- Whisper (and AI will still hear you)
- Flying around a tornado (VIDEO)
- Scared? Do this.
Today I remind everyone that I'm a geek and software is my comfort zone. 🤓
Thing 1 - Whisper (and AI will still hear you)
OpenAI just changed the speech recognition game.
Skip to "What to expect now" if you don't care about the nitty-gritty.
Setting the stage
Google Assistant, Alexa, Siri, Otter.ai... all recognize what you're saying pretty well. So... What gives?
You can't use those to create your own thing without
- some heavy roadblocks
- active internet connection
- privacy concerns
God forbid you try to make some money from all the effort you put in.
There must be other tools, Z, you've heard of open source before?
Sure, there's Sphinx, Kaldi, Julius, Firefox speech recognition, and others. Brew a cup of coffee and get some steel nerves if you're to set them up. And spend an eternity on customization to get more than meh results.
How is Whisper different?
- Runs on your computer - offline and private
- Takes <15 min to set up
- Works like a Swiss watch - impeccable
- Free as in speech
- Free as in beer
So you've tried it, Z?
Yes, I have. And the others I mentioned.
"Fix" a CUDA error with
--device cpu, if you get one.
- Hears better than I do
- Detects jargon words based on context
- Translates on the fly
- ...is not real-time... yet
I'm impressed and delighted that OpenAI published the code and the model under the MIT license. This means you have no restrictions on what you do with it, but don't go crying back if it eats your lunch.
What to expect now (THE NON-GEEK PART)
Everyone and their uncle just got the ability to
- put sound into words
- detect which language the words are in
- translate languages to English
This is a mega-transfer of power from a few giant companies to every hobbyist who wants to talk to their computer.
Every time this happens we get a plethora of innovations from directions you never expected.
If you want to create maximum progress, give people tools that help them do things they couldn't before. They'll find ways to use them in ways you haven't predicted.
Thing 2 - Flying around a tornado (VIDEO)
David Evans followed some hawks in his glider on a flight one day. (That's a thing you do when flying a glider)
As they circled in a thermal stream, a tornado formed in the center.
Here's what he saw through his window:
Thing 3 - Scared? Do this.
Unlike the cool-headed David, the rest of us experience fear every now and then.
What's more important than what happens to you in life?
How you react to what happened.
Fear serves a purpose. It sounds an alarm that something you value may be in danger.
The ideal response is not to dismiss the alert. It's to troubleshoot.
Emotional intelligence involves checking if the threat is real and choosing an effective reaction.
- Adam Grant
You can treat fear the same way you treat hunger. It's your body giving you some information.
It doesn't have to take over your cognitive process. You can set fear aside the same way you set aside hunger, and deal with it when it's appropriate.
- Alex Honnold
Alex must've mastered his reaction to fear. How else do you explain his free climbing adventures?
Note to self: Like most things in life, this is a skill that you can practice. Put yourself into situations where you feel fear. Stay relaxed. Manage your reaction. Be safe.
Till next week,
cheers, Zvonimir 💻 🤓