Wednesday, April 29th 2020

A Proposal for Cloud Translation Layers

We have reached peak cloud. It's time we standardize productivity data.

Productivity Infrastructure Today

I believe that digital work today has four fundamental pillars: files, people, messages, and knowledge. For each pillar, tens or hundreds of services exist with only minute differences in execution or focus.


Dropbox, Google Drive, OneDrive, Box, S3, and others. Files also live in places that aren't explicitly 'storage' like Slack, Gmail, CRMs, and even our calendars.


Email addresses, Facebook accounts, CRM entries, iPhone contacts, Slack members, LinkedIn, GitHub. Our digital footprints comprise so many different addresses and profiles.


Text: LinkedIn, Email, Slack, Messenger, Text, SMS, IRC, WhatsApp, Discord, Twitter, Drift, Intercom, iMessage, Twilio, Signal, SnapChat.

Audio/Video: Slack, Messenger, Zoom, Google Meet, Discord, SIP, VoIP, GoToMeeting, WhatsApp, Skype, FaceTime.

Fundamentally, users are looking for ways to connect with others. Every day, I access between five and ten messaging services in order to stay in touch with the people in my life. If I wanted to send a digital message, I bet I could find more than 100 ways to do it. Some of those messages might have blue bubbles, or grey, or green, some will be plaintext, and others will have richer formatting. A message is a message is a message. It's the same story with audio and video; I could start a real time call with anyone through a dozen different services. There are far more tools that can send text, voice, or video than we need. All of them are different takes on the same fundamental interface for connecting with others through messages or calls.


Notion, Slight, Slab, Evernote, Zendesk, Crunchbase, Wikipedia, Confluence, Freshdesk,, Trello, Basecamp.

While knowledge can certainly be stored in files, it’s a distinct pillar because of the ways in which it is organized. For example, a flat list of files does not itself contain explicit knowledge, but a folder hierarchy does. What sets knowledge apart from arbitrary data is the effort people have made to organize and link it for others to navigate and consume.

Productivity Infrastructure Tomorrow

Across each of the four pillars of digital work, there is an opportunity to universalize their underlying infrastructure by abstractly understanding the atomic concepts upon which today’s productivity tools are built. This understanding can power translation layers that seamlessly move data to and from existing services. With these layers in hand, we can tightly link or even replace all our tools while simultaneously eliminating context switching. This new infrastructure is the foundation upon which we can radically rethink productivity.


With my team at Rethink, I have already successfully created a universal infrastructure for binary data. It is now possible to access a file anywhere on the cloud with just a single pathname. To do this, we created a translation layer that can abstractly read and write data over any API.


The translation layer concept is equally applicable to other data. We can develop a system that understands an atomic 'person' and can link the many addresses at which they can be contacted such as Slack, Email, phone, social media, and even physical addresses. By integrating the many tools through which we connect with people, we can build the most comprehensive possible understanding of our networks. This universal contact book reduces friction and removes the onus of knowing which of the many communication tools to use in order to reach someone.


Universal messaging is where these translation layers get really interesting. We could send messages across services, and create communication systems that are not concerned with where messages come from or where they’re going.

When combined with a universal understanding of the people with whom we communicate, this concept becomes exponentially more powerful. Any person you need to reach is immediately accessible to you through one tool instead of ten. We can build a single inbox for everything: emails, texts, slack, social networks, and many others.

Integrated messaging platforms have already been proven to be phenomenally effective.

Intercom does it for customer communication channels. Through their interface, I can reply to messages coming in from email, on-site chat, or Facebook. This enables me to focus on helping customers instead of where I should go to type my response. This is an incredible tool and it astounds me that no one has created the same kind of translation layer for any message. I would use this universal messenger for my personal and work communication in a heartbeat because I am tired of being spread over so. many. different. services.

Audio and video can also be wrangled. Infrastructure has become so commodity that the very same universal messenger could support real time communication just as easily as text content.

A translation layer for communication would remove context switching entirely—the greatest single threat to productivity—and instead allow people to focus solely on communication itself. No longer do we need to consider the text box into which we're typing our message or the color of the bubble in which the message will be sent. We could even understand physical destinations. There exist APIs today that can send a paper and ink postcard using just an HTML formatted message and a POST request. Because it is trivial to support any communication API in our translation layer, hitting send in Gmail could result in the arrival of a letter if we so desired.


While the concept of a unit of ‘knowledge’ is a bit more nebulous and harder to reduce than a file, person, or message, each of the other translation layers are already facilitating knowledge storage and retrieval, albeit unknowingly. So, there’s an opportunity to build a fourth translation layer for knowledge based on an understanding of the structures and organizational systems used to identify arbitrary information as ‘knowledge’ (such wikis, tables or kanban boards). Furthermore, by integrating all four translation layers, we can power a universal knowledge base that can self-populate and self-organize in response to any new information that is created or received.

Why Now?

Today, our dependence on an overwhelming number of siloed tools is starting to work against us. Tens of thousands of businesses have thought deeply about modern productivity and put forth a plethora of products that help us to connect with one another or create, consume, and organize information. Studying the thinking captured in their products has shown that it’s possible to reduce the problems they solve to these four fundamental concepts: files, people, communication, and knowledge. The creation of translation layers will help us to cut out the overhead of navigating disjointed tools and scattered data. They can also power a completely new breed of cloud productivity applications and pave the way for something far more important: a new operating system for the cloud.