I'm wondering if everyone enamored of this meme is thinking that software will allow industries to change the very definition of "core". I don't know anything a

I'm wondering if everyone enamored of this meme is thinking that software will allow industries to change the very definition of "core". I don't know anything about the bakery business, but scratching my head over your very good question led me to posit some "what if's" to try to find ways to change the "core" of a bakery business. Hopefully these corny off-the-cuffs will get the gist across of how some think this trend will redefine the "core" of various businesses. What if the bakery was built around an automated production system driven by software that lets a master baker key in the latest recipe tweaks to a baked good product that works great in a small batch, and automatically works out the changes necessary to scale it up (or maybe it is so automated that it can mass produce in small batches), sets up A/B testing in limited distribution areas, and measures the results...the very same day that the tweaks are made? Or it can permute variables in the first production run, yielding a few dozen samples laid out in an array, let the master baker (or apprentices) taste test, tell the system what result to modify, and dynamically adjust, permuting in a day of testing what would take a conventional bakery a week or more to work out what to do in production? What if the automated system is so modular and flexible that it can do whatever a human baker can do, simply by watching a baker perform a single rendition (with appropriately-instrumented tools to sense pressure, flow, orientation, etc.)? What kind of competitive advantage would it confer (if any) to allow a bakery to start mass producing baked goods on demand for special promotions and occassions, in the time it takes a conventional bakery to produce a single prototype? What if the marketing department has software that enables them to set up A/B testing for a different package artwork, and tweak it day-by-day, comparing against historical sales data and adjusting for economic conditions? What if the sales department has software that can deduce the relative stock levels of competitors from the phone cam video taken by delivery and stocking employees of the bakery that stock at grocery retailers, and pounce to deliver more stock when the competitors are running low on stock but won't know until a couple days from now on their regularly scheduled stock deliveries? These what-if's do not really apply to the baker that only wants a small corner bakery in the European style, serving only the neighborhood within a walkable distance, and taking a craftsman's approach to their pursuit of their art and livelihood. For that context, I believe you are right: the core business is all that really matters, and software is merely a tool for reducing frictional costs of doing business. I suspect Andreessen, Kirkpatrick, and Rao are thinking of mid-size and larger-scaled businesses when they discuss this observation/trend. The huge assumption here is that the flexible fabric of software bytes is translated across into some directable and malleable bits in the real world. That happens, but it many times doesn't come cheap. I wonder if the proponents of this trend are vastly underestimating those costs.