Agentic AI Is Not Even Close To Ready. Lots of Shorts Hiding In Plain Sight.
After extensively using OpenAI, Anthropic. Gemini (Google) and now Replit (not so extensively), based on each’s lack of ability to debug code, there is a wide chasm between investor perceptions of Agentic AI and reality - especially as it relates to coding.
None of the AI agents are ready for prime time. None are close. This truth combined with valuations suggests that all of the previously-mentioned companies are shorts. This also goes for the many other companies that have rode the “AI” valuation tailwind, such as Oracle (ORCL), Microsoft (MSFT), and Palantir (PLTR), as well as every other company that experienced a significant valuation boost without a commensurate revenue boost.
I’ve felt this way for two/three years and the more I use agentic tools, the more strongly I feel about AI not being close to ready.