If you like this newsletter, make sure you subscribe; otherwise, you won’t hear from us next month.
This is a16z’s newsletter for all things enterprise and B2B. From AI to open source to software-as-a-service, enterprise software to company building, we share what we’re seeing, hearing, and talking about in our own hallways.
Open source is more valuable than ever before. When Sun Microsystems bought MySQL for $1B back in 2008, a $1B exit was the exception; the valuation of most open source companies, such as RedHat, MySQL, and XenSource, were dwarfed by their proprietary counterparts (Microsoft, Oracle, and VMWare). But new business models — most notably software-as-a-service — have changed things. And in just the last two years, Mulesoft, MongoDB, Elastic, Cloudera, and GitHub have all had multibillion-dollar exits.
With open source, however, competitive advantage isn’t code — it’s community. So how exactly do you turn developer word of mouth into enterprise sales? What metrics matter at different stages? Can community become a moat against the threat from public clouds? Here’s our framework and a podcast with Ali Ghodsi, CEO of Databricks, and Armon Dadgar, CTO of HashiCorp, on these questions and more.
Natural Language Processing (NLP) is having its ImageNet moment, which means that software is getting surprisingly good at understanding language in the same way that it got good at figuring out what’s in a picture. State of the art language models (many named after Sesame Street characters like AllenNLP’s ELMo and Google’s BERT; OpenAI broke the naming convention by calling theirs GPT-2) have gotten so good that researchers have raised the benchmark standard for NLP from GLUE to SuperGLUE.
And just as more accurate computer vision models unlocked new use cases (self-driving cars, analyzing factory work), more accurate language models will unlock new use cases around understanding language, from figuring out if customers are getting mad during customer service calls to determining if a specific document is relevant for a trial to the literally life-saving models of Crisis Text Line that can predict which people are the biggest suicide risks.
What does this all mean for businesses? Just storing text-based data or recording calls is no longer enough. Now is the time to “read” that data in search of customer and product insights.
Enjoying the newsletter, subscribe to get it monthly.
Check out the following a16z podcasts and videos to hear more on:
1. Why AI is the fastest moving trend to hit enterprise. “No trend has grown as fast in terms of real spending, in terms of headcount, focused on whatever measure you want to use.”
—Paul Daugherty, CTO, Accenture in Automation + Work, Human + Machine
2. Why AI is something we design, not something coming for us. “With careful, thoughtful, empathic design, we can enable ourselves to live longer, safer lives. We can create jobs where we are doing more creative work. We can understand each other better.”
—Frank Chen in Better Together: Humanity & AI
3. Why AI will automate tasks, not jobs, and how to find common ground with machines. “When computers move into a workplace, they never take anybody’s entire job. They take away certain tasks. All of us have a certain percentage of our time that is spent on very codifiable rules-based knowledge work that given the right algorithm, could be turned over to a computer.”
—Julia Kirby, Author in Automation, Jobs, & the Future of Work (and Income)
4. Why AI is a copilot that can make us more focused and better at what we do. “AI is augmenting your ability to make decisions. Think of it as you are in a race car and there’s someone next to you that knows what’s around the curve and is whispering in your ear, what to be ready for, what’s the best next thing you could do.”
—Oleg Rogynskyy, Founder, People.ai in AI in B2B
5. Why AI will turn every worker into an analyst. “They need to know right now whether or not to respond to a competitor’s flash sale and what segments of their customers to target to make sure to not lose any customers.”
Recently, security topics have reigned on the a16z news show 16 Minutes, where we shared the latest takes on SIM swaps, iOS exploits, BEC fraud, and the CapitalOne data breach (and others like it). In related mobile security news, Apple exploits are now cheaper than Android; business email compromise (BEC) has become more expensive than ransomware in cyber insurance claims; and the CapitalOne breach highlighted we need a new security approach to multi-cloud configuration.
On that last point, while some argue that the future is single cloud, the future is already here and multi-cloud for two reasons: 1) The average enterprise has over 90+ SaaS applications. These applications run in a mix of public cloud and vendor-built clouds. 2) The cloud is not an abstraction — it is a set of servers in a location that relies on internet from an ISP and power from a power grid. Only saving to one cloud creates vulnerabilities by depending on a single vendor. Smart enterprises will run something in one cloud and back it up in another.
Two big trends: first, cloud-native technologies, most notably Kubernetes, are gaining traction on GitHub, as evidenced by the number of issues opened. Second, developers seem to be gravitating to open source AI projects: real-time voice cloning, image-to-image translation, and automated anime video remastering were all among the most starred GitHub repos.
Yah, we get it. Here’s a photo from a recent (private!) event we hosted for portfolio founders and other builders. While the event was private, the presentation and panel discussion are now open to the public.
We hope you enjoyed the first edition of the a16z enterprise newsletter — subscribe to receive future editions in your inbox.