Two δ×1 tubes crossing at a shallow angle θ overlap in a thin parallelogram of area about δ2/θ — the more nearly parallel they are, the longer they run side by side. Adding up the characteristic functions of all N tubes into a single 'multiplicity function' f, the integral ∫f2 exactly counts, with multiplicity, all pairwise overlaps ∣Ti∩Tj∣. Because the N directions are spread δ-evenly across a range of angles of size ∼1, summing δ2/∣θi−θj∣ over all pairs produces a harmonic-series divergence — a single factor of logN — rather than anything worse.
∣Ti∩Tj∣≲min(δ,∣θi−θj∣δ2),f=i=1∑NχTi
Detailed analysis
Order the tubes so θi≈iδ for i=1,…,N. Fix Ti: for each k=1,…,N there are O(1) tubes Tj with ∣θi−θj∣≈kδ, each contributing overlap ≲δ2/(kδ)=δ/k (capping at ∣Ti∩Ti∣=δ for k=0). So ∑j∣Ti∩Tj∣≲δ∑k=1Nk1∼δlogN. Summing over all N choices of i, ∫f2=∑i,j∣Ti∩Tj∣≲NδlogN∼logN (using Nδ∼1). Restoring the more standard length-N, radius-1 tube normalization used in Guth's lecture notes (equivalent by rescaling) gives the frequently quoted form ∫f2≲N2logN.