Training Computers to See
Over the past few decades, computer vision has held the promise of making the world a better place, from aiding the blind to helping doctors better analyze medical imagery. But as it turns out, teaching computers to see has some unintended consequences. Joseph Redmon, a researcher at the University of Washington and computer vision researcher, tells the story of the history of this quickly evolving technology, as well as his own experience seeing a something he built be put to uses he’d never envisioned — applications that might, quite literally, be used to kill.
Find out more at rawdatapodcast.com
More episodes from Raw Data
Technically Sweet
Chocolate is beloved by...well, most humans, it would seem. But this sweet treat that, for many of us, brings instant happiness, has a nasty secret: …
The Worst Cyber Attack (For now)
In June 2017, something weird — and very alarming — started happening at a company in Copenhagen. It seemed that hackers had shut down the company’s network, and were demanding a ransom. …
How We Found Ourselves
Get out your smartphone, and you can almost instantaneously know where you are — and find out how to get where you want to go. Which, when you think back on the history of human navigation …
About Us, But Not For Us (Repost)
Originally broadcast in April 2019. As we approach the end of 2019, the Financial Times recognizes Shoshana Zuboff's "The Age of Surveillance Capitalism" as one of the best business books …
Prediction (BYTE)
When humans predict something, it’s basically an educated guess, based on our experiences. When a machine makes a prediction, it uses data and math. And we are increasingly relying on …
The Digital Divide
Imagine a life without the Internet. No email, no Instagram, no texting, no Google maps, no Netflix...what would you do? A “normal” life would be …
How you can listen to this podcast
You can listen to episodes right here on the website, or if you prefer, in a podcast app. Listening in an app makes it easier to keep track of what you’ve already heard, listen without using your data plan and many other conveniences.