A post originally shared by crypto analyst Steph Is Crypto, which has since been deleted, sparked discussion within the digital asset community due to an experiment involving Elon Musk’s artificial intelligence chatbot, Grok.
The post featured an image populated with several recognizable figures from the cryptocurrency industry. It detailed a prompt given to Grok, asking it to identify and remove the "worst CEO" from the depicted group. Grok's response, which resulted in the removal of Ripple CEO Brad Garlinghouse, became the central point of the post before its subsequent deletion.
While the immediate implication of Grok's action suggested that the AI system had identified Garlinghouse as the least effective chief executive among those presented, a closer examination of the image's composition and the professional roles of the individuals involved reveals significant context that complicates this interpretation.
A Mix of Roles, Not a Uniform Group of CEOs
Although Brad Garlinghouse is unequivocally the chief executive officer of Ripple, several other individuals featured in the image do not hold CEO positions. For instance, Anatoly Yakovenko, one of the co-founders of Solana, is positioned near Garlinghouse. In decentralized or semi-decentralized projects, a co-founder role does not automatically equate to the position of chief executive officer.
Another prominent figure included in the image is Vitalik Buterin, a co-founder of Ethereum. Ethereum does not operate under a traditional corporate structure, and Buterin has never held the title of CEO for the network. While his influence is widely recognized, it operates outside the conventional executive framework implied by the prompt given to Grok.
The image also contained Michael Saylor, who previously served as the CEO of MicroStrategy. Saylor has since stepped down from that role and now holds the position of executive chairman. At the time relevant to the image, he was no longer the company’s chief executive, rendering his inclusion under the label of "CEO" inaccurate in the context of the prompt.
Questions Regarding the Accuracy of the AI's Outcome
When considered in light of these differing roles, Grok's decision to remove Garlinghouse appears less like a comparative assessment of executive performance among peers and more like a flawed response stemming from an imprecise prompt.
If the majority of the individuals depicted were not CEOs, then the fundamental premise of identifying the "worst CEO" was inherently flawed. Garlinghouse was one of the few participants in the image who clearly met the criteria specified in the prompt.
We are on X, follow us to connect with us :- @TimesTabloid1
— TimesTabloid (@TimesTabloid1) June 15, 2025
This situation raises questions about whether the AI system actually evaluated executive performance or if it defaulted to selecting a recognizable CEO without properly accounting for the diverse roles of the other individuals present in the image.
Understanding the Limits of AI Interpretation
The now-deleted post from Steph Is Crypto ultimately served to highlight the limitations of employing AI tools for judgments that are heavily reliant on contextual understanding and role-specific accuracy. Grok’s response demonstrated how easily an AI-generated output can be misinterpreted when the underlying assumptions of a given prompt do not align with the actual reality of the situation.
Rather than functioning as a definitive evaluation of leadership quality, this incident illustrates how AI-generated conclusions can reflect deficiencies in contextual comprehension, particularly within an industry where titles and responsibilities are not uniformly applied across different projects.