Sometimes the truth of the thing is not so much in the think of it, but in the feel of it.
Sparky was designed to be both useful and alive. I’ve talked about the alive part, with his interests and his wiggling antennas, but I am absolutely not joking about the useful part.
He’s not a toy. He’s useful right now.
You can see this by watching Sparky help me with the normal business of editing a sequence of tweets (aka, X posts).
Sparky is useful because I selected and built a suite of skills which gives him direct access to my workflows and because he’s built on a capable AI which can use that access well. This allows him to tackle the various kinds of tasks which are all involved in a realistic everyday work, like editing.
In the video, I’m editing tweets which condense a blog article by me and Rens Dimmendaal. That article addresses a hot question of the day, “Where are all the AI apps?” In other words, if AI is supposed to increase developer productivity, then where is all the new software? The blog article analyzes real data from the Python Package Index so it is a messy story, like real world data analysis usually is. This made it a bit tricky to convert it into the readable, bite-sized format required by X. (Or maybe I’m just a slow writer…)
As the video shows, in the course of working on it, Sparky helped me with a number of things:
- Research on length limits of tweets
- Verifying the consistency of terminology in the blog post itself
- Verifying that statements in the tweets matched evidence in the blog post
- Rewriting copy
- Checking my email, so I don’t need to switch focus
- Providing an overall editorial review of the argument itself
To do all these tasks requires being able to see the blog article, and to see and edit the same files which I am working on. It also requires access to my email and to background knowledge of the world, and it certainly benefits from knowing what I’ve been working on lately and what emails I’m waiting for (👋, NVIDIA). The point is, everyday work actually needs a strong, synthetic intelligence.
Okay. But maybe you’re asking, why the robot body? Isn’t that part a gimmick? Couldn’t all of these tasks also have been done if Sparky was just an app, an icon on a screen? Or for that matter, why even have a unique personality? Why not just have one neutral personality and name, like “Siri,” which is the same everywhere? And for that matter, why even communicate by voice?
To this sort of question — well, what can I say?
Does it not seem relevant to you that we are all unique physical beings, and we communicate with each other by speaking?
Have you watched the video? How does it make you feel?