AI Optimism, Home Computing

AI Optimism

I am incredibly bullish on the potential of AI.  David Brooks does a nice job articulating why AI will be good for us, I was pointed to this by Jeff Raikes.   I like his remarks towards the end:

But A.I.’s ultimate accomplishment will be to remind us who we are by revealing what it can’t do. It will compel us to double down on all the activities that make us distinctly human: taking care of each other, being a good teammate, reading deeply, exploring daringly, growing spiritually, finding kindred spirits and having a good time.

This is generally the case for all technology — remove the tedium and bullshit from our lives, and allow us to be better human beings. Which is kind of the theme of this newsletter: using technology to make us better.

Home Computers

I joined Microsoft in the 80s because of its mission statement – “A PC on every desk and in every home.”  I was excited about the personal power and creativity of PCs; I wanted to be a part of that revolution.  And by gosh, it happened.  There are PCs on every desk and at every home.  

Now, while the industry succeeded in getting PCs into homes, these machines remain largely indistinguishable from their office counterparts, offering little in the way of truly ‘home-oriented’ design. PCs are great productivity and business tools, and they are optimized for desktop use.  But for home, well, there is nothing particularly home-oriented.   There are no material feature differences between your work laptop and what you buy for home.  You can buy some home systems that have funky backlit keyboards for gaming, but other than that, a home PC is a business PC.  

We predominantly use home PCs to do business-like things — communicate, connect to work systems, manage lists, create documents, etc.  Games are a unique home use but don’t drive any hardware differences, particularly now that every PC is getting AI inference jammed into it.  And much game and entertainment use has been siphoned off to consoles, smart TVs, streaming devices, tablets, and phones.  There is just very little “home” about a home PC.

Is there a case for unique home PC hardware?  Well, there are certainly unique tasks and settings in the home.  The business laptop is designed for a single user at a desk.  This is a typical setup at home as well, but we also sit at dinner tables and kitchen counters and share meals, have conversations, work on homework, play games, talk to grandparents and kids, share stupid dachshund videos, and other social activities.  Could this warrant a different hardware design?

Well, Apple is rumored to introduce some sort of home command center — the “tabletop robotic home device” will be a combo “smart home command center, videoconferencing machine and remote-controlled home security tool.”   This would be a unique form factor for the home.  Apple’s history suggests it will be a standout product and help establish the market.   I certainly hope so!  

A great-looking, stylish product that uses the robotic arm to provide a differentiated experience when multiple people are sitting around the table talking, playing games, working on homework, or sharing stupid dachshund videos could be great.   A deep focus on games, entertainment, education, and family fun would be valuable.  A great product would launch with day one support from top family board games — Catan, Ticket to Ride, Carcassonne, etc.  And day one support for some leading education products, both for kids and adults.  With a big enough catalog of fun family activities, this device could be a slam dunk.  

I am less excited about the idea of a smart home command center.  I have used boatloads of smart home equipment.  Most of them are in the trash.  The space is rife with products that don’t connect well, are not robust, and have terrible UXes.  The display of smart home info has never been a problem for me, and I am not sure a screen waving around on a robot arm will add much. The issues are all distributed – power, installation, connectivity, naming, discovery, geospatial location, uniform interaction language, etc.  And even if this was all solved, most people don’t want to spend time on the “ops” side of their home.  We just want all that stuff to work and work well. 

The demand for a unique home PC design isn’t going to arise because of better control of your house; it is going to come from the creative and enriching uses of the device.

AI workflow

I’ve used all the current and prior generation end-user AI tools.  I’ve monkeyed with prompts and context.   Much as I like playing around with all this, a lot of this usage is falling by the wayside; it just takes too much additional time on top of my current workflow to use the AI tools.

So, I really enjoy the tools that put AI into the context of my existing workflow.   Github Copilot is great; it works right in VS Code, which I was already using, and it works right in line with my editing process.   I don’t have to think about prompts or context; my current work IS my prompt and context.   I love this model.

Grammarly also works well for me (thanks for the recommendation, John).  It works in every editor on my Mac — Notes, browsers, whatever.  I don’t have to change anything about my workflow.  I have long ago learned to do all substantial drafts in a tool like Notes, where I cannot make the error of sending them out too early.  I only transfer over to email, blog editor, or newsletter editor when I am materially done with my writing.   I have learned too often that email drafting can lead to sorrow when you mistakenly send a draft.  So, I prefer to isolate my editing in an editing sandbox and want AI to work there.

For this reason, something like Google Gemini doesn’t work for me. It’s in the wrong spot in the workflow.  By the time I paste my content into a mail message, all the heavy editing lifting has happened.

People only have 24 hours a day, and their days are already pretty full, and no amount of innovation will change this.  So, asking them to spend time with a whole new experience doing whole new tasks is a pretty huge ask!  For broad horizontal tasks like writing, the winning tools will slip into existing workflows, which means the incumbents have a huge advantage.  Brand new tools and new companies will have a better chance of success for focused tasks in vertical markets – and there is a ton of opportunity there.