📊 Full opportunity report: Engineering Is Automated. Research Is the Residual. on ThorstenMeyerAI.com — validation score, market gap, and execution plan.

Recent empirical data and expert analysis confirm that AI systems are now capable of automating the majority of core engineering tasks in AI research, while the research process itself remains less automated. This shift could significantly accelerate AI development timelines and reshape institutional strategies, making engineering a largely automated process.

Thorsten Meyer’s review of recent benchmarks reveals that AI models have reached near-saturation levels in core engineering skills, such as reproducing research experiments and competing in Kaggle-like challenges. For example, the CORE-Bench, which measures research reproduction ability, improved from 21.5% in September 2024 to 95.5% in December 2025, with one author declaring it ‘solved.’ Similarly, the MLE-Bench, assessing performance in Kaggle competitions, rose from 16.9% to 64.4% over sixteen months, surpassing mid-tier human performance.

Experts attribute these advances to improvements in AI model capabilities, including better kernel design and automated code conversion, which are now integrated into production-grade systems. The pattern across multiple benchmarks indicates a rapid approach toward saturation, suggesting that the engineering aspect of AI R&D is becoming fully automated. However, the research process—generating new hypotheses, designing experiments, and interpreting results—remains less automated, with the structural question of whether research can be automated at scale still open.

Impact of Automation on AI Development Speed

The near-complete automation of core engineering tasks in AI research implies that the bottleneck in AI development may shift from engineering to research itself. This could lead to faster iteration cycles and more rapid deployment of AI systems. However, it also raises questions about the role of human researchers and the potential need for new institutional responses to manage increasingly autonomous AI-driven research processes.

Recent Advances in AI Engineering Capabilities

Over the past two years, multiple benchmarks and research efforts have documented rapid progress in AI’s ability to perform engineering tasks. The CORE-Bench, MLE-Bench, and kernel design research demonstrate that AI models are now handling tasks traditionally performed by human engineers, such as reproducing scientific experiments, competing in data science competitions, and designing optimized hardware kernels. These developments suggest that engineering in AI R&D is approaching full automation, a trend highlighted by Thorsten Meyer’s analysis of Clark’s empirical work.

“The pattern across benchmarks indicates a rapid approach toward saturation, suggesting that the engineering aspect of AI R&D is becoming fully automated.”

— Thorsten Meyer

Unresolved Questions About AI Research Automation

While engineering tasks are rapidly becoming automated, it remains unclear how much of the research process—such as hypothesis generation, experimental design, and interpretation—can be automated at scale. The structural question Clark leaves open is whether research itself is a form of engineering at scale, which could mean the residual gap closes faster than anticipated. Additionally, the institutional and strategic implications of fully automating engineering are still under exploration.

Next Steps in AI Automation and Research Development

Researchers and institutions will likely focus on developing methods to automate research activities further, including hypothesis generation and experimental design. Monitoring the evolution of benchmarks and real-world applications will be crucial to assess whether research automation progresses at a similar pace. Policymakers and industry leaders should prepare for potential shifts in research workflows and the role of human researchers in AI R&D.

Key Questions

How close are AI systems to fully automating AI research?

While engineering tasks are nearing full automation, the automation of research activities such as hypothesis creation and interpretation remains uncertain and is an active area of investigation.

What are the implications for human researchers?

If engineering is fully automated, human researchers may need to focus more on strategic oversight, hypothesis development, and ethical considerations, rather than routine experimentation.

Could this accelerate AI development timelines?

Yes, automating core engineering tasks could significantly reduce the time required to develop and deploy new AI systems, but the pace depends on progress in automating research activities.

Are there risks associated with fully automated AI research?

Potential risks include loss of human oversight, difficulty in understanding AI-generated research, and ethical concerns about autonomous scientific discovery. These issues are still under discussion.

Source: ThorstenMeyerAI.com