If people love to watch novel use cases of dashboard software like Grafana etc, I highly recommend this video about it being used at a water treatment plant: https://www.youtube.com/watch?v=wucMZ9tb1I0
We used to (and still do) have things that could run commands and interpret them. These things would sometimes forget key parts to run or even forget to run them at all. So we invented a system where you could give instructions (code) and schedule when they would be run (cron etc). Those things were called humans.
There is a great article called "Manual Work is a Bug" [0]. The idea is that you have humans doing a lot of random things so you should:
- first make a list of the things they are doing
- then update the list with the commands they have to run for each step
- some of the steps won't have commands b/c it's things like "ask Bob what the limit should be"
- over time, the commands become scripts
- then the "ask Bob" becomes an API call
- one day, the whole thing is an automated system that runs code
People like to think that LLMs can do all of the above. I don't get this b/c code is deterministic and can be run repeatedly basically "for free" (at least compared to token spend).
I do think that LLMs can greatly accelerate the creation of the code/system etc and can also help with maintaining it but the whole "we will just version control the prompt" was clearly hogwash.
While watching it, I thought about AI generated content.
I have never personally met anyone who worked on any of the Toy Story movies. I know, from documentaries etc, who Brad Bird and John Lasseter are. I've also watched the video [0] about how Toy Story 2 almost got deleted (which I highly recommend if you are in storage, DevOps or SRE).
There are other movies like the Wild Robot that:
- had big emotional impact on me (b/c I'm a parent)
- are 100% animated
- I have zero idea who made them
I say all of this b/c at some point, most people don't care if the movie was hand painted cels, CGI done by humans or fully AI generated with human text prompting. If the feeling is strong, people will have a "bigger" reaction which in turn will make it more memorable. It's all basically on a spectrum of "humans using tools". They care about how the movie makes them feel. It started with humans spitting charcoal at a wall [1] and now it's linear algebra.
Apparently in the military, special forces in particular, this is even worse.
If you have a terrible soldier in a Tier 1 unit and you are trying to get rid of them via transfer to another unit, when asked why he is transferring you can just say "sorry, that's classified". The receiving unit then has no idea of the problems associated with that soldier.
There used to be great writing in magazines and people paid for those.
In fact, Playboy is famous for having excellent authors write articles or do serials. This is my favorite example of "you can be 'good' at two wildly different disciplines"
Atul Gawande, of Checkslist Manifesto [0] fame, has a great article about the Cheesecake Factory [1].
He mentions that that they have a model that can predict both guest numbers and revenue:
“We have forecasting models based on historical data—the trend of the past six weeks and also the trend of the previous year,” Gordon told me. “The predictability of the business has become astounding.” The company has even learned how to make adjustments for the weather or for scheduled events like playoff games that keep people at home."
And this was in 2012!
I highly recommend both the book and the article as both talk, in detail, about how to build systems even in very dynamic environments.
I can only imagine how accurate the models have gotten now.
I'm always fascinated by these tricks of game theory.
e.g. in business school, the dean of the undergraduate school had this story:
"When I was a practicing lawyer working on wills and estates, people would often ask me to cut someone completely out of their will.
I would always say that a better option was to write something like 'To my daughter Susan, I leave $1,000. She always said that she wanted to be financially independent from me so this is an amount to show her I lover her.'
Clients would always think this would send the wrong message and I would replay:
'No, no. If Susan fights the will and says she should have gotten more, the judge will say: but she clearly left you something and pointed out that she loved you AND took your wishes into account' "
I wish there was a book or collection of these types of tricks to study.
I was born with no sense of smell [0] and I always wondered if I could combine that with my tech skills to be CTO at a place like the screwworm factory or possibly Waste Management.
One of my favorite recently learned facts about Congress:
Federally mandated parental leave (paternity and maternity leave) polls at about 80% in favor with the US adult population. This is regardless of political affiliation, by the way. Democrat and Republican voters both support it.
Upon reading this, you might be surprised as to why it's NOT federally mandated given how popular it is.
One group it's NOT popular with is corporations. And corporations donate a lot of money to politicians. And it's cheaper to donate to politicians who are against parental leave than it is to pay people for that parental leave.
I enjoy sharing this b/c it's a reminder that there are groups who spend a lot of time and money to get their way. At first, that might feel overwhelming. You might be surprised to know that when you call your local congressperson, those calls gets tallied b/c they want to know what their constituents care about. So give them a call and let them know.
So I was an intern at Merck MANY years ago and they had this interesting comparison.
Most companies only publish medical research findings on blockbuster drugs once both are true:
1. Production has started
2. A patent has been filed
The reason for this is that they want to maximize the amount of production time under patent b/c that maximizes revenue.
If you are the researcher, that means you have to wait until all of the production setup is ready to go.
Merck took a different stance.
There, the patent was filed as soon as the researcher was ready to publish. This meant that there was less time under patent for production but was much better for the researcher as they got their findings out earlier.
The thinking was that being able to publish earlier would attract better researchers and in turn would lead to better drugs, more revenue, more profits.
This was in the late 1990s so not sure how this plan worked out as I haven't been in pharma since that era.
Would be interesting to hear from other folks more knowledgeable.