AI researchers warn against delegating creative processes to automated tools.
Highly recommended reading.
Don't offload your learning. Don't offload your creative process.
"You can offload a task, or even a job, but you can never offload your learning."
http://x.com/i/article/2065582894790365184
Many users endorsed Satya Nadella's warning against offloading learning and creative processes to AI because retaining personal understanding and judgment builds lasting skill and control.
what to learn in this AI era tho, that is the hard part for a lot of people. There are so many avenues now opened with AI that it is hard to have a clear vision.
I think we'll soon seen an increase in "roadmaps" and "how to become ____" PDFs soon, people will choose what they want to be, like they've done before. If I wanted to be a frontend engineer, I could've looked up 1000 roadmaps.
Now we'll get a new set of roadmaps but it will be even more useful in all this info cluter.
@omarsar0 Offloading the task is fine, offloading the judgment that evaluates the output is what quietly makes you dependent. The people who get the most from AI are the ones who can tell immediately when output is wrong, which requires actually knowing the domain.
"You can offload a task, or even a job, but you can never offload your learning."
http://x.com/i/article/2065582894790365184
@omarsar0 this is the line i keep coming back to. i can offload the typing, the boilerplate, the first draft. the understanding of why it works is the one thing that doesn't transfer. offload that and you're renting a capability you can't debug when it breaks.
@MartinisLeonard We are solving this here: https://academy.dair.ai/
Lots more to build still, but I agree with the comments. We want to build a learning platform that personalizes to your needs, skills, and interests.
@omarsar0 well said ...
@omarsar0 the quote hits hard but "highly recommended reading" without naming the book is criminal
whats the rec?
@omarsar0 and never, ever offload your truth - document it, control it, recall it
https://github.com/tcballard/requirements-as-code
@omarsar0 Offloading tasks is fine. Offloading understanding is where you lose control of your craft
@omarsar0 the hard part is knowing the difference between offloading execution vs understanding
most people treat both the same
@omarsar0 Building this @liquicent
@omarsar0 @omarsar0 here's the other side of the ledger.
what do you think?
@omarsar0 The real compounding edge isn't data access — it's the reasoning scaffold you build by doing the work yourself. Offloading that is just renting someone else's map.
@omarsar0 i mean it doesn't make sense bro
@omarsar0 Agreed
@omarsar0 The line that matters most here is that "institutional memory should be queryable, not just stored".
Most firms treat knowledge as an archive. The ones that win will treat it as a living system that compounds, every decision making the next one sharper.
@omarsar0 the counterpoint people miss is that offloading *part* of the work changes what you learn you learn to verify and evaluate instead of build from scratch, which is a different skill
@omarsar0 Strong agree.
If you outsource the thinking, you outsource the growth. AI should compress the boring parts so you can wrestle with the hard ideas longer.
@omarsar0 What does off loading learning mean? Here
@omarsar0 the offload-tasks-not-learning line is the whole thing - the people speedrunning $20-to-$40k threads are offloading exactly the part you can’t get back