Smash Notes
Slack CTO Cal Henderson
Workplace chat and collaboration technology company Slack is going public on Thursday, with implications for a wide variety of companies in the Seattle region and beyond. Microsoft is Slack’s primary competitor with its Teams service, and Slack is a big Amazon Web Services customer. And many companies that sell technologies to businesses will be watching…
People
Notes
-
Cal Henderson has been working with the same set of people and co-founders for a few decades now. In 2000s they'd made a video game on the internet, which turned into Flickr and got acquired by Yahoo. A decade later they tried another online game called Glitch, which again did not work, but turned into Slack.
-
Slack is bringing hundreds of different work applications under the same roof, enabling companies to use a unified interface for all their communications and data.
While email was built on a one-to-one model of communication where new information was pushed on you, Slack enabled a one-to-many pull model, where anyone can subscribe to any information of interest.
Information in email is only born once an account is opened, and is as good as destroyed once the email account is closed. Meanwhile, Slack enables information to live on forever, so new team members never start with an empty slate, nor is their knowledge lost with transition.
-
It could be very addicting to dive deep into a technology and build something for massive scale, only to never see it come true. When starting Slack, the team was unconstrained by what the technology would become, instead they focused on finding the product market fit. The job was to build something that worked and to iterate fast.
-
It would have been much-much harder. In his first company, Cal had to buy physical servers and manually drive and install them in data-centers, which sucked up a third of his time. Moving to infrastructure-as-a-service enabled a whole new generation of startups.
-
Slack CTO Cal Henderson tells GeekWire that a Slack launched today would make fewer product mistakes, because the company now has the data. The cloud lesson is the one he cares about. He would move further up the stack, away from raw virtual machines and toward platform services, including Kubernetes and higher-level Amazon Web Services. The goal is to avoid undifferentiated work that every company does and that does not make Slack better at Slack. At Flickr, he says, a quarter to a third of the team's time went to that work: buying servers in Canada, talking to customs, racking them, then realizing they were short on network cables. Hosted services give that time back.
-
Slack is focused on ways to make engineers more effective and productive, lowering the technical barrier for DevOps. The idea is that if a team could build a feature, they should also be able to support it.
By embracing the latest cloud technologies like Kubernetes, they are simplifying deployment and maintenance for the whole organization.
-
Opportunity cost and engineering time are super expensive, and it makes sense to optimize for those, rather than try and save costs on web services.
-
GeekWire asks Slack CTO Cal Henderson whether a company at Slack's size can switch between Amazon Web Services and Google Cloud. He says the big clouds are more alike than they used to be, so a move is possible, and also a lot of work. As the service gets more complex, it depends on the small operational details of one provider, such as how a queue behaves under load or when it fails. Engineers build around those details. Broadly the services match. The cost is in the differences. Henderson adds that Slack would rather spend engineer time, which is expensive, on fast iteration than on keeping an easy exit to another cloud.
-
AI can certainly make small interactions better, thereby making your whole experience better.
It will be a while until AI is able to make drastic improvements in our use of software, probably not until it can achieve human-level intelligence.
-
Breaking-in is a strong term. Cal was just very interested in the game that Steward was building, so he changed the default admin password on Glit's mailing list to get some insights, with the intention to build an API on top of the game. It wasn't the greatest move, but it worked out at the time.
-
People you hire in week one of the company are very different from the people you should hire with 1500 employees. More and more work has become knowledge work, with ability to collaborate at its core.
-
How do you take software that you already use everyday, and use Slack to enrich the usage of those applications? Slack would like that, and they even have a Slack Fund to fund those.
-
On GeekWire, Slack CTO Cal Henderson says that for almost any software company that is not already a giant, the first job is finding product-market fit quickly. Slack was not designed on day one for more than 10 million daily users. The first version put in people's hands was the wrong one, and the early years were spent learning which parts of a tool that worked for small engineering teams would work for everyone else. He warns that building for massive scale before that fit is addictive for an engineer and a way to get stuck. Cloud hosting is what made the fast loop possible. At Flickr, he says, they bought servers and racked them themselves.
Related notes
- What Are the Key Ingredients of Starting a Business
- What Are the Downsides of Hosting a Website on the Cloud
- What Is Joe Heitzeberg's Advice to Every Entrepreneur