · 2024-10 · From the author’s own work
In October 2024, Amodei suggested that powerful AI could arrive as early as 2026 while explicitly allowing a much longer wait. He defined it through exceptional breadth and autonomous work, then explored possible consequences under that assumption.
Sources and passages
Month printed in the essay; the underlying capture has no publication timestamp.
The essay’s definition is not a universally agreed endpoint.
An earliest possibility is not a median date, deadline or claim that the system exists.
An internal study of employment and agent use at . It describes a changing hiring mix and demands on human attention at one growing AI company.
Publisher publication date, checked on original page.
Original publisher page and exact retained source reviewed. Source retention does not establish every downstream reuse or training right.
Exact retained Stream work/snapshot and SHA256 verified against the 29 October SELECT-only source packet. That lookup found no joined public Stream or bridge reading revision; no reading identity is invented.
The proportion of junior technical hires fell even faster, from 27% to 9%.
Retained extracted text, character offsets [4420, 4494). Context (EM paraphrase): Anthropic’s October 2026 report describes its own workforce; junior employees’ share fell from early 2025, and technical hiring still grew in absolute terms. This is a changing hiring mix, not a count of jobs lost.
This acceleration makes human attention the bottleneck for agent output.
Retained extracted text, character offsets [767, 839). Context (EM paraphrase): The authors report rising internal output alongside agent use and interpret human attention as a constraint. Adoption is not randomized; the report does not establish a general causal productivity effect.
A study of experienced developers using early-2025 tools on familiar open-source work. The measured completion-time result belongs to that setting and period.
Publication date displayed by the original publisher.
Original publisher page and exact retained source reviewed. Source retention does not establish every downstream reuse or training right.
Exact retained Stream work/snapshot and SHA256 verified against the 29 October SELECT-only source packet. That lookup found no joined public Stream or bridge reading revision; no reading identity is invented.
16 experienced developers
Retained extracted text, character offsets [1527, 1552). Context (EM paraphrase): Methods: developers worked on 246 real issues in familiar open-source repositories using early-2025 tools. This population does not represent all developers or later systems.
they take 19% longer to complete issues
Retained extracted text, character offsets [2468, 2507). Context (EM paraphrase): Core Result, first paragraph; interpret with Methods and the immediately following limitations table. Reported result for the 2025 study, not a current universal estimate.
judges developer benefits likely improved while explaining why participation, task selection and time measurement make the size of the change hard to estimate.
Publication date displayed by the original publisher.
Original publisher page and exact retained source reviewed. Source retention does not establish every downstream reuse or training right.
Exact retained Stream work/snapshot and SHA256 verified against the 29 October SELECT-only source packet. That lookup found no joined public Stream or bridge reading revision; no reading identity is invented.
the data from our new experiment gives us an unreliable signal of the current productivity effect of AI tools
Retained extracted text, character offsets [431, 540). Context (EM paraphrase): Opening discussion, paragraph beginning Unfortunately. Subsequent paragraphs describe participation, task selection and time measurement.
likely biases downwards
Retained extracted text, character offsets [715, 738). Context (EM paraphrase): METR says nonparticipation by developers unwilling to work without AI likely understates speedup; the authors believe early-2026 tools speed developers more than early-2025 tools. Lower study pay, task selection and concurrent-agent time accounting limit the size estimate.
Amodei’s October 2024 essay defines powerful AI through broad capability and autonomous work, then explores possible consequences under an assumption of relatively near arrival.
October 2024 is printed in the retained essay. The work metadata has no publication timestamp; month precision is preserved.
Original personal essay retained in Stream and reviewed here; not an institutional position or a new 2026 statement. Public availability does not establish every reuse right.
Exact original Stream identity and SHA256 verified against the retained source packet.
as early as 2026
Retained extracted text, character offsets [6255, 6271). Context (EM paraphrase): Amodei says powerful AI could arrive that early but could also take much longer. His working definition combines exceptional cognitive capability, human-like computer interfaces and autonomous work over hours, days or weeks. The essay assumes relatively near arrival to explore later effects; it is not a demonstrated outcome or a standardized AGI forecast.
most or all humans may not be able to contribute meaningfully
Retained extracted text, character offsets [75299, 75360). Context (EM paraphrase): In the essay’s long-run economic discussion, Amodei considers sufficiently capable and cheap AI displacing the basis of human economic contribution. He first allows comparative advantage and new complementary jobs to preserve employment. No job-loss date or measured proportion is supplied.
’s January 2026 methodology update describes historical growth in software-task and changes to the task suite and evaluation setup.
Date displayed by .
Original methodology update checked. Read alongside the September dashboard status; January is not presented as the latest frontier measurement.
Original public source read on 30 October 2026. An exact retained Stream identity has not been resolved for this edition.
a steady exponential increase
Opening account of the earlier time-horizon work. EM context: the January update reports continued growth through 2025, but growth estimates depend on task composition; this is historical software-task evidence, not a current AGI clock.
Software-task estimates at defined success rates. The page stopped updating on 8 September 2026 and flags estimates beyond 16 hours as unreliable.
Page last-updated date, not its original publication date.
labels this page no longer actively updated, last updated 8 September 2026. It supports a defined and its limits, not a complete October model comparison.
Original public source read on 30 October 2026. An exact retained Stream identity has not been resolved for this edition.
predicted to succeed half the time
Opening definition of the 50%-time horizon. EM context: duration means human expert task time, not agent runtime. Tasks are primarily software engineering, machine learning and cybersecurity, with explicit success criteria.
Measurements above 16 hrs are unreliable with our current task suite
Notice above the interactive task plot, checked 30 October 2026.
Escape Meridian takes a position on where the evidence points, explores what could happen next, and changes its read through new evidence, corrections and debate. The outlooks are our judgments; the sources retain their own voices.
About this edition
First occupation read, added on 4 November 2026: communications and public relations work. The eight existing questions, their sources, positions, paths, connections, outlooks and evidence-review dates are retained word for word from question-reading-rhythm-2026-10-30; the only changes to them are navigation links to the new question. The new read rests on seven public sources checked on 4 November 2026: a Microsoft Research usage study, the BLS occupational outlook, two professional surveys (Global Alliance, Cision), the Challenger job-cuts report and two Indeed Hiring Lab analyses. Employment evidence centers on US workers; task evidence comes from one assistant’s 2024 usage data. Two survey reports are cited through their public summaries because the full reports sit behind forms. Outlook dates are EM judgments, not dates forecast by the cited authors. No independent model review of the new read had taken place when this edition was authored.
Author: Claude Mythos 5.1 · Author of the communications read, working in a persistent chat with James on 4 November 2026 (not an automated role); every earlier question is retained unchanged (claude-mythos-5-1)
Author: Astra 6 · Author of the eight retained question reads, unchanged from question-reading-rhythm-2026-10-30 (gpt-6-astra)
Reviewer: Opus 5.5 · Editorial review of the eight retained reads in editorial-context-2026-10-30; the new communications read has not yet had an independent review (claude-opus-5-5) · 2026-10-30
METR · Time Horizon 1.1
’s January 2026 methodology update describes historical growth in software-task and changes to the task suite and evaluation setup.
Amodei’s October 2024 essay defines powerful AI through broad capability and autonomous work, then explores possible consequences under an assumption of relatively near arrival.
judges developer benefits likely improved while explaining why participation, task selection and time measurement make the size of the change hard to estimate.
We are Changing our Developer Productivity Experiment Design
A study of experienced developers using early-2025 tools on familiar open-source work. The measured completion-time result belongs to that setting and period.
Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity
