Frontier AI labs are increasingly utilizing autonomous systems to handle their internal development workflows. Anthropic's research pace on defined tasks doubled over the prior year due to extensive internal model usage, while power users at OpenAI sharply escalated their spending on coding agents. Similar automation previously enabled models to solve open Erdős problems and secure Olympiad gold medals through verifiable proof methods. Experts indicate this rapid technical momentum could produce systems that match elite human specialists across all domains by 2030 or even sooner.
Key points
Frontier laboratories are doubling their research velocity by integrating autonomous codi…: Frontier laboratories are doubling their research velocity by integrating autonomous coding and research agents into daily operations.
Current machine capabilities advance most rapidly in verifiable domains by finding counte…: Current machine capabilities advance most rapidly in verifiable domains by finding counterexamples and automating complex technical proofs.
Comprehensive automation of research workflows could compress historical development time…: Comprehensive automation of research workflows could compress historical development timelines, yielding expert-level cross-domain intelligence by 2030.
Sources checked by our analysis pipeline: 1
Many thanks to Max Farrens, Gabe Fried, and Thomas Larsen for pairing with me through significant parts of writing and researching this essay.
You can watch or listen to me narrate this essay on YouTube, Spotify, and Apple Podcasts.
If you asked me earlier this year how far away we are from superintelligence, I’d probably have said at least the better part of a decade. But over the last three months, I’ve updated my views a ton. I’ve been debating people about this question, both on and off the podcast. And I’ve been watching the progress of AI on verifiable domains and thinking about what that implies for AI research itself.
I’ve now come to the view that we’re only a couple of years away, and possibly only a couple of months away, from AIs that are as good as the best humans at AI research. I’ve also come to believe that automating AI research will compress so much progress into a single year that, soon thereafter (so by 2030, but plausibly as early as next year), we’ll have AIs…
The feature introduces agentic commerce by allowing artificial intelligence to execute financial transactions directly on behalf of consumers. To safeguard funds, the checkout process mandates biometric confirmation and masks sensitive credit card data through digital tokens. However, adoption hinges on whether external merchants update their fraud systems and refrain from blocking automated bot traffic on their platforms.
Software industry leaders at the AIE Code Summit 2026 have introduced a set of guiding principles to counter the spread of unvetted automated development. The initiative aims to alleviate the asymmetric review burden placed on engineering teams while curtailing corporate pressure that prioritizes sheer output over quality. By re-establishing human accountability, the authors hope to prevent machine-generated errors from compounding across modern codebases and knowledge systems.
Investors are pouring fresh capital into developer infrastructure as enterprises transition from experimental machine learning models to production-ready autonomous systems. LangChain is positioning its management platform as a vendor-neutral safeguard that prevents businesses from being tethered to a single cloud or model ecosystem. This major cash infusion reflects surging corporate urgency to deploy and control autonomous workflows at scale.