📊 Full opportunity report: When AI Builds Itself: Inside Anthropic’s Evidence on Recursive Self-Improvement on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
Anthropic’s new report provides data indicating AI systems are now capable of automating significant parts of AI research and development. While full recursive self-improvement is not yet achieved, the evidence suggests it could happen sooner than expected, raising important questions about AI progress.
Anthropic’s new report presents concrete data indicating that AI systems are now capable of automating substantial parts of AI research and development, a development that could lead to recursive self-improvement if certain remaining gaps are closed. While the authors emphasize that this is not an inevitability, the evidence suggests such a scenario could arrive sooner than most institutions expect, making this a critical moment for understanding AI progress.
The report from Anthropic’s Institute highlights that AI models are increasingly performing tasks traditionally done by humans in AI development, such as writing code and conducting experiments. Public benchmarks show rapid improvements, with models now handling complex tasks that once required days of human effort. Inside labs, data indicates AI is already accelerating the pace of research, with engineers producing eight times more code per quarter than in previous years. The core argument is that AI is closing the gap in the ‘doing’ of AI research, but the ‘deciding’—selecting which problems to pursue—remains a human-led activity. The authors caution that full recursive self-improvement depends on automating this decision-making process, which is not yet achieved, but the evidence suggests it could happen soon if current trends continue.When AI builds itself
Anthropic is delegating a growing share of AI development to AI. Taken far enough, that points to a system that designs its own successor — recursive self-improvement. Not here yet, not inevitable. But the case isn’t speculation: it’s data on what AI is doing to AI development right now.
The curve that hasn’t bent
METR tracks the length of tasks AI can reliably complete on its own. That horizon is doubling roughly every four months — up from every seven. Anyone can check this in public data.
Task horizon — how long a job AI can handle solo
Each model handles dramatically longer tasks than the one a year before. The line keeps going up.
![Claude AI for Beginners Bible: [5 in 1] The Ultimate Guide to Automate Your Work, Save Hours Every Week, and Use AI for Real-World Results](https://m.media-amazon.com/images/I/415+fSJacsL._SL500_.jpg)
Claude AI for Beginners Bible: [5 in 1] The Ultimate Guide to Automate Your Work, Save Hours Every Week, and Use AI for Real-World Results
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Two kinds of work, one persistent gap
Building a frontier model splits into engineering and research. Across both, the pattern is the same — and so is the one thing AI still can’t do well.
Code, infrastructure, training
Claude can take an underspecified problem and find a method. Humans supply the goal; they no longer need to supply the method.
Which experiments, what they mean
Claude can match or outperform skilled humans at executing a well-specified experiment. But choosing which experiment still needs a human.
The same ladder Anthropic employees climb with experience
AI VoiceWriter – Smart Dictation & AI Writing Assistant for Windows & Mac | USB Dongle & Mobile App for Voice Input, Proofreading, Rewriting & Multilingual Support
🎙️ Hands-Free Voice Typing for Windows & Mac – Powered by iOS & Android dictation technology, AI VoiceWriter…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Watch the human share shrink, rung by rung
Walk up the four stages of AI development. At each, the human/AI split shifts — and the real internal numbers show exactly where AI has reached parity, gone superhuman, or still trails. Tap a rung.
The human role across the development loop
The doing now costs almost nothing in human time. What’s left is the deciding.
LangChain for Life Sciences and Healthcare: Innovation Through LLMs and Generative AI Agents
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Agents ran an open research project end to end
April 2026: the first demonstration of Claude running an open-ended research project from hypotheses to findings — on a real AI-safety problem.
Can a weaker model reliably supervise a stronger one?
Agents were left to solve it: proposing hypotheses, testing them, sharing findings across parallel agents, iterating. Measured against the gap between a “floor” (weak supervisor alone) and “ceiling” (strong model trained on correct answers).
(humans: ~23% in a week)
· ~$18,000 compute
the agents themselves
Absolute Zero Reasoner: Self-Improving AI Systems (Toward Artificial SuperIntelligence Book 4)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Picking a better next step than the human
Real research sessions where a human took a wrong turn. Models saw only the work before the detour and proposed a next step; a judge that knew the outcome scored them. The day-to-day of research is this chain of next-step calls.
“Can the model pick a better next step than the human?”
Share of moments where the model’s next step was judged better. The amber line is the practical ceiling (an ideal answer that could see the whole session).
It depends on whether the trend continues — and what we do
The piece refuses a single prediction. It lays out three scenarios, and is clear about which it finds most likely.
