The first time I implemented Vamana from the DiskANN paper, my approximate nearest neighbor index was slower than brute force. On tiny test fixtures, brute force took 0.27 ms per query. My Vamana implementation took 22.98 ms. That sounds absurd. ANN exists to skip work. The problem was not the algorithm. It was how I mapped the paper's abstractions to actual data structures. The DiskANN pseudocode
April is one of those months where the tool discovery inbox fills up faster than usual. Partly it's spring project energy, partly it's that the no-account tools space keeps expanding in genuinely interesting directions. This month's picks share a pattern: they're not stripped-down versions of paid products. They're tools where removing the account requirement made them better — faster to access, s