The Human Edge in the Age of Machine Intelligence
Why human expertise still outperforms AI models in offensive security and ethical hacking — and how human operators and AI combine into a defence greater than either alone.
Whitepapers, security briefings and original research from the BytePatch team — on cyber security, data privacy, India's DPDP regime and building software that's secure by design.
PDF · 16 pages
Why human expertise still outperforms AI models in offensive security and ethical hacking — and how human operators and AI combine into a defence greater than either alone.
PDF · 9 pages
A plain-language briefing decoding India's Digital Personal Data Protection (DPDP) Act, 2023 and the draft DPDP Rules, 2025 — key obligations, consent & notice, the 72-hour breach window, data-retention timelines and what every business must do to stay compliant.
PDF · 40 pages
This project evaluates deep-learning-based cryptanalysis against the Speck32/64 lightweight block cipher (p. 4). It compares a classical differential baseline against a deep residual neural distinguisher (p. 4). The trained network achieves test accuracies of 0.885 at 5 rounds and 0.715 at 6 rounds (p. 4). Finally, the work implements a vectorised last-round subkey recovery attack and evaluates few-shot transfer learning (p. 4).










