Software eats, AI augments?
Software Eats, 4.0; plus, MBAs ❤️ semis, genAI and... AMC
Friends - this past week was the final week of classes in Strategy for full-time MBAs. For first-year MBAs at Berkeley Haas, fall and spring semesters are divided into A and B half-semesters, and Strategy comes in Spring B, along with Ethics. This gives the class a bit of a capstone role, and I do try to reference concepts introduced in other Core courses, such as Marketing, Operations or Macroeconomics.
Across two cohorts (Oski and Axe cohorts) we had 24 study groups, and thus 24 final projects. In a sign of the times:
Out of 24 topics, three were directly semiconductor-related; and four were on generative AI-related topics.
In 2023, with about the same number of project groups, we had one chip-related topic. (That year, a team - I see you Albert Deng - that made a compelling case that Intel should exit foundries; apparently Intel didn’t get the memo and persists with 5N4Y and IFS.)
Some topics are evergreen - every year I’ve taught Core Strategy, student groups have proposed topics related to Amazon, Apple, and Tesla…. and, wait for it — AMC Entertainment Holdings.
Yes, there’s something about $AMC that’s just hard to let go. This year, a thoughtful team of MBAs (hat-tip Jenny Hays and team) assessed the cost of conversion of theater spaces and alternative uses for them, such as hosting e-sports events. A quick look at comparables for AMC on Capital IQ highlights the challenge AMC faces as a very physical company with semi-fixed operating costs competing with Netflix as a substitute good.
A warm welcome to new subscribers! Some of you are current or recent students; others came via Babbage🙇♂️ . In this newsletter I generally cover themes related to my instruction:
Strategy for the Networked Economy
Clusters: Locations, Ecosystems and Opportunity
Business in Japan
Competitive Strategy
Thanks for being here! Looking forward to your feedback.
We are now almost 13 years on from Marc Andreessen’s famous WSJ editorial: Why Software is Eating the World (PDF here). I’ve incorporated Software Eats as a theme in Strategy for the Networked Economy since 2016.
Over time, I’ve versioned this a bit, into Software Eats 1.0, 2.0, and 3.0.
1.0: value accruing to the OS and productivity software, from hardware, as embodied by Windows and Android
2.0: will be familiar to readers of Ben Thompson and Fred Wilson (Web 2.0 era, Etsy and Tumblr-era Fred Wilson, perhaps). This is the disintermediation and re-aggregation enabled by the Internet.
3.0: this describes sharing economy entrants with variable physical capacity. A simple example: it’s a lot easier for Airbnb to add capacity during Dreamforce Week than it is for a traditional hotelier. A San Francisco resident can choose to host *just that week*. And players like Airbnb have had the impact you might expect - hotels can’t surge price (as much), and overall consumer surplus has been, generally, increased. (This last point is in the aggregate; absolutely circumstances may vary at the the regional level.)
I experienced this regional variance recently during a trip to Washington DC - taxis, which use to be abundant (if not always pleasant) in my period of regular travel to the District (2004-2010), were non-existent, but Uber/Lyft weren’t great either. Lyft functioned better and that’s what we relied on. (I guess hailing with a click was better than searching for a hotel with a cab stand. But only marginally.)
At the firm level, responses to Software Eats may vary. IBM, for example, has over the decades divested a lot of its hardware businesses, including semiconductors.
This March, I had the pleasure of joining Foundation Capital for their annual LP event. Foundation general partner (and proud EECS alum) Joanne Chen presented Foundation’s view on AI and the enterprise and what’s at stake.
The $13B under Sales & Marketing refers to the size of Salesforce’s core CRM business (Salesforce overall is now around $34B in revenue). Conversely $1.1T is what is spent on sales and marketing. What percent of this becomes software, or software-assisted? Suffice it to say it’s time to add a Software Eats 4.0 category to the taxonomy above - AI Eating, or Augmenting, Software. Whether this is Eating, which implies substitution, or Augmenting (e.g. CoPilot-y services) and Accelerating may vary.
In natural resources, Jevon’s Paradox, first posited to describe increasing use of coal as coal got cheaper to extract, gets to the heart of this - usually if the cost of doing something goes down, we do more of that thing, not less. On that hypothesis, then, perhaps 4.0 is: AI Augments? Please let me know your thoughts in the comments!
To help us discuss this topic, in Week 9 of 10 in Strategy for the Networked Economy (back in March), we hosted Kakul Srivastava, CEO of Splice. Splice provide audio creation tools and samples to audio creators and producers. (Regular readers may remember this post, from when Kakul joined class in 2023.)
Kakul presented a timely message - that software, and subtle hints from it, can assist human creators and make it easier to give voice to their ideas. That what we need is not computer-generated music, but rather, more humans making music. Splice refers to this as AI that empowers.
There’s a pyramid in music, from those who enjoy (the wide bottom of the pyramid?); to those who play an instrument themselves; to those who actually aspire to create and share their craft; and then those who try to make a living doing so. This last category is the narrow top tip of the pyramid. Can this segment below the top - those who create and share - be better served? Can barriers to creation be lowered? If so, this is indeed AI that empowers creators.
It was an uplifting talk. And the right talk for the moment. I was moved to see one of my students move rooms during our session together on Zoom. After the break, she shared she had been inspired to spend some time on her instrument.
Coming up this week - the RSA conference in San Francisco. Excited to attend, and looking forward to seeing my faculty colleague Andrew Reddie moderate a panel on Confidence Building Measures for AI in National Security. Excited also that various secretaries - State, and DHS among them - will be among the speakers this week.
Onward and upward,
Jon