Phil Smeeton's Tech Thinking Blog

What makes a good personal assistant? Do they make you the perfect coffee (Tesla optimus) Do they create perfect documents (Microsoft co-pilot) Do they order you stuff (Alexa) Do they drive you places (Autonomous vehicles) Do they clean and tidy up work workspace (robo vacuums and Tesla optimus) Do they book your holidays (Web) Do they check the weather (App) Do they take notes and perform translations into another language?(Google translate)

The association of the individuals in silicon valley during the use of virtual meeting software by users for example computer experts mainly think about the speed and quality of their internet connection, the responsiveness and stability of the software application, security of the operating system etc, other users will think more about the human element e.g. the workers in silicon valley that have been involved in the project. In social science and in particular human relationships and networks there is the idea of strong and weak ties. An example of a strong tie would be a parent, an example of a weak tie might be a Deliveroo rider that you interact with occasionally. My point here is that relationships formed and maintained using virtual meeting software have a strength of tie, stronger than telephone and email and potentially weaker than in-person ties. Within the realm of virtual meetings a certain strength of tie potential exists that a person might want to adjust or maintain. Wa...

Camera views: Front Back Side 1 Side 2 The win of using multiple cameras during a virtual meeting: Increased recognition of social cues, facial expressions, micro expressions, body language, hand gestures. The additional views can be included in the virtual meeting software, the small wireless cameras should be positioned accordingly, the increased bandwidth/speed of internet connections make the multiple streams of video viable.

Users are becoming used to reading smart output from generative AI and they are enjoying the benefits of the advice they are receiving from more and more advanced LLMs.  One idea I had was to periodically create messages that report system and application status, for example your application could do with updating or you should check your system for disk space. Whilst this is not revolutionary information users might act on the information like renewing hardware more often. Like an agent that keeps an eye on your system and tell you when it's time for a new computer.

With many decades of history, digital computers have been through some interesting times and forms, zooming straight into the present, I think that quite a considerable domestication has taken place with people often taking technology for granted and 'making do' (as Hall would put it) with the tech they currently own instead of investing and learning new devices. Even with the introduction of groundbreaking generative AI the tech industry has a lot of work to do by becoming less talented at its job. With all these super reliable faultless, secure and fast networks, users have become used the IV morphine of computer services. The expectations of the users must fit into a psychology of the privilege of a working fast and secure computer system, and not treat it as a given. Should the computer(s) require fixing, this is a premium paid service that sometimes incurs a delay due to demands on the technician(s). I think the lessons learned recently include being over efficient at our ...

The way users often view hardware is that it is mostly similar to another configuration, when thinking about the equipment involved in online virtual meetings we can see a screen, camera, keyboard etc. I think most tech professionals would agree that a certain hardware configuration produces certain results as the hardware configuration differs, so do the results to an extent. One point I think is of value (particularly to western people) is the symmetry, hair line and aging of users appearance on virtual meeting calls. I propose a software application that runs either as part of the virtual meeting software or as a separate application. The function of the application is to make subtle but relevant adjustments to the appearance of the users, for example making the users face 5-10% more symmetrical without over doing any particular enhancement.

With a good number of users using generative AI to help them solve problems and explain complex ideas as well as for code generation and content creation, how are these users getting on with it? Some early research by Microsoft suggests that the technology is a success with a high percentage of users reporting they feel more productive and capable in the workplace whilst using Gen AI. I recently attended a tech conference where AI LLMs where compared (there are quite a few to choose from currently) and one point that was raised was how your company's LLM ought to be designed to produce (Gen AI) advice in alignment with organisational policy. So it's important to know how your Gen AI LLM is biased and you might want to train in your organisations own policy/ideology/bias so that your staff users are given the right advice in the context of the organisation. Also at the same conference the topic of speed appeared in the context of Gen AI, I'm personally less concerned about t...