“I can do this with any LLM”, but you didn’t that’s why you’re broke (OpenClaw)

Quick Overview

The speaker argues that the reason many people fail to replicate his business success is not due to the superior capabilities of advanced Large Language Models (LLMs) like GPT-4, but rather because they lack the necessary execution, such as running constant A/B tests, building comprehensive systems, and maintaining high output volume, which he demonstrates through his own detailed revenue tracking and operational routines.

Key Points: The speaker refutes the idea that advanced LLMs like GPT-4 are the sole differentiator for success, stating that execution, like constant A/B testing, is the real factor. He shares his January revenue update, showing total revenue of $3,235,661 across three startups, with $110,520 in recent 30-day revenue. The speaker highlights that his AI influencer system, which uses a local LLM, has doubled impressions daily since its inception, proving the effectiveness of self-built systems. He contrasts his comprehensive list of 12 daily tasks, which includes research, scraping, downloading, and mounting, against the common, simpler approaches of others. The speaker emphasizes that he is transparent, sharing his detailed revenue breakdown publicly, unlike others who only share curated successes. He notes that even with basic tools, execution (like running 30 scripts daily against competitors' Facebook Ad Libraries) is what drives results, not just the tool itself.

Context: The video addresses comments the creator received regarding his previous assertion that he could achieve similar results using any LLM, not just the most advanced ones like GPT-4, implying that people who fail are broke due to poor execution rather than technological limitations. The creator uses his own business performance data and a detailed list of his daily operational tasks to prove that consistent, high-volume execution is the critical missing piece for those who claim they cannot match his results.

Detailed Analysis

The speaker begins by addressing comments on his previous statement, "I can do this with any LLM," asserting that the reason people are broke is not the LLM they use, but their lack of execution. He shows his recent revenue update from Trustb8, noting total revenue of $3,235,661 across three startups, with $110,520 in 30-day revenue and $132,071 in estimated MRR. He details the performance of his three active startups: Indexsy ($859,520 total revenue), Advise.so, and LocalRank.so ($469,473 total revenue). He then reviews his Jan 31 episode notes, focusing on the list of 12 tasks he performs daily, which range from running a custom 'whitening digest' tool to building custom WordPress plugins and performing extensive research. He points out that his AI influencer system, which uses a local LLM, has doubled impressions daily since its inception, emphasizing that self-built, integrated systems outperform relying on basic, readily available tools. He criticizes the mindset of those who only point to superior technology (like Cloudbot) or who make false promises about what simple tools can achieve, arguing that execution—like scraping competitor ad libraries or running 30 scripts daily—is what separates success from failure. He concludes by stating that his transparency allows others to see the actual work involved, contrasting it with the common habit of only sharing curated successes.