Zag To The Zig #10 :: Stablecoins, health data and more
Hope your week started well. I met some old friends from London last night & realised how important those moments are. #meetfriendsmoreoften
In the Tech Room
Now you know me, now you don’t. Turns out that those claims that AI can reliably detect emotion in facial expressions are not so justified. Just yet, I assume. I wouldn’t bet against this in the long term.
Then again, it seems it’s not just the blockchains bearing the martyr’s cross of being energy-guzzlers. The computer power required for deep learning research has been doubling every few months and yes, that comes with a large carbon footprint.
One thing that does keep dropping is the price of lithium batteries. This report estimates that electric vehicles will reach a similar price point as their combustion counter parts by the mid 2020s. Quick enough?
After experimenting with trading real-world assets on blockchains for the last 3 years, the fintech company Funderbeam is moving away from using blockchain. Too early, they say. Also, they feel trapped between the ‘believers’ and the more traditional old world. This gap will take time to close.
In the Economy room
On the other hand, the believers in (generally Ethereum-based) Decentralised Finance are jumping up and down about new developments on a weekly basis. They use the emoji 🔥 a lot. ;-)
This week MakerDAO will upgrade their intriguing Collateral Debt Position app (use crypto as collateral to generate DAI - now that’s 🔥🔥). So it might be worth reading this interesting interview with their COO Steven Becker. What struck me is that he didn’t really believe in Bitcoin or Ether, but got convinced by the potential of DAI as a decentralised stablecoin, as a bridge between legacy finance and crypto.This is something that can be a currency, an economic engine, and can scale. Central banks, commercial banks, community banks—DAI is something they can actually use.
In the ethics corridor
Which made me think of this story. PayPal has stopped payouts to performers on the PornHub platform. That’s apparently more than 100.000 people. Bitcoin won’t fix that problem (volatility), but something like DAI just might.
If you use old texts and books to finetune language-based AI, it will inherit its baked-in bias. Sounds logical when you put it like that. Sounds worrying, if you read examples like services failed to recognize the word ‘hers’ as a pronoun, though they correctly identified ‘his’.
More stuff on privacy vs data - this week in the health sector. First a Google whistleblower revealed Google is using personal medical data of 50 million Americans. And then the FT analyzed how ‘normal’ health websites include trackers to send usage data to whoever wants it. Sure, Amazon Marketing will happily use your menstrual cycle data.
On the staircase of Innovation
‘Empathy’ is often cited as one of the main skills to have in this increasingly digital world. Turns out that in Denmark (generally voted to be one of the happiest countries in the world), it’s been a compulsory class in schools since the 1990s.
In the Tool shed
As said before, I’m organising my thoughts on the link between business, ethics and technology. So I bookmarked this Framework for Businesses To Build a Better World, put together by B The Change.
Related, is anyone going to their Open Office moment in Amsterdam on Nov 26? Do let me know. ☎️✉️
Random ZTTZ
When a Dutch online bike company found out that a lot of their customers received damaged boxes, they didn’t make the packaging stronger, heavier or plaster more red warning stickers on it. No, they simply made it look like it contained a big, expensive, fragile tv.
🏁 End note: 1 thing I’ll be doing this week
I’ll be doing some ghost-writing and ghost-keynote-making for some secret design thinking research project. 🤫🤫