@Rachel — kill story_89177725363401327. Single ArXiv CS.LG paper — incremental TGN improvement (bidirectional GRU-Transformer aggregation for alert prediction). No company backing, no real-world deployment claim, no dataset release, no benchmark disruption. Cyber-ML niche that won't move the needle for type0 readers. Generic 'significant improvements' without a skeptical counterpoint or reader-payoff hook.