According to @LexSokolin, AI-driven automation will erode moats for fintechs built on manual processes, increasing competitive pressure and likely attrition among such startups, which is a material risk factor for investors; source: Lex Sokolin on X, Jan 19, 2026. According to @LexSokolin, large language models can generate middleware in seconds, compressing software differentiation and shifting value away from labor-intensive fintech operations toward automation-native architectures, a dynamic traders should factor into positioning; source: Lex Sokolin on X, Jan 19, 2026. According to @LexSokolin, firms integrating AI at the protocol layer will be more defensible than those relying on manual workflows, implying a relative advantage for API-first, automation-centric platforms and related crypto protocols with embedded AI logic; source: Lex Sokolin on X, Jan 19, 2026. According to @LexSokolin, for trade setup, this view favors exposure to fintech and crypto infrastructure that embed AI at the protocol layer and underweights operations-heavy models that depend on manual processes as automation advances; source: Lex Sokolin on X, Jan 19, 2026.

According to @milesdeutscher (source: https://twitter.com/milesdeutscher/status/1963334195046223935), he has been exploring ways to automate and streamline his crypto research process using large language models. According to @milesdeutscher (source: https://twitter.com/milesdeutscher/status/1963334195046223935), he notes that ChatGPT and Grok are strong options in his testing. According to @milesdeutscher (source: https://twitter.com/milesdeutscher/status/1963334195046223935), Google’s Gemini surprised him the most among the LLMs evaluated. According to @milesdeutscher (source: https://twitter.com/milesdeutscher/status/1963334195046223935), the update specifically focuses on automating crypto research workflows, which is directly relevant to trading research processes in crypto.

According to Andrej Karpathy, cryptocurrency trading platforms should revise their documentation to include command-line instructions like curl commands. This change aims to enhance usability for traders utilizing large language models (LLMs) for automation and efficiency, potentially leading to increased trading activity and user satisfaction.