Abstract: A random stranger approaches you as you wait for a train, and makes a comment about your body or attire, and walks away. Call this a hit-and-run encounter. Such encounters are common on social media like Twitter, with its open platform, searchable hashtags, and potential for lurking, mobbing, and flooding. Online hit-and-run speech is all about impact without speaker responsibility. On most accounts of language (Austin, Brandom, Grice, Lewis, others), speakers undertake responsibility when they speak, but hit-and-run online speech tends to evade such responsibilities. Anonymity makes such evasion easy, but the rapid pace of most social media platforms, the sheer speed of transmission, combined with the flood of transmissions, creates challenges for holding speakers to account for their utterances. This paper addresses questions of responsibility for speech online, with an emphasis on hit-and-run cases, contrasting them with non-anonymous chat spaces that are more like IRL conversations. Tracking speaker and hearer responsibilities in these different contexts will offer insights about the fraught issue of online counter-speech.