May 21st Newsletter
A thought: Go confidently towards what you want, trust your heart to guide you.
I worked in colleges for a decade, colleges have a fundamental problem. College has been totally financialized, turned into a consumer product for purchase. For many, if not all, the expected return is a job.
It wasn’t always this way; in the early days of college, pre New Deal, college was a finishing school for the elite, it was a place for children of the wealthy to go play and get prepared for executive positions leading the country.
It changed to a tool of economic expansion as the liberal state expanded and since then it’s expanded to include a wide variety of premium services from housing to campus activities to student mental and emotional support (i’m not here to debate whether this needed to happen). Commensurate increases in fees followed to where college has become a huge investment for most families.
I believe, personally, that there’s still enormous value for most families.
What is not debatable is that with this kind of expenditure, college has become a quantity to be purchased and thus there is expectation of return.
I was thinking of this reading an article about students pushing back about professors trying to limit their ChatGPT usage.
I think most people on campus, on some level, realize the students are customers. The old saying is that the customer is always right.
Universities may succeed in pushing back against the ChatGPT’ing of education, but if the students aren’t willing to march along, the students/customers will likely win.
The tension is palpable.
As we’re watching people get fired across the government and the installation of AI, it’s worth asking what efficiency is.
The hair trigger answer is probably doing things faster.
The right question is what things are getting done faster.
If you do it faster and it’s wrong, that’s not efficient. If you do the wrong thing fast that’s bad too.
Efficiency isn’t faster, efficiency is solving the right problems for the right people. Faster isn’t better if it’s wrong, it’s just faster.
So efficiency is solving the right customer’s problems correctly and quickly.
I’ve experienced this quagmire talking to customer service people. Sometimes it seems like their goal is to get you out of their face or off the phone as quickly as possible but they aren’t all that focused on actually solving your specific problem quickly.
To solve the correct problem, you have to listen. You have to ask questions that are discovery minded.
Often people don’t know how to articulate what problem they have or they jump to a solution that makes sense to them. You may have an easier solution they don’t know about. You’ll never know unless you ask questions and listen and figure out what their actual problem is. Only when you understand the problem will you be able to sort out the right solution.
So, back to efficiency, is firing a bunch of people and replacing them with AI actually efficient?
Well, are people getting their problems solved? Time will tell on this one.
I see a lot of people on Bluesky screaming and yelling about AI.
I get it, it’s frustrating when new things come to market and tech tends to do a lot of “break things and see what happens”. Not here to comment on copyright infringement, not a laywer.
However, I think it’s dangerous to dismiss AI.
Also, AI is not google circa 2015.
If you’re using your AI to search like you do on Google, congrats, you’re old.
I get it, I’m old too.
Sam Altman from OpenAI was discussing recently that adults were using it more like Google and young people were using it as a life decision assistant. That’s just fundamentally different.
Let’s put aside whether you should be using AI as a life decision assistant, you should definitely be using it as an assistant.
The future of AI will be AI agents, where you’re telling the AI to run around and do tasks. People will manage multiple agents.
Here’s how you can start doing this now.
Go on your favorite AI, create one thread for things you do at work, more than one might be helpful, maybe based on particular functions of your work. Then just start discussing things in that same thread. Let it get better at that particular job.
Congrats, you’re now training your AI assistant. It’s going to get better, just like you’ll get better at teaching it.
This is the future. Might as well join up and stop whining.