67 Comments
User's avatar
Miranda's avatar

Thank for you writing this essay. After 5-months of relying on AI for creativity, problem-solving, chatting, I realized how much I couldn't handle human interaction, as humans are too slow to respond to me. I've used to AI's fast feedback and highly aligned what I'm thinking in mind. But meanwhile I gradually realized that I started to lose the ability to connect with people. When I chatted with a person, I started to analyze, their logic loophole, lack of structure, clarity and precision. This got me impatient. I noticed that I became more judgmental than before as I use AI cognitive ability as a benchmark to compare with human's. Until last week my date offered me a piece of feedback after I found his words were inconsistent. He said: "you're very analytical". I didn't know how to respond. I used to see analytical as a compliment in work. But on a date, it might mean I'm acting like a robotic scanner analyzing someone rather than connecting.

Anu's avatar

Thanks for sharing this story! The details are so human and so informative of the experiences people are silently having and how it’s changing them

Ondrej David's avatar

"I couldn't handle human interaction, as humans are too slow to respond to me"

You just describing talking to a very smart engineer. Or an average New Yorker. :)

Tom White's avatar

“A blank piece of paper is God’s way of telling us how hard it is to be God.” — Sidney Sheldon

Anu's avatar

phenomenal, stealing this for future fodder

David Shamash's avatar

“It seems 60% good so you start editing it, only to spend your time reconciling its argument instead of your own.“

This.

Anu's avatar

many such cases! learn from experience …

Samesh Lakhotia's avatar

thanks for such a great read anu!

more importantly, i appreciate everyone you attract in the comments section. i love engaging here until i find more real life 'intellectual sparring partners' haha :)

Anu's avatar

ah great compliment, thanks for reading! every essay is hopefully a conversation prompt

Louise Ankers/Pixel Sisterhood's avatar

This was brilliant. I shared it with two smart absolutely human friends ✅

Before I got to that exhortation!

Anu's avatar

i’d be concerned if you shared it with two nonhuman friends so this is a major win

Maurizio's avatar

PG says, "Instead of good writers, ok writers, and people who can't write, there will just be good writers and people who can't write. Is that so bad? Isn't it common for skills to disappear when technology makes them obsolete? There aren't many blacksmiths left, and it doesn't seem to be a problem."

So, what skill is exactly "disappearing"?! I don't get it... People that could and could not write always existed, I'd guess since writing was invented. Also, the lack of blacksmith is obviously a problem, but you only realize that when you need one.

Anu's avatar

I think the implication is that thinking is the underlying skill — what is the equivalent of going to the gym if you don’t do manual labor?

Maurizio's avatar

My point is that nothing has changed! Does he think all people were good thinkers pre-LLM?

On the gym analogy, I see it as people using dumbbells vs people using machines. Both build muscles.

Anu's avatar

You could argue everyone had to write essays in school or write emails at work but now it’s all doable by machine, so the muscle isn’t required to be flexed.

The dumbbells analogy then asks what is the dumbbells here versus just riding in a car or being carried around (not exerting yourself at all) …

Ondrej David's avatar

The middling middle writers will disappear. When AI can write as well as 85% of the writers (to the eye of an average reader) the market allows less of them in existence. Same in code. Same in design.

Jonas Valatkevičius's avatar

Great thoughts! Could easily restack half of your text :)

Anu's avatar

thank you! I’m most amused on behalf of @Substack to see that the word “restack” has caught on :)

Jonas Valatkevičius's avatar

Well, if the word could make us use less words, it usually sticks :) I’m working on a piece of using AI for creative work, I’ll direct to your text from there - you really touched on some important points

Anu's avatar

fair enough! look forward to it

QuestionClass's avatar

Maybe the better question isn’t how much we use AI, but how we approach it. The blank box can be a slot machine of endless suggestions or a mirror that forces us to face our own thinking. The difference comes from pausing to ask something worth wrestling with — a question that makes the box go quiet for a moment so you can hear yourself think.

Outside the Box's avatar

This made me burst out laughing at work - LLMs should ask if you're still thinking or just prompting into the abyss. (Or maybe they should ask how long it’s been since you’ve talked to a human.) --this was such a thoughtful and thought provoking read. Thank you.

Anu's avatar

I’m glad to bring some levity even with this topic, thanks for sharing!

Angela Hollowell's avatar

"Creative work requires communion with yourself first."

As someone who's been vibe coding and occasionally having AI assist me with the messy middle of writing (refinement and outlines) i very much appreciate this thoughtful article. I recently went through my Notion workspace and started to think intentionally on how I could make space for more thoughtfulness about how I use AI and that was super helpful for shaping what I hope will be a healthier relationship with it.

Anu's avatar

great anecdote - writing this from my own experience of introspection with tools as well

Sam Dix's avatar

Cy Twombly 🐐

Anu's avatar

the people love Cy!

Khyati Yadav's avatar

very very interesting take

Elin Häggberg's avatar

Loved reading this! I’m writing a series of essays on this same topic and this was very inspiring and added both confirmation and new input to my own thoughts on the subject! Thank you! 🙏🙌

Anu's avatar

Thanks for reading! glad it resonated

Ruchir Vyas's avatar

I love this. It reminded me of the time 2 years ago when I had to quickly write a YouTube script on-board a flight. GPT-3 had come out by then and obviously I could not access it on flight nor was it any good for writing back then. I was slightly scared that I’ll miss the deadline but on flight my brain worked magically. I wrote the entire outline on a paper napkin and almost half of the script and it was one of my best writing and thinking sessions.

Anu's avatar

Constraints are so good, sadly deadlines as constraints too. Technology takes it away but hopefully gives us new ones

Ruchir Vyas's avatar

Totally! And I have a hunch that thinking and writing without using LLM models will become a new trend soon.

Shriya's avatar

thank you so much for articulating so brilliantly the unease i have been feeling for sometime now!

Anu's avatar

Ty for reading! Glad it resonates even if with unease 😓

Justin's avatar

I appreciate the front and ends in addition to the middle part. And yet at the same time in other domains having the AI do the middle part can be helpful so that the human can focus more on the coming up with the idea and then the testing things out.

But I can definitely see how in writing the front and back end of things our areas that people are more likely to give AI a try while you determine how much of the actual middle middle process you’re wrestling and struggling with.

For a vibe coping in particular, I think the huge opportunity is that the middle middle part can be done decently , but the challenge for educators is to scaffold the process so that the students can learn bit by bit how well the middle middle was done. And thinking about creating useful things for people, than the front end of understanding the problem and problem framing for real life local needs, and then the back end of testing can be spent more time in,

Anu's avatar

for sure it’ll vary by task and what the learning goal is — also why a general purpose tool will have trouble catering to the right level of friction at the right steps!

Justin's avatar

That gives me the idea to have a friction knob on a web app 😅

F equals mu times N

The two ways of viewing the barbell is helpful for my educator world. Thank you again!!

Anu's avatar

look forward to seeing it in action!

Justin's avatar

it won’t be heavy yet, but at least an instantiated form of an idea 😅