Predictions | Notion

<aside> 🤪 This is a placeholder document. I like the idea of sharing predictions, though I don't have much to write now.

</aside>

Blurring distinction between (different types of) live and recorded audio

The lines between these will blur. Live radio, late or delayed radio, voicemails, podcasts, live stream, notifications (”you have a new message from Alex”, “rain is forecast in 30 minutes”), AI-generated content (”The Spurs have just beaten the Suns 99-98 in…”, headlines read to you blending in with other notifications). 📻
Already (2022)…
- A YouTube video playing in the background is a lot like listening to an audio “podcast”. And “podacst” is increasingly going to be used to include videos also (see Spotify’s changes).
- A gated “podcast” to just a few friends is a bit like a voicemail in a group chat.
- A written article read by a text-to-speech engine is a bit like a “podcast episode”.
- A recorded Clubhouse or Twitter space is a lot like a “podcast episode”.
- As such, labels like “podcast” are less distinct and less helpful than before.

Soon: The end consumer need not even know whether you wrote or spoke the message; they can choose to consume voice or text as suits them.

Right now, your decision to text or send a voice message to a person is determined by:
- What you're doing
  - Cooking? Driving? You might use your voice.
  - In a meeting? You'll surely text.
- What you think they're doing
  - Are they driving or cycling? Better not send them a long text.
  - They're in a meeting? Send a text; they won't listen to a voicemail.
Transcription (speed-to-text) is now accurate and cheap. Speech-to-text is also now cheap, and finally human-like. Soon everybody will have the option to have text-to-speech in their own voice, that really sounds like them, for approx $0. As these both continue to improve, you (the creator/sender) can focus on what you're doing, and not worry about what the recipient doing.
- Do they hate audio, find it time-consuming, and prefer to read? No problem, record a voicemail and trust the transcription.
- Are they driving? No problem, type your text in your meeting, and it'll be read aloud to them (maybe in your voice, thanks to AI voice models).

As the cost of collecting more health data falls, this will have profound effects on privacy and new business models will ermege.

You won't use any and every toilet, because you'll be nervous they'll analyze your stool and sell your data learn things you don't.
You won't casually toss your empty coffee cup in a public bin, for the same reason: you'll be nervous about people analysing your saliva.
In 2021, you choose to get your own genome sequenced. But you can't choose to not be on Facebook. Even if you choose to not have a Facebook account, your friends might still upload pictures of you, and FB will create a ghost profile for you. Even if you never opt in to taking a photo and uploading to Facebook, Facebook knows who you are and what you look like. In future, it will be easy for other people to know your biometric data. Even if you never opt in to 23andMe, others will go ahead and sequence your genome.
In the meantime, while biometric data is considered rare and valuable, there will be a new category of products and business models. For example, a gym that's free to attend, but the Terms of Service allow them to collect detailed data (e.g., analyzing your heart rate and sweat)
Sperm donors will have a hard time remaining anonymous. Similarly for foster parents. Services will spring up to monetize genetic data, to sell insurance and differentiate pricing. Laws will be invented to counteract this.