Sunday Grab Bag - Random VC Thoughts
All Things Venture #103
I haven’t posted anything in awhile so I figured I’d share some of the things I’ve been thinking about, at a high level, about the VC industry. The format is pretty rough and really just an exercise to help me organize my thoughts but I figured since I wrote things down, I might as well share them out with you all as well.
Feedback is welcome. If you hate it, let me know. If you love it, let me know. Regardless, I hope you all have a great upcoming week & happy NYC marathon for all those who participated or had a loved one participate :)
Vertical AI Consensus - The Hype Continues!
We’re fully in a period of consensus surrounding the vertical AI opportunity. The cat’s out of the bag and capital is searching for entrepreneurs who agree with the broad based thesis of “selling work.” I have to give credit where credit is due here, so we should all really thank Sarah Tavel at Benchmark for her cogent thoughts all the way back in, *checks calendar* August of 2023. Since then, and I am guilty of this, the hive mind of the venture capital industry has churned out think piece, after think piece, after think piece with variations of Tavel’s initial premise that, “An AI-driven product with the consistency and SLAs it can achieve, should be vastly superior to an outsourced offering - a 10x and cheaper opportunity” Sell work, not software, secure the bag, this is the way
Non-Consensus - But Right, Hardware
Given this rotation, it feels pretty obvious that the net new, non-consensus but right opportunity is in hardware. Once we have AGI/ASI or whatever people want to water down superintelligence to, these mechanical brains are going to get put into objects. This is a pretty huge deal and, for what it’s worth, this already exists. Waymo, the self-driving car division of Google/Alphabet, 10x’d their paid weekly trips YoY to 100,000. That’s absolutely nuts. Cars are literally driving themselves and we’re all just kinda like meh. Hardware, like services, are going to be a pretty interesting category to watch in VC over the next decade and it’s going to push the world into a pretty wild place before we even know it
Computer Use, from Anthropic
I thought the Computer Use announcement from Anthropic was pretty sweet. The long and short of the announcement is that now, via the Claude API, you can create an RPA-like action via natural language. The example they provide in the announcement video is a pretty benign data entry task, but when you take a step back I think this is pretty incredible. I have been saying for nearly a year now that, “translating unstructured data into structured formats is the lifeblood of the professional services industry” and computer use is a perfect example of that
I think of Computer Use as a tipping point for AI native companies, it’s another feature set that helps weight the initial customer interaction to a wow, magic moment that ultimately leads to both A) conversion and B) more importantly, retention
Memory Lane, GPS enables Uber, a historical “tipping point”
A further thought on the “tipping point” idea. I think Uber, and the importance of GPS to that businesses’ success is super important/representative of the concept of a “tipping point” in product. Uber is obviously one of the largest VC backed winners of the 2000s and it’s a definitive example of mobile computing’s value creation. There’s also a pretty neat through line that demarcates Uber being possible and not possible
GPS wasn’t available until July of 2008 with the launch of the iPhone 3G, 8 months or so later, in March of 2009, Uber was launched. Pre-July 2008? Uber = not possible. Post-July 2008? Uber = possible
Herein lies the tipping point: For Uber, mobile computing + GPS = category defining business. For AI native agents, foundational LLMs + Computer Use = category defining business (maybe).
Thinking about it more, if I had to put a definition on a tipping point in the context of VC, I would define it as an emergent product feature that is readily available and undifferentiated but when combined with an emergent technology trend creates product market fit