AI slop has a smell: "If you tweet at me with AI, I will block you" | Peter Steinberger

Quick Overview

Peter Steinberger states he blocks anyone who tweets at him with AI-generated content because he values raw, human-created content and finds AI-generated text, especially when it smells of AI, to be fraudulent and lacking the nuance he appreciates in human interaction.

Key Points: Steinberger has a strict policy of blocking users who tweet at him using AI, stating, "If you tweet at me with AI, I will block you" (0:09). He prefers raw, human-written content over AI-generated text, even if the human writing contains errors or typos (2:24). He finds AI-generated content, particularly when it 'smells like AI,' to be fundamentally 'fake' and 'fraudulent' (2:21, 2:55). Steinberger mentions experimenting with AI for creating blog posts, noting that the resulting diagrams and information were good, but the process was ultimately too much work and he is allergic to AI-generated code (2:51, 3:56). He expresses hope that the human experience will be empowered, not damaged, by AI tools, stating that tools should augment, not replace, human value (5:48). He finds it fascinating how allergic he is to even a little bit of AI slop in video images or text (4:08). He notes that while AI can accomplish tasks like generating diagrams, it's not going to be 'all your style' (3:15).

Context: This clip features Peter Steinberger in a discussion, likely on the Lex Fridman Podcast (as indicated by the intro graphic), where he articulates his strong negative stance regarding the use of Artificial Intelligence (AI) for generating social media content, specifically tweets directed at him. The conversation centers on the perceived lack of authenticity and value in AI-generated text versus genuine human expression, despite the imperfections that human creation might entail.

Detailed Analysis

Peter Steinberger strongly asserts that he blocks anyone who tweets at him using AI-generated content, emphasizing his preference for raw, human-created content, even if it contains typos or broken English (0:09, 2:24). He describes AI output that 'smells like AI' as 'fake' and 'fraudulent' (2:21, 2:55). Steinberger recounts an experience experimenting with AI to create blog posts where the diagrams were useful, but the overall effort was too much, and he feels an immediate negative reaction ('allergy') to AI-generated content, even in video imagery (3:56, 4:08). He clarifies that while AI can be a tool to accomplish tasks, such as drawing diagrams, it cannot replace the personal style or human element (3:15). He concludes with a hopeful note that AI should empower human experience rather than damage it, suggesting that the value lies in human interaction, which AI currently cannot replicate fully (5:48).