INTRO

It’s Wednesday, you know what the means — it’s time to do a deep dive into the life and tech stack of some of the coolest founders out there.

In today’s edition, we have Kwindla Hultman Kramer, founder of Daily.co, a real-time voice, video, and AI service for developers. Enjoying this newsletter? Don’t forget to forward it to your friends!

MEET THE MAKER

Kwindla

Hi! I’m Kwindla Hultman Kramer 👋 

I am one of the founders of Daily and I serve as the company's CEO. We develop infrastructure and SDKs for video and audio. If you are creating a product or an app that has video or audio features, we can probably help. The Internet is increasingly a video-first medium, and we think of ourselves as building the infrastructure for the future of our collective digital experience.

I have been interested in large-scale networked systems and real-time video since I was a graduate student with Mitchel Resnick at the MIT Media Lab.

Before Daily, I helped to found Oblong Industries. We built an operating system for spatial, multi-user, multi-screen, multi-device computing. Our must successful product was a collaborative big-data environment. 

Links:

• https://twitter.com/kwindla

• https://www.daily.co/

• https://www.linkedin.com/in/kwkramer/

Kwindla’s Stack

General communication internally at Daily.

Notion | Slack | Github | Linear | Superhuman | Daily

Looking at this list, it's interesting how complementary all of these are. They're all local maxima for what they do, at least for me or for us as a team. Probably every product on this list is familiar to people reading this except for Daily. We use our own WebRTC stack for all our video meetings. Eat your own dogfood.

Tools for writing code

VS Code | Emacs

I'm a long-timer user, near-religious devotee of Emacs. I have hundreds of lines of Emacs Lisp customizations I've built up over the years. I feel most productive, individually, in Emacs.

But ... the world has moved on. Most of my colleagues use VS Code. Using VS Code means that I probably have the right linter settings, etc, for collaborating on our shared codebases.

I still feel scratchy about this, and think I need to get an inspirational poster to put up on a wall that I can look at whenever I'm tempted to complain about VS Code: "If you want to go fast, go alone. If you want to go far, go together."

LLMs

GPT-4 | LLama3 | Groq

I'm using LLMs a lot, both as part of writing code and as increasingly important parts of our product stack. We care a lot about latency for most of what we do. So in addition to GPT-4, which is still the undisputed best general-purpose LLM, I'm also using Llama 3 more and more. Especially Llama 3 on Groq, which is blazingly fast.

Video editing

ffmpeg | Descript | Deepgram

I write low-level video code, so when I need to trim or resize a video I often just use the command line tool ffmpeg. When I first tried Descript, a year or so ago, though, it completely blew me away with its totally different, "transcript first," editing workflow. I'm a big fan of Descript and it's influencing a whole new generation of video tools. Whenever I need to do batch transcription I use Deepgram.

LLM engineering platform

Baserun

Everybody who writes LLM-related code is aware that we're going to need a new category of dev tools to support "AI Engineering." None of our old prototyping, CI/CD, or QA tools help much as soon as you have an LLM prompt or two in your codebase! Lately I've been using Baserun to test and iterate on prompts.