That would be much quicker by just opening the actual calendar. In the next steps they show how using an API key will make an online AI help the user to create pictures for the presentation and how summaries of a voice recording look like.
That does not connect any dots and is a bit short before the webinar expecting the webinar host to be ready that morning, to be honest. In this case they could use a local AI RAGed on presentation techniques, technical aspects of the webinar in the weeks prior to create that presentation or, considering that it is the very morning connect dots, combine data such as GPS position, traffic, prior meetings added after that Webinare was scheduled or GPS positions of coworkers needed for the meeting to understand the departure time or if the meeting is under jeopardy.
That there is building a pipe to a large and general AI model. That there is hardly more than a Co-Pilot called system used to enable the use of natural language changing user behaviour away from clicking.
With Open Source Ollama models, Linux and Python embedded tailored systems can be created for the needs of individuals and teams adding to the general AIs of the large companies especially if rules are added taking away processing weight from the AI.
If these Webinars all use the same files, a script can check of those are in the correct folder and warn if any is missing. Than an AI could start searching the networks it has access to including sending an email to a defined contact.
Consider that there above the first layer and wonder what is missing in an individual case of someone traveling in need of team support, to foreign destinations creating very different needs than for an individual being from a team, but not in need of that there and that keeps frequenting the same spots.
Cyberdecks will come.... the online AI will tell about the Diner restaurant and the local keep a record about it separating the AI guys from the AI guys, potentially.
#cyberpunkcoltoure
