Money is changing. Machine Learning is broken. China is winning foreign investment. Apple is becoming a spy and venture capital is exploding in all directions. Another That Was The week.

David Marcus heads up Facebook’s journey into digital wallets, stable coins, and payments infrastructure. The journey so far has taken him far from his starting point, but he has not lost his taste for a big vision that in my opinion would represent a net plus for most of us on the planet.

Having initially made the error of too closely aligning Facebook with the Libra Project and so alerting many national governments to the threat posed by a 2.85 billion-member super-organization with its own currency and global payments infrastructure, free of the need for banks and other custodians, Marcus has evolved the idea. Facebook is now simply a wallet provider (called the Novi wallet) and also an adopter of Diem (the token formally knows as Libra). He sits on the Diem Association Board too.

The Diem organization is taking the lead on software development, government relations, and regulatory issues.

Facebook as an adopter and payment provider will bootstrap and accelerate the adoption of Diem as its Novi wallet is issued to Facebook members in countries where regulatory approval permits.

In a major piece in Medium Marcus spells out that he is focused and patient but maintains his end game of a global, free, initially p2p and later merchant payments system using Diem as a stable coin.

His key statement is:

That end game is - for me - inevitable and generally good. The bad guys - from my point of view - are the myriad of middlemen who prey on those who want to manage and control their own money. And due to its cumbersome, analog, characteristics need to charge large sums for processing payments or holding funds, not to mention for loans or other debt. Interest paid on stored funds is pathetic compared to the possibilities with P2P loans currently being process on the Defi networks like Celsius and BlockFi.

A genuine global stable coin or coins would enable all elements of the human experience with money to be improved.

We have some other significant contributions this week. Cory Doctorow writes clearly and at length about the “rubbish in - rubbish out” problem of the data underlying machine learning. Data Analytics cannot be trusted unless the data and algorithms the analytics is founded upon can be trusted. Machine Learning is one step beyond analytics and will simply compound errors in the data it uses. So, how good is the data? Doctorow uses an example of duplicate voting analytics where the initial conclusions were over 90% inaccurate due to poor data. His concerns are entirely correct and are a major limit on much data-dependent automation. The fix is good data.

Edward Snowden writes about why Apple (or really anybody) can be trusted to place spying software on our personal devices, even in the aid of a good cause like the protection of children. He is right…

I included a long section on venture trends. These trends are fascinating to insiders. Fun fact of the week. Coatue has a 108% internal rate of return on its most recent fund and Series A pre-money valuations are at $37m. For those who care, venture capital’s evolution to a larger and more predictable asset class is well underway. In my role at SignalRank Corporation, I calculated this week that the 2021 cohort of unicorns has had over $136 billion of investments during the life of the 400 or so companies in the class (that includes companies that exited in 2021 above $1bn as well as private valuations over $1 bn). The investors at the angel, seed-stage would require $9 billion in pro-rata capital in order to maintain their equity stakes in those unicorns. The lion’s share of the value created is not being taken by the early investors. It is being transferred to large private equity investors from the B and C rounds onwards. This ecosystem is ripe for disruption. More in the video..