Good read. Tried something similar to the first workflow before though it was more to get better results since better results require better prompts. Will try the rest of the workflows and see how it goes. keep up the good work ruben!
Great post! A question/reflection - surely the MIT study is inherently flawed in 2 ways. 1) measuring how smart/dumb someone [or an activity] is by measuring brain activity surely isn’t correct. If leveraging GPT allows you to achieve the same output while using far less of your brain that would in many circles be seen as optimal in the same way using a calculator is instead of mental math?
2) seeing MIT using sentence recall as a way to measure learning is a massive misstep - this is arguably the key vector that is wrong with the current education system. Exact recall is a very poor indicator of learning or indeed level of intelligence.
Ruben, I'm happy you came to Substack. My friend Michael Simmons told me about you. I think you know him. Subscribed, and getting a lot out of this newsletter! Keep up the great work.
You did a great job summarizing the research and providing practical applications. I enjoyed he article and found you offered wonderful insight. I'm currently tracking research that examines AI-Human co-creation. I'm about to post on about Resumes and AI.
Very interetsing. I've found the "Draft, then ask" workflow to be critical to getting the most out of any LLM that I'm using. Unless you and the technology are mutally aligned in understanding your context and goals you will not get the results you're looking for.
My impression is that most people go straight to "give me an output" without taking the time to make sure they even understand what they want!
Thanks also for sharing your conversation, it's fascinating to see how others interact with ChatGPT
Great article. As an AI user myself, I also had to contend with my self-concept, including ethical practices. And yes, I realized that working with AI is different from asking AI to work for me. I look at AI as my online coach, giving me marginal feedback on my drafts...at the end, I take control of what I would consider my final draft.
Great article with regorous flow of how to take full advantage of AI instead of letting AI make us dumb. Many thanks Ruben for your effort <3
It was a different kind of blog.
I enjoyed it!
I wouldn’t say it’s made me dumb, but it’s made me numb somehow.
How come?
Because I don’t have to think or feel emotion. I just ask AI
It just made me impatient, especially if I want a quick fix/answer to something. No matter how personalised my input was.
Good read. Tried something similar to the first workflow before though it was more to get better results since better results require better prompts. Will try the rest of the workflows and see how it goes. keep up the good work ruben!
Good prompts + good understanding of how LLM works is the only requirement. But people tend to forget knowing how LLMs work.
Do you advise loading final copy into ChatGPT so it learns more about your tone, editorial decisions etc? If so what is the best way to do this?
That's a great idea indeed.
Especially for the next time you need something similar.
It's best to do it before than after. But after works (a little).
Nice one, Ruben.
Thanks!!
Great post! A question/reflection - surely the MIT study is inherently flawed in 2 ways. 1) measuring how smart/dumb someone [or an activity] is by measuring brain activity surely isn’t correct. If leveraging GPT allows you to achieve the same output while using far less of your brain that would in many circles be seen as optimal in the same way using a calculator is instead of mental math?
2) seeing MIT using sentence recall as a way to measure learning is a massive misstep - this is arguably the key vector that is wrong with the current education system. Exact recall is a very poor indicator of learning or indeed level of intelligence.
I agree to both.
The brain still needs a bit of “hard work exercice”, that's why the task is just an essay (and could have been much harder).
Ruben, I'm happy you came to Substack. My friend Michael Simmons told me about you. I think you know him. Subscribed, and getting a lot out of this newsletter! Keep up the great work.
Glad to meet you, Tom!
I've checked your work too, and I've been enjoying it a lot.
I'll follow along.
https://www.linkedin.com/posts/nataliekosmina_after-our-preprint-which-we-ourselves-originally-activity-7342548168150048768-43Z8?utm_medium=ios_app&rcm=ACoAAAEqRQcBmR4RIFeNYNTN2NsFTUtkwR5vaqo&utm_source=social_share_send&utm_campaign=copy_link
Yes?
Recent post from the author of the MIT study which I found interesting. Is the link accessible for you?
I should have shared it, sorry.
I'll edit my blog once I'm back in front of a computer.
Sometimes everyone feels like this. The effects of AI are still unclear and we are all, learning as we go, for sure 😃
Feels dumb? Aha
I wouldn't say it made me dumb. But kind of impatient.
You made me see it for what it really is.
Love how useful this write up is. I guess AI sometimes is just training us to be more human :)
You did a great job summarizing the research and providing practical applications. I enjoyed he article and found you offered wonderful insight. I'm currently tracking research that examines AI-Human co-creation. I'm about to post on about Resumes and AI.
Very interetsing. I've found the "Draft, then ask" workflow to be critical to getting the most out of any LLM that I'm using. Unless you and the technology are mutally aligned in understanding your context and goals you will not get the results you're looking for.
My impression is that most people go straight to "give me an output" without taking the time to make sure they even understand what they want!
Thanks also for sharing your conversation, it's fascinating to see how others interact with ChatGPT
It is about how you use it, I too follow the similar approach, Plan my approach first, then get it evaluated by AI, then work with AI to get it done.
Great article. As an AI user myself, I also had to contend with my self-concept, including ethical practices. And yes, I realized that working with AI is different from asking AI to work for me. I look at AI as my online coach, giving me marginal feedback on my drafts...at the end, I take control of what I would consider my final draft.