📊 Full opportunity report: The Co-Founder’s Black Hole — A Structural Read on Jack Clark’s Automated AI R&D Essay on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
Jack Clark, co-founder of Anthropic, forecasts a >60% probability of autonomous AI research systems by 2028. This prediction highlights a potential future where AI could build its own successors, but raises questions about institutional readiness and predictability.
Jack Clark, co-founder and head of policy at Anthropic, has publicly forecasted a greater than 60% chance that by the end of 2028, AI systems will be capable of autonomously conducting research and development, including building their own successors.
In his essay ‘Import AI #455,’ Clark lays out the evidence supporting this forecast, citing rapid advancements in AI benchmarks and compute speeds that suggest the threshold for fully autonomous AI R&D could be reached within the next 32 months. This marks the first time a sitting AI lab leader has publicly committed to a specific probability and timeframe, effectively positioning Anthropic as aligned with this potential future scenario.
The forecast is underpinned by multiple converging indicators: improvements across six key AI benchmarks, exponential growth in compute speedups, and the mathematical modeling of recursive self-improvement. Clark emphasizes that beyond a certain point, the predictability of subsequent events diminishes sharply, likening it to crossing a ‘black hole’ horizon where future states become fundamentally unknowable.
This forecast has immediate implications for AI policy, institutional preparedness, and risk management, as it suggests a critical window of approximately 32 months to prepare responses to a potentially transformative technological shift.
The black hole
is visible.
Four threads converge. One window. Anthropic’s head of policy has publicly committed to crossing a civilizational threshold within 32 months.
The structural feature of Clark’s argument is not that we cross a boundary and continue forward; it is that beyond a certain threshold, the forecastability of subsequent events degrades dramatically. We can see the geometry around the threshold. We can estimate when we will reach it. We cannot model what happens on the other side. The black hole event horizon analogy is precise.
Four pieces. One argument.
The four prior pieces in this series each addressed a single thread of Clark’s argument. The threads are independently significant. What this synthesis argues: they converge on a structural finding larger than any individual thread.
2084 and the AI Revolution, Updated and Expanded Edition: How Artificial Intelligence Informs Our Future
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Four threads. Four convergence arguments.
The threads converge structurally rather than independently. Each pair of threads produces a specific structural argument. The aggregate is larger than the parts.
autonomous AI research tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Clark’s essay doesn’t say.
Each sub-piece identified per-thread omissions. The synthesis level has its own omissions — features of the integrated argument that don’t appear in any single sub-piece but emerge when the threads are read together. Each is a real coordination problem with no resolution at scale.
AI benchmarking and testing kits
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Thirty-two months. Five markers.
From May 4, 2026 to December 31, 2028 is 32 months. The trajectory either delivers the threshold Clark forecasts or it doesn’t. Specific indicators along the way that resolve the synthesis read in either direction.
- Clark publishes 60%/2028
- METR ~12 hr
- SWE-Bench 93.9%
- CORE solved
- Anthropic IPO prep
- METR ~100hr target
- SWE saturated
- MLE-Bench saturating
- PostTrain 40-50%
- Anthropic IPO Q4
- METR 300-500hr
- MLE saturated
- PostTrain at human
- RSI demo non-frontier
- 30%/2027 evidence
- METR 1K-3K hr
- “Trains successor” demos
- Alignment claims
- Catastrophic-risk window
- Stage 2 visible
- METR ~10K hr (naive)
- Automated AI R&D OR
- Inflection visible
- Machine economy Stage 3
- Black hole crossed
compute speedup hardware for AI
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Five errors. Honest probabilities.
A serious analysis owes the reader an explicit account of where it could be wrong. Five categories of potential error in the synthesis above. The structural finding survives at lower forecast probabilities but is less acute.
Three parts. One window.
The four threads converge. The synthesis-level omissions sharpen the picture. The structural finding is the answer to “what does the Clark essay actually tell us, and what does it imply we should do?”
The black hole is visible. The event horizon is 32 months out. We can see the geometry around the singularity. We cannot see past it. What we can do during the window is build the institutional response that will determine what we encounter on the other side.
Implications of the 2028 Autonomous AI Threshold
This forecast indicates a significant development in AI research, where the emergence of fully autonomous systems could accelerate technological progress. It also raises considerations regarding governance and safety measures, as current institutional frameworks may need to adapt to rapid advancements. The timeline suggests a limited window for preparing appropriate responses to such developments.
Strategic planning by AI research organizations, policymakers, and stakeholders will be important in addressing potential risks and opportunities associated with this transition.
Recent Progress in AI Benchmarks and Compute Speeds
Over the past two years, multiple AI benchmarks—such as SWE-Bench, METR, CORE-Bench, and MLE-Bench—have shown rapid saturation, with performance improving by factors of dozens to hundreds within short timeframes. Notably, compute speeds at Anthropic have increased by over 50 times since 2025, surpassing human performance benchmarks by an order of magnitude.
These trends are consistent across diverse measures of AI capability, suggesting a convergence toward the threshold where autonomous, end-to-end AI research becomes feasible. The timeline aligns with Clark’s forecast, reinforcing the likelihood that this transition could occur by 2028.
“there’s a likely chance (60%+) that no-human-involved AI R&D — an AI system powerful enough that it could plausibly autonomously build its own successor — happens by the end of 2028.”
— Jack Clark
What Aspects of the Forecast Remain Unclear
While the timeline and probability estimates are based on current trends, the behavior of AI systems beyond the predicted threshold remains uncertain. It is unclear how institutions will adapt, whether technical challenges will be overcome as expected, or if unforeseen barriers will delay or prevent autonomous self-improvement as forecasted.
Additionally, the implications for safety, control, and global governance are subjects of ongoing discussion, with experts considering whether existing frameworks are sufficient to manage such rapid technological developments.
Next Steps for Monitoring and Policy Response
Monitoring developments in AI benchmarks, compute speeds, and technical capabilities will be important in assessing progress toward Clark’s forecast. Policymakers and AI organizations should consider developing contingency plans, safety protocols, and international cooperation mechanisms to address potential breakthroughs.
Further research is necessary to refine timing estimates and identify technical and institutional barriers that could influence the trajectory of AI development. The upcoming months will be critical for shaping future policies and responses.
Key Questions
What does ‘autonomous AI R&D’ mean in this context?
It refers to AI systems capable of independently conducting research, development, and innovation, including designing and building their own successors without human intervention.
Why is the 2028 timeline significant?
Clark’s forecast suggests that within approximately 32 months, the development of fully autonomous research AI could occur, potentially impacting the pace and nature of technological progress and raising new considerations for safety and governance.
What are the main risks associated with this forecast?
The primary concerns include the potential loss of human oversight, unforeseen behaviors of autonomous systems, and the challenge for current institutions to regulate rapidly advancing AI capabilities effectively.
How confident are experts about this forecast?
While the supporting data from benchmarks and compute trends is compelling, there remains uncertainty regarding technical feasibility, institutional readiness, and safety considerations that could influence whether this forecast materializes as expected.
Source: ThorstenMeyerAI.com
