5 Comments
User's avatar
JMS TechNova's avatar

AI's future won't be decided by who builds the smartest model, but by who solves the most valuable business problems. Insightful analysis.

Tiny Tim's avatar

Great article. Rooted in reality. This doesn't mean that LLMs are not valuable. They're fantastic and powerful tools that have changed productivity significantly for those that use it appropriately. But this accurately support the claim that AI has been and is over hyped.

Natalie Hanson, Ph.D.'s avatar

Fantastic summary of Benedict Evan’s latest publication “AI Eats the World”. The comparative, longitudinal perspective is insightful. And the reframing of tasks versus jobs gave me a lot to think about.

Daniel Ionescu's avatar

900 million users is a lovely headline until you see how lightly most of them use it.

A lot of AI companies may be building for a habit that still doesn’t really exist.

AnalysisSocial Stocks's avatar

Great summary and a detail ‘nobody is talking loudly about’: AI infrastructure is obsolete in less than 2-3 years , if not faster, while Telcos’ infrastructure can lasts decades without major updates. The GPU/CPU AI CAPEX will need to keep coming to keep the cycle alive… same GPU/CPU CAPEX…