In this week’s episode, we consider the rapid progress and recent impressive demos in the realm of speech recognition technology. We consider the difference between transcribing and understanding language, and work out a thought experiment of what might change when full transcription is widely and cheaply available. We talk about the challenges facing current generation technologies and speculate which are likely to be improved soon and which are sticky. We wonder whether even machine-readable transcription might be enough to help search engines do things like jump you directly to a movie quote’s location in a film, or to help YouTube and Facebook mine your private videos for marketing purposes. We also cover the effects on lifelogging, surveillance, interface design — and of course, jobs. At the end of the episode, we debut a new listener mailbag feature and respond to your comments.
Relevant Links
- Review the Future 033: What’ll be the Impacts of Self Driving Cars?
- Review the Future 041: What are the Reasons to Protect Privacy?
- Strange Days on IMDb
- Transcriberly transcription service – machine and human
Other Recent Content by Review the Future
- LET GO: an indie sci fi graphic novel
- Guest Spot: Robot Overlordz podcast, episode 138
- Guest Spot: Futurology podcast episode 19