How does Amazon CTO Werner Vogels — a man worth untold millions, who during the COVID-19 pandemic outright bought the small Central Amsterdam Airbnb he'd been living in — spend his days? From the looks of it: building AI-powered meeting-summarizing apps. Go figure.

According to a blog post this week by Vogels, he and his "OCTO" (Office of the CTO) team created an open source app called Distill to transcript and summarize their conference calls. Distill takes a recording of a meeting in formats such as MP3, FLAC and WAV, analyzes it then generates a summary and a list of to-do items. It can optionally spit that summary and list out to platforms like Slack via custom integrations. One might expect the CTO of Amazon to design an application which takes heavy computation lifting from paid Amazon products and services. It is for this reason that AWS Transcribe carries out the transcription performed by Distill, while Amazon S3 provides storage for meeting audio files, and Bedrock, a development suite for generative AI from Amazon, would do summarizing.
But what is the point of writing a meeting summarizer when there are countless tools out there that would do the trick? Well, I have to think Vogels must have thought, why not? He has tons of resources at his disposal and, seemingly enough spare time for hobbyist programming projects. According to the blog, he's already trying his hand at porting Distill's codebase from Python to Rust. (Being the CTO is nice work if you can get it.)

One thing that makes Distill unique is you can specify which model of AI does the meeting summarizing. By default, it's Sonnet-a midrange model in the Claude 3 family by Anthropic. (Amazon's large stake in Anthropic might have had something to do with that design decision.) Any model hosted in Bedrock, however, should work, including Meta's Llama 3 and models from AI startups Mistral AI, AI21 Labs, and Cohere.

And by extension, Vogels further cautions that Distill does not guarantee not to err along the line.

"So long as I have AI to do the grunt work, sometimes those summaries we get back … that are full of errors and need manual alteration. Okay? Because it still speeds up our processes. It's simply a reminder that we must still be discerning and involved in the process. Critical thinking is as important now as it has ever been," he writes.

I'd say that in order to be "involved" in summarizing kind of defeats the purpose of an automatic summarizer. One may as well retain a stenographer. You'll never catch Vogel badmouthing the tech his employer's selling. And that, I'd wager to say, is why he's still CTO.