Recursive Superintelligence emerged with a $650 million round and a $4.65 billion valuation to pursue recursive self-improvement rather than a conventional product roadmap.
何が変わったか
Richard Socher and Tim Rocktäschel lead a research-heavy team split between London and San Francisco.
The company’s premise is that AI can propose hypotheses, design experiments, write code and change its own training loop.
なぜ重要か
No public benchmark or product has been shown, so investors are underwriting people, compute access and a very open research problem.
次に見るべき点
見るべきなのは見出しの数字だけではなく、その製品、契約、プラットフォーム転換が今後2四半期で再現性を持つかどうかだ。
参考資料: CocoLoop