Welcome to my crib.

Overcast weather but a fun first ski tour.

2 Lab 2 Leak

So it looks like Omicron was (another) lab leak.

Key arguments:

Difference in the phylogenetic tree between Omicron and all the other variants. Natural selection typically occurs one mutation at a time. We would expect to see intermediate variants between the mid-2020 ancestor of Omicron, and Omicron itself. Here’s the money plot:
Not every mutation will be active. Mutation in the wild often leads to “silent” mutations—differences on the genome that don’t have a noticeable phenotypic effect. You can track the “normality-factor” of a variant by looking at the ratio of silent to active mutations. Almost all the mutations on this variant are active: implying that these mutations were artificially selected in the lab.
Current mainstream consensus is that Omicron jumped to humans from mice. We know lots of places where mice are given infections… they’re called labs.
The more you look for them, the more common lab leaks seem to be. The original SARS virus escaped from a lab multiple times. And Delta has already escaped from a lab in Taiwan (though by then Delta was already spreading worldwide).

Ok, so perhaps you agree Omicron was probably a lab leak. Given this new information, what should we expect?

Increased likelihood of lab leaks of other covid variants, and potentially other pathogens
Increased likelihood of cover-ups—since the OG Wuhan lab leak hasn’t even been fully acknowledged by the mainstream yet.
Due to 2, acceleration in mistrust in “experts” and institutions like the WHO, CDC, and Mainstream Consensus Opinion generally
(Potentially) a prolonging of the pandemic. If virulent variants with immune escape are leaked, then our existing vaccines will only confer limited protection against them. Only social distancing measures or new vaccines will stop the spread of the new variants.

What, I wonder, are the other consequences of Omicron being a lab leak?

Leave a comment

Three youtube creator marketing strategies

Cash money dollar dollar bills yall. A youtube sponsorship is when a brand pays a youtube creator to promote their product in an ad integrated within one of their videos.

At Wren, Youtube sponsorships on boring channels consistently perform better than sponsorships on entertaining channels. Even when you take into account that you have to pay more per view for more entertaining channels. Today I realised why.

Narrower audience targeting. People watching boring videos must be really interested in the content. If the content is in your niche, then it’s quite likely they could be a prospective customer.

The most-viewed youtube videos are appealing to almost everyone. This definitionally means a smaller proportion of the total viewership is part of your target audience. So on a per-view basis, they should perform worse. It doesn’t make sense to advertise on a Mr. Beast video unless you have an extremely broadly appealing product.

In fact, the most viewed videos should perform a lot worse, because they also have the most companies clamouring to sponsor them. In addition to companies like Skillshare and Brilliant who run performance based ads, they’ll have companies running brand advertising. Brand marketers can pay much more than direct response marketers, because they don’t need to worry about immediate return on investment.

Here are the three youtube sponsorship strategies that are working for us at Wren:

Sponsor boring videos in your target niche
Sponsor videos you think might go viral in your target niche
Get creators in your target niche to make custom dedicated videos very closely related to your product

1 works because of narrow targeting. 2 works because of underpricing: if you’re good at identifying potentially-viral videos, you can pay less for them than others. 3 is also a form of narrow targeting: only people interested in your product will watch the video. But you have to pay a lot for a creator to cede editorial control of their content, so making dedicated videos less likely to be successful than 1 or 2 if you can do them well.

Audience segmentation is the most challenging part of the job for creator marketing. Unlike with targeted ads like on Facebook, you can’t know the last thing a viewer bought, or how much money they’re making. You have less hard data to base your decisions on. Right now, I think of our target niche as a set of different interest-based and demographic factors that tilt people’s interest towards our product. Here’s an example:

The most interesting thing about this Venn diagram is how different the 2nd and 3rd order connections are. Trade union members or small biz owners are very unlikely to buy consumer climate products. But WEIRD people and the liberal intelligentsia both are.

You can sort and rank youtube channels and videos based off your chosen dimensions. Are people watching this content likely to be Left wing/educated/western/have money? Channels that meet more than half your demographic qualities are good candidates to sponsor.

Share

At Wren, in addition to climate change focused channels, we’ve found solid success with left/liberal cultural commentary channels (Left/Western/Educated), SpaceX news channels with liberal hosts (Western/Educated/Liberal), and Western science education channels (Western/Educated/climate related).

Soliciting book recommendations

I’ve been given a bountiful learning and development budget by my work, and I’m looking for help spending it. Here are some prompts:

What book have you read in the last few years that changed the way you see the world, but you wouldn’t necessarily recommend to anyone else?
Are there any books you’d like me to read and share my learnings from?

For my part, here are a few books I read last year that I recommend:

The Sports Gene: good book on nature vs. nurture in high level sports. Spoiler– it’s both.
The Courage to be Disliked: When Eddie Jones, Marc Andressen, and Shreyas Hariharan all recommend the same book, you know it’s gonna be a good read.
Draft No. 4: Good book on the writing process by a master. To write well you first have to write very poorly.
The Road to Wigan Pier: Amazing ethnography of miners by Orwell that also has an analysis of the mid-war British class system. Also, he hated processed food before it was cool. Short and very well written.
Sourdough by Robin Sloan: Amusing book about Bay Area bread baking.
Chaos Monkeys: Fear And (Self) Loathing In Big Tech Cos.

I don’t recommend the Contrarian by Max Chafkin. Terrible book.

Leave a comment

Between the cracks

#21 Books, youtube, and Omicron lab leak