Thanks for really good thoughts. I think intuition is just sub-conscious analysis. That is, the older and more experience you get the better is your gut feeling as a guide in life. It gets more and more common for me to just trust my intuition. I used to think it was laziness that made me skip overthinking, but I have come to realize that using intuition ex ante often score the best ex-post
Great stuff Louie! I think sites that generate their own data like Stack Overflow and Google will still be okay because LLMs like GPT-3 and Chat GPT-3 are constrained by the knowledge they were trained with (which isn't beyond 2022). My long term prediction is that it's likelier sites like Stack Overflow will want to have its own GPT-3 model fine-tuned on its own data to offer a better user experience.
Lots of gems here Louie! My favorite: “[Procrastination] is a strong signal that we aren't working on the right thing or the thing we are working on is too ambiguous. It's crazy to ignore this signal and override it all the time.”
This resonated with me because when I was working my corporate job, I would procrastinate so much on the menial tasks that seemed to have to purpose. Working for myself though, I’m so excited to get up and tackle the next thing!
> 2. I changed my mind about trusting my intuition.
> 4. Rules of thumb can beat out complex trained models with access to tons of data.
Before I was educated, I used to think simple rules are enough.
Then I got educated, I was taught that simple rules are never enough. There's always something that meets the eye.
I believed in it and trusted nothing. Not even myself.
Now, I have survived yet another year working for myself many, many years after my last formal education stint, I now know that pure rules, no matter how complex or simple, are never enough.
The best combination is a good intuition honed by experience and some simple rules for my conscious left analytical brain to use or keep busy with.
I, too, have come full circle back to simple rules.
Aren't all simple rules susceptible to be automated by AI?
The factual know-what simple rules, yes. Now, even the complicated know-how simple rules too.
But the simple rule hardest to automate are those to do with timing and content. The Know-When kind of simple rules.
That's Art. And I'm not that sure that can ever be automated by machines.
Happy to hear we're not the only ones with this too-simple philosophy. We purchased all apartments that we're currently renting based on this rule.
> "But this year, I learned that procrastination is a signal. "
It is! Daniel also popularized this idea, and I started to look at procrastination as a signal, just as you did. I noticed that I'm abandoning side projects much faster without overthinking whether I should continue them. :D
Thanks for really good thoughts. I think intuition is just sub-conscious analysis. That is, the older and more experience you get the better is your gut feeling as a guide in life. It gets more and more common for me to just trust my intuition. I used to think it was laziness that made me skip overthinking, but I have come to realize that using intuition ex ante often score the best ex-post
Great stuff Louie! I think sites that generate their own data like Stack Overflow and Google will still be okay because LLMs like GPT-3 and Chat GPT-3 are constrained by the knowledge they were trained with (which isn't beyond 2022). My long term prediction is that it's likelier sites like Stack Overflow will want to have its own GPT-3 model fine-tuned on its own data to offer a better user experience.
I knew procrastination was about more than laziness.
But I've never heard anyone say ambiguity can be a key reason for it. Makes total sense!
Lots of gems here Louie! My favorite: “[Procrastination] is a strong signal that we aren't working on the right thing or the thing we are working on is too ambiguous. It's crazy to ignore this signal and override it all the time.”
This resonated with me because when I was working my corporate job, I would procrastinate so much on the menial tasks that seemed to have to purpose. Working for myself though, I’m so excited to get up and tackle the next thing!
Thank you for sharing!
I like the rule of thumb topic.
> 2. I changed my mind about trusting my intuition.
> 4. Rules of thumb can beat out complex trained models with access to tons of data.
Before I was educated, I used to think simple rules are enough.
Then I got educated, I was taught that simple rules are never enough. There's always something that meets the eye.
I believed in it and trusted nothing. Not even myself.
Now, I have survived yet another year working for myself many, many years after my last formal education stint, I now know that pure rules, no matter how complex or simple, are never enough.
The best combination is a good intuition honed by experience and some simple rules for my conscious left analytical brain to use or keep busy with.
I, too, have come full circle back to simple rules.
Aren't all simple rules susceptible to be automated by AI?
The factual know-what simple rules, yes. Now, even the complicated know-how simple rules too.
But the simple rule hardest to automate are those to do with timing and content. The Know-When kind of simple rules.
That's Art. And I'm not that sure that can ever be automated by machines.
I totally find myself in your last meme :D
Great stuff, Louie!
> "Would I live there myself?"
Happy to hear we're not the only ones with this too-simple philosophy. We purchased all apartments that we're currently renting based on this rule.
> "But this year, I learned that procrastination is a signal. "
It is! Daniel also popularized this idea, and I started to look at procrastination as a signal, just as you did. I noticed that I'm abandoning side projects much faster without overthinking whether I should continue them. :D
Finally getting to reading this. Your point about procrastination resonated well. I had forgotten about this and thanks for the timely reminder.