meta_pixel
Tapesearch Logo
Log in
Tech Brew Ride Home

Fri. 04/17 - Y Combinator Gets Picky

Tech Brew Ride Home

Amalgamated Internets, LLC

Tech News, News, Technology

4.71K Ratings

🗓️ 17 April 2020

⏱️ 20 minutes

🧾️ Download transcript

Summary

Microsoft has some AI tech that can catch bugs 99% of the time, Y Combinator is going to be more picky about who it invests in, macOS is getting a smarter battery management system for your laptop, and of course, the weekend longreads suggestions. Sponsors: DoubleUp.agency CognitoHQ.com Links: AI spots critical Microsoft security bugs 97% of the time (VentureBeat) Changing policy, Y Combinator cuts its pro rate stake and makes investments case-by-case (TechCrunch) Google's fast-growing Meet video tool getting Zoom-like layout, Gmail link (Reuters) Apple changes default MacBook charging behavior to improve battery health (Six Colors) Apple CEO Talks Covid-19 Crisis, Return to Work Plan at Company-Wide Meeting (Bloomberg) 'Needle in a haystack': Reborn tech offices may need distance and mass testing (Protocol) The coronavirus pandemic turned Folding@Home into an exaFLOP supercomputer (Ars Technica) Can Comic Books Survive the Coronavirus Era? (NYTimes) LIDAR: Peek Into The Future With iPad Pro (Halide) The Devastating Decline of a Brilliant Young Coder (Wired) In Half-Life’s improv scene, anyone can speak for Gordon Freeman (Ars Technica) Learn more about your ad choices. Visit megaphone.fm/adchoices

Transcript

Click on a timestamp to play from that location

0:00.0

Welcome to the tech meme ride home for Friday, April 17th, 2020. I'm Brian McCullough today.

0:09.7

Microsoft has some AI tech that can catch bugs 99% of the time it says.

0:15.0

Why Combinator is going to be more picky about who it invests in.

0:18.7

Mac OS is getting a smarter battery management system for your laptop and of course the weekend long read suggestions.

0:25.0

Here's what you missed today in the world of tech. Microsoft claims it can now use AI to distinguish between security and non-security

0:38.8

bugs 99% of the time and it can identify high priority security bugs 97% of the time. It plans to open

0:46.8

up the methodology behind all of this to make it possible for you to maybe do the same in coming months quoting venture beat the work

0:55.7

suggests that such a system which was trained on a data set of 13 million work items

1:00.5

and bugs from 47,000 developers at Microsoft stored across

1:05.0

Azure Devops and GitHub repositories could be used to support human experts.

1:10.8

Corologics estimates that developers create 70 bugs per thousand lines of code

1:16.2

and that fixing a bug takes 30 times longer than writing a line of code.

1:20.4

In the US, $113 billion is spent annually on identifying and fixing product defects.

1:26.0

Microsoft model leverages two techniques to make its bug predictions.

1:30.0

The first is a term frequency inverse document frequency algorithm, a T.F. I.

1:36.0

An information retrieval approach that assigns importance to a word based on the number of times it appears in a document

1:42.0

and checks how relevant the word is throughout

1:44.5

a collection of titles.

1:46.0

Microsoft says that its bug titles are generally very short containing around 10 words.

1:51.5

The second technique, a logistic regression model, uses a

1:55.0

logistic function to model the probability of a certain class or event existing.

1:59.7

Microsoft says that the model is deployed in production internally and that it is continually

...

Please login to see the full transcript.

Disclaimer: The podcast and artwork embedded on this page are from Amalgamated Internets, LLC, and are the property of its owner and not affiliated with or endorsed by Tapesearch.

Generated transcripts are the property of Amalgamated Internets, LLC and are distributed freely under the Fair Use doctrine. Transcripts generated by Tapesearch are not guaranteed to be accurate.

Copyright © Tapesearch 2026.