That’s a deceptively easy and tempting explanation to reach for. But I would argue for many of the knowledge workers it’s something else entirely: we are sick of everyone blabbering when we just want to be told “use that special argument in that function in the library X to solve your problem”.

I have struggled with ADHD for a number of years before I finally started beating it, and when I started truly pushing it back, I realized something (that is depressing for many): for a lot of people ADHD comes to existence because they don’t want to listen to most people around them (and to most on the internet).

I like books. A lot. But with time technical books became harder and harder to read because it happened no less than 15 times to me that the true insights started at 65% of the book and lasted to about 80%. The rest was either a setup (i.e. “why do we need this?”) or meaningless story-telling (“I had this problem when I worked for Oracle back in 1998…”).

Nuance matters. It’s definitely not all to be attributed to the good old “kids these days have no attention span” trope. Most people simply don’t have interesting things to say, or write bad books.

I am beginning to think I’ll start asking LLMs to summarize chapters after I finish them and observe if I agree with them. Most of the time I am after concrete knowledge. Deep technical reading is very enjoyable and welcoming… but not in this phase of my life and career. I need answers on how to advance with A or B, not a plethora of new paradigmae and philosophies.

Case in point: I need to start memorizing various PostgreSQL incantations because I periodically need them… I don’t need to be told that SQL sucks and that humanity missed out on something great by choosing it. Cool, very nice, but I still have work to deliver tomorrow.

People who don’t want to think just found one more huge excuse not to think. Nothing new.

LLMs objectively helped me by accelerating learning and in overcoming the blank canvas terror.

Honestly I’m quite lost right now, but I just cant stand books anymore when I started programming about a decade now books were amazing, but that’s in part because I knew almost nothing, so reading was exciting. Now that I am at a decent pace, I just find alot of books to have fluff in them and its not because I know everything in said book, if I did I wouldnt pick them up, its just sometimes I don’t have patience to traverse the chapters, when I really just want something to help me build intuition on the topic, rather than hold my hand. I only want to read books that get to the point

Right now, I just use LLMs mainly chatgpt & claude, mostly the former. Tbh I haven’t shopped around for other models or local LLM’s, but llms just strike the perfect balance for me as I can get it to explain things without the fluff, ofc the downside is you have to infer whether the LLm response is a complete lie and as we know with elixir LLMs only have information about the most popular topics, so you still end up looking through other materials like docs etc, you also don’t get the authors personal anecdote like you get in a book

I largely avoid tutorial videos , only watch conference talks or podcasts. So for me right now it is

1. llms
2. books
3. videos

I do wonder if there is a good LLM that can interact with a book, anyone know?

No, I have not changed. I can only learn be doing stuff myself, and the stuff needs to be a problem I genuinely need to solve. I can never follow tutorials;. I read some just so that I know what are possible.

Ive actually stumbled upon this topic before and I have tried detox, for me it usually takes 1 week from quitting social media, to start seeing the benefits. Also another good method is to find the best study time and length, for me its early in the morning and I can sit & study for ~5 hours pomedoro style. I can only do this on the weekends though because of the day job

But I dont think thats my issue, my main issues with STEM books is they feel a bit like a facade/selling you a dream, reading from cover to cover wont get you ahead, its the experience of trying things out and failing. I just wish every book had like a no nonsense summary booklet or something, that I could quickly read over and then try infer information.

I would love an LLM made for books, that can summarise and give me keynotes. I saw something earlier about an LLM making up bite sized court notes for people on various topics and I tried it with BEAM vm, it was great at making questions which felt like I was actually learning something, it also had various styles of questions from crosswords, to pick & place etc