An AI agent needs a loop to reconsider what comes next after each result, not to repeat the same step, because the situation may have changed. The loop gives the agent a chance to look again before acting again, instead of blindly following fixed sequences. Microblog @antgrasso
@antgrasso
-
Evaluating AI Model Efficiency Beyond Parameter Count
By
–
You found a new AI model with billions of parameters and you like what you see, but now you are wondering whether it is also efficient. Before getting excited about the size, ask one more question: what does it take to run that model at the scale you need? Microblog @antgrasso
-
Human-centered AI design starts with human objectives
By
–
If you decide what AI should do before deciding what people need to achieve, you are designing the system backward. Human-centered AI begins with the human objective, not with a model looking for a use case; technology gets its role after the goal is clear. Microblog @antgrasso
-
Early access to Tavus Griffin conversational AI agent
By
–
I had early access to Griffin from Tavus, and what impressed me was how quickly the interaction stopped feeling like a sequence of prompts and answers. The conversation simply kept moving.
-
Early access to Tavus Griffin Human Interaction Model
By
–
Tavus gave me early access to Griffin, and after a few minutes talking with Ashley, I understood why they call it a Human Interaction Model.
— Antonio Grasso (@antgrasso) 1 octobre 2026
The continuity of the conversation is what surprised me most.@tavus partner. https://t.co/rE1XsKe4J8Tavus gave me early access to Griffin, and after a few minutes talking with Ashley, I understood why they call it a Human Interaction Model. The continuity of the conversation is what surprised me most. @tavus partner.
-
Agentic banking: Customer AI interacting with Bank AI
By
–
The next shift may happen when Customer AI begins to interact directly with Bank AI. Agentic banking could change not only how banks operate, but also how customers engage with them.
-
AI agent delays shift from code to coordination as platforms converge
By
–
Enterprise platforms converge on data, models and orchestration layers. Choices on architecture now steer how fast AI agents act across processes, so delays shift from code to coordination. Source @Gartner_inc Link https://
gtnr.it/4tKg5vC via @antgrasso -
AI reduces instinct in decisions, human responsibility remains
By
–
AI enters decision flows once guided by judgment, shrinking the space for instinct. Teams must question results because responsibility remains human, gradually reshaping how trust forms within
-
Retrieval Augmented Generation adapts AI to business context
By
–
Retrieval Augmented Generation helps AI move from generic answers to answers tailored to the business context. It does not consist of integrating all of the company's knowledge into the model, but rather connecting the AI to approved sources so that the
-
Levels of AI adoption in Europe widen gaps
By
–
AI adoption in Europe is divided into levels, with some countries advancing faster than others. Gaps are widening as skills and workflows progress unevenly, forcing slower regions to react rather than plan.