I suspect it’s not even a time-saver: reviewing AI code requires building the same mental models I would have built while writing it, but without the creative satisfaction. I’m starting to think “doing it yourself” is the more sustainable way to stay in the zone.
I’ve realized I enjoy programming less when I rely too much on AI. It’s great for skipping boilerplate, but delegating the core logic kills the flow state that made me love the craft.
Screenshot highlighting the quote of Alan Turing: „One might be tempted to define thinking as consisting of "those mental processes that we don't understand." If this is right, then to make a thinking machine is to make one which does interesting things without our really understanding quite how it is done.“
About the quote, when googling it it’s often cited without specifying a source. It’s from a 1952 BBC radio broadcast titled “Can Automatic Calculating Machines Be Said to Think?“. The transcript is here: turingarchive.kings.cam.ac.uk/publications...
The quote is on p. 19
„One might be tempted to define thinking as consisting of "those mental processes that we don't understand." If this is right, then to make a thinking machine is to make one which does interesting things without our really understanding quite how it is done“ - Alan Turing
LLMs might be that machine
I know that feeling Karpathy describes. When I’m overwhelmed or confused, that’s usually when there’s actually something to learn. Same with reading—it’s not passive. You need to wrestle with it.
This session w/ Dwarkesh & Andy Matuschak shows that really well: youtu.be/OFuu4pesKf0
Source: x.com/karpathy/sta...
Karpathy on learning: “Learning is not supposed to be fun. […] the primary feeling should be that of effort. […] You want the mental equivalent of sweating. It's not that the quickie doesn't do anything, it's just that it is wildly suboptimal if you actually care to learn.”
In the book „The Scaling Era: An Oral History of AI“ by Dwarkesh Patel, Carl Shulman says we’re “growing” AI. Not building—growing. That one word hints at a shift: from making and designing to something more emergent.
Experienced devs: I stopped using Copilot — it made me forget basics like filter and reduce, and reviewing LLM code often means duplication, broken patterns, and needless complexity.
I now just use LLMs in the browser for learning.
Anyone else feel this? Or am I missing out?
A screenshot from Twitter. The original tweet reads: "Rapper Kenderick Lamar wore size 29 women's jeans to Super Bowl LIX." The tweet garnered over a thousand replies, most negative. I (Derek Guy) screenshotted that tweet and wrote: "so funny to see history repeat itself. men who wear slim fit and stretch denim—both of which were once considered strictly for women—criticizing kendrick for wearing flares bc they were on the women's aisle. in ten years, you'll be in flares clutching pearls about something else" Then another user replied to my tweet, saying: "Wearing women’s clothes has never been in style, homo."
Not true. I'll give you some examples of when styles crossed over from womenswear to menswear, and how men have worn straight-up womenswear or just feminine styles in cool ways. 🧵
"I will, in fact, claim that the difference between a bad programmer and a good one is whether he considers his code or his data structures more important. Bad programmers worry about the code. Good programmers worry about data structures and their relationships".
-- Linus Torvalds
Ridley Scott's legendary 1984 Superbowl commercial for Apple.
Elon Musk speaking to the audience at Germany's far-right AfD rally.
1984 / 2025
One programming bitter lesson that Factorio really drills into you is that efficiency and scalability and fundamentally at odds. Maximum efficiency exploits the local structure of a context, and that structure changes when you make the context bigger
What does it mean for something to be Turing complete?
I answer this question, and more, through a series of fully interactive Turing machine simulations! Play, pause, step forwards and backwards, and even write your own Turing machine programs in my latest blog post.
samwho.dev/turing-machi...
Reading Dream Machine by Waldrop—great book! Learned von Neumann & Goldstine explored building complex programs from simpler blocks 🧩 in 1947!
How much progress have we made since?
Planning to read their work: Planning and Coding Problems for an Electronic Computing System. #computingHistory
