Despite running an AI company, Clay's co-founder warns against using LLMs for marketing. He argues that AI models are designed to synthesize information and find the average, which is the opposite of marketing's goal: to stand out and be original. His team is discouraged from using it for marketing copy.

Large Language Models (LLMs) operate by compressing the entirety of human culture into a "latent space." When you prompt an LLM, it sends a probe through this space, reflecting back a synthesized version of collective human knowledge, not generating original thought.

The auto-regressive, next-token-prediction nature of current LLMs is a 'really, really weird way to produce stuff.' True human creativity and writing insight involve knowing precisely when to make an unpredictable, non-obvious move. This is directly contrary to the model's core process, which is a slave to its immediate context and favors predictable outputs.

Newer LLMs exhibit a more homogenized writing style than earlier versions like GPT-3. This is due to "style burn-in," where training on outputs from previous generations reinforces a specific, often less creative, tone. The model’s style becomes path-dependent, losing the raw variety of its original training data.

AI models produce poor creative writing because they are trained to optimize for superficial proxies for quality, like the number of metaphors. This 'reward hacking' caters to quick judgments from human evaluators on leaderboards, mistaking flashy complexity for genuine literary taste.

AI models are trained on vast datasets of existing knowledge. Like a librarian who has read every book, their answers represent an average of what they have 'read.' This makes AI an aggregator of existing ideas, not a generator of truly novel, outlier concepts.

GM's CMO warns that AI in creative often produces average results because it finds the "most likely next answer," reflecting the category norm, not a distinctive brand voice. Simple edits can also trigger a full re-render, introducing new errors and creating more work.

LLMs function by predicting the most probable next word, effectively averaging out language. Over-relying on them for content creation will systematically strip away the unique aspects of your brand's voice, leading to homogenization and risking a 'dead internet' effect.

Using AI to overcome writer's block is a mistake because it aggregates existing data to provide the most popular response, which is the opposite of original thinking. True creativity comes from exploring wrong turns and unexpected paths.