The exponentials turn out to be S-curves
Maybe taste can’t be scaled into existence; maybe the constraint is the supply chain — chips, grid, interconnect — not intelligence. Even so, the world still changes: Glasswing’s Mythos found 10,000+ critical vulnerabilities in weeks, and a 100-person firm does the work of 1,000.
included for completeness · they doubt itDevelopment automates; humans still steer
100-person companies doing the work of tens of thousands — revolutionary, but turnable to harm (population-scale surveillance, tailored manipulation). Bound by Amdahl’s law: speeding one part shifts the bottleneck — which is exactly why human code review became Anthropic’s new chokepoint.
★ they think we’re likely heading hereAI designs and refines its own successors
Progress paced only by compute. Humans move to oversight of an expanding “virtual lab.” The future they understand least — especially whether alignment holds, or whether rare misalignments compound as models build successors, until control slips.
the one they’re most uncertain aboutBuild the option to slow down — verifiably
The piece ends on policy, not product. A unilateral pause just changes who leads; what’s missing is the ability to verify others have actually slowed.
Why a credible pause is hard — and worth building toward
A slowdown that only lets the least cautious catch up leaves everyone less safe. So the goal is the option: systems that let frontier labs verify others have genuinely stopped. Anthropic says if such systems existed and peers paused verifiably, it expects it would too.
Detection beats verification — and even that’s tough
Training runs are easier to conceal than missile silos, inputs are general-purpose, and whoever continues while others pause inherits the lead.
We’ve done it before — slowly
Regimes like the INF Treaty built verification and trust over decades. The authors’ blunt line: “We don’t have that long.”
Reading it in proportion
- This is one lab’s account of its own internal data — much previously unreported, not independently audited.
- The soft spots are stated in the original: lines-of-code overstates productivity; the self-reported 4× is probably high; the headline research result didn’t transfer to production scale; the next-step test used cherry-picked moments.
- “More autonomous” is not “fully autonomous” — every standout result still had a human framing the problem and defining success.
- That the authors surface these caveats themselves — against their own incentive — is part of what makes the document serious.
Implications of Accelerated AI Self-Development
This evidence indicates that AI systems are already significantly impacting their own development process, potentially leading to rapid, autonomous improvements. If AI can automate both the technical and strategic aspects of research, it could drastically shorten development cycles and reshape the future of AI innovation. This raises questions about control, safety, and the readiness of institutions to manage such capabilities, making it a crucial development for policymakers and researchers alike.Recent Trends in AI Capability Benchmarks
Over the past few years, public benchmarks like METR, SWE-bench, and CORE-Bench have shown consistent, rapid improvements in AI performance on tasks related to coding, bug fixing, and research reproduction. For example, models have progressed from handling simple tasks to managing hours-long projects within a year. Inside labs, data shared by Anthropic reveals that AI now contributes a large portion of code development, with over 80% of code merged by AI in May 2026, up from single digits in early 2025. These trends suggest a pattern of accelerating capability that could underpin future recursive self-improvement if the strategic decision-making gap is closed.“The data from Anthropic indicates that AI is already automating substantial parts of its own development, which could lead to a rapid loop of self-improvement if the decision-making gap closes.”
— Thorsten Meyer, AI researcher
Uncertainties Surrounding Autonomous Goal Selection
It remains unclear whether AI systems will soon be able to autonomously select research goals and design their own successors without human input. The current evidence shows progress in executing tasks but significant gaps persist in strategic decision-making capabilities. The timeline for closing this gap is uncertain, and experts warn that unforeseen technical or safety challenges could delay or prevent full recursive self-improvement.
Next Steps in Monitoring AI Self-Improvement
Researchers and institutions will likely focus on tracking further internal data from labs, especially regarding AI’s role in strategic decision-making and goal setting. Public benchmarks may continue to improve, but the key indicator will be whether AI begins to autonomously design, evaluate, and optimize its own architectures. Policymakers and safety researchers will also scrutinize these developments to assess risks and prepare appropriate regulatory responses.
Key Questions
What is recursive self-improvement in AI?
Recursive self-improvement refers to an AI system’s ability to autonomously improve its own capabilities, potentially leading to rapid, exponential progress in its intelligence and functionality.
Are current AI systems capable of fully automating their own development?
No, current systems are improving in automating specific tasks like coding and experiment execution, but they have not yet demonstrated the ability to independently set goals or design their own successors.
Why does this development matter for AI safety?
If AI systems can autonomously improve themselves, it could accelerate the pace of development beyond human control, raising safety, control, and ethical concerns that require careful management.
How soon could recursive self-improvement happen?
The report suggests it could occur within a few years if current trends continue, but the timeline remains uncertain due to technical and strategic challenges.
What should institutions do in response to this evidence?
Institutions should monitor internal AI development data closely, invest in safety research, and develop policies to manage the potential risks of autonomous AI self-improvement.
Source: ThorstenMeyerAI.com
