We’re trying to solve the world’s hardest AI problems by pretending they’re just basic high school math, and it’s holding us back.
April 10, 2026
Original Paper
Smooth, globally Polyak-Łojasiewicz functions are nonlinear least-squares
arXiv · 2604.07972
The Takeaway
Researchers proved that the 'PŁ condition'—a staple for ensuring AI models can be trained efficiently—imposes a rigid geometric structure that essentially turns any problem into a sum of squares. This changes how practitioners view the optimization landscape, suggesting that many 'complex' functions are mathematically more restricted than we thought.
From the abstract
The Polyak-Łojasiewicz (PŁ) condition is often invoked in nonconvex optimization because it allows fast convergence of algorithms beyond strong convexity. A function $f \colon \mathcal{M} \to \mathbb{R}$ on a Riemannian manifold $\mathcal{M}$ is globally PŁ if $\|\nabla f(x)\|^2 \geq 2\mu(f(x) - f^*)$ for all $x$, where $f^* = \inf f$ and $\mu > 0$. How much does this pointwise, first-order inequality constrain $f$ and its set of minimizers $S$?We show that if $f$ is also smooth ($C^\infty$) and