Amazon’s CTO created a meeting-summary app for some purpose
2 min readHow does Amazon CTO Werner Vogels – a person value hundreds of thousands who purchased out the tiny central Amsterdam Airbnb he was residing in throughout the COVID-19 pandemic – spends his days? By its appears to be like: Building AI-powered meeting-summary apps. go determine.
one in Post This week on Vogels’ private weblog he particulars the distillate open supply App he created along with his “OCTO” (Office of the CTO) group to transcribe and summarize their convention calls. Distill takes an audio recording of a gathering (in codecs like MP3, FLAC, and WAV), analyzes it, and produces a abstract with an inventory of to-do objects. It can optionally spit out that abstract and record to platforms like Slack by way of customized integration.
![to forge](https://techcrunch.com/wp-content/uploads/2024/05/Screenshot-2024-05-09-at-11.46.29 AM.png)
As one would possibly count on of an app from Amazon’s CTO, Distill clearly depends on paid Amazon services and products to do the computational heavy lifting. AWS Transcribe transcribes distills; Amazon S3 gives storage for assembly audio information; And Bedrock, Amazon’s generative AI growth suite, handles the summarization.
But why create assembly summaries when numerous instruments exist that may serve the aim? Well, I’ve to think about the Vogels thought, why not? After all, he has loads of assets and sufficient free time for hobbyist programming tasks. According to the weblog, he is already attempting his hand at porting Distill’s code from Python to Rust. Being a CTO is a superb job if you will get it.
One distinctive factor about Distill is that it permits you to select which AI mannequin summarizes the assembly. By default, that is the Sonet, a mid-range mannequin in Anthropic’s Cloud 3 household. (Amazon’s heavy stake in Anthropic could have one thing to do with that design determination.) But any mannequin hosted in Bedrock will work, like Meta’s Llama 3 and fashions from AI startups Mistral, AI21 Labs, and Foghere.
Vogels doesn’t promise that Distill won’t make errors.
“Remember, AI is not perfect,” he writes. “Some of the summaries that we get back… have errors that require manual adjustment. But that’s okay, because it still speeds up our processes. It’s just a reminder that we still need to be smart and engaged in the process. “Critical thinking is as important now as it ever was.”
I argue that having “includes” in summarization defeats the aim of the automated summarizer. You may also rent a stenographer. But you will by no means see Vogels badgering the know-how being offered by his employer. And I’d say that is why he is nonetheless the CTO.