2024-02-05
I was listening to Fall of Civilizations podcast and they referenced the size of Lake Victoria, the source of the Nile river, in Africa as approximately the US State of Georgia.
It turns out this is incorrect. Lake Victoria is 23,146 square miles while the US State of Georgia is 57,513 square miles. However, it is about the same size as the Country of Georgia, which is 26,830 square miles or the state of West Virginia (24,038 square miles). Still incredibly massive, and it's not even the biggest fresh water lake! That honor goes to Lake Superior, which is 49,300 square miles, about the size of Alabama or Greece and holds 10% of the world's fresh water (!).
In either case, Lake Victoria has shot up my list of places to go. I'm sure I knew that it was the source of the Nile river at some point, but there's something romantic about being reminded about all that distance and all that history.
2024-02-04
I’ve been noticing a particular strong point in my use of LLMs: repackaging knowledge.
I’ll use a specific example here from my work at Macro Oceans: I need to understand the research around specific polysaccharides that can be extracted from seaweeds 1) so I can sell them and 2) so I can do further product development with them. To do this, I do a pretty basic lit review. I get a lot of help from GPT in this stage understanding things that are beyond the chemistry and/or biology that I took. But I actually need to distill this into something I understand and have an intuition for, so it's pretty essential for me to take the times to put this into my own words.
At the end of this process, I’ve effectively got a 1 page memo that explains how the polysaccharides work and what their benefits are. From there, I need to turn this into:
* 2 slides for the team meeting to help teach internally
* 4-5 slides for a sales training deck
* a blog post for content marketing purposes
* a paragraph to a specific customer for a deal where it's relevant
This is where I think the LLM really shines. The context is set and hallucinations are rare. Perhaps more important, I understand the content well enough to catch things that aren’t quite right. The important part is mutating the form. In tasks like this, I’m frequently able to cut half the time more out of the work compared to what it would take me starting from a blank page.
An implication of this is that the value of proofreading and comprehension as a skill is shooting up tremendously. I’ve always been impressed with the ability of some of the senior executives I’ve worked with to read a memo or a slide deck and immediately pick out the critical issues at hand (Neal Mohan, CEO of YouTube, is fantastic at this). The difference between an average manager and the top ones on this dimension is startling. The average managers fall into nodding along but the really top tier people are engaging with and testing the material they’re being given.
I think this comprehension skill is going to play the role in the next two decades that being a strong writer played in the previous two. Of course these two skills are closely related, but as more people move into a role where they’re reacting to more text than they’re creating, it will become even more important.
2024-02-02
A compelling case that plants and animals without brains have memories.
“The neuron is not a miracle cell,” says Stefano Mancuso, a University of Florence botanist who has written several books on plant intelligence. “It's a normal cell that is able to produce an electric signal. In plants almost every cell is able to do that.”
On one plant, the touch-me-not, feathery leaves normally fold and wilt when touched (a defense mechanism against being eaten), but when a team of scientists at the University of Western Australia and the University of Firenze in Italy conditioned the plant by jostling it throughout the day without harming it, it quickly learned to ignore the stimulus. Most remarkably, when the scientists left the plant alone for a month and then retested it, it remembered the experience.
Found via The Browser.
2024-02-02
Speculative from someone who knows the industry well:
* The changes Apple made to Ad Tracking have totally ruined the playbook for DTC beauty brands, who can no longer grow using targeted ads on Facebook (this is well known)
* A second order effect of this is that it’s making it harder for new ingredients to get adopted, because the DTC brands were usually the most willing to experiment here.
* This in turn slows the pace of innovation in the ingredients industry, especially towards cleaner / more sustainable / more natural alternatives.
I can't prove or disprove this, but it does make intuitive sense.