Link: https://youtu.be/GXJwX3tpMe0
That is not a joke, but rather a serious problem solved as a side quest and thought of just shipping things. You never know what could work and what won’t.
Source: techstructive-weekly-70
programmer by day, programmer by night.
Link: https://seated.ro/posts/rabbit-hole-learning.html
This is a good idea, but it kind of breaks the cycle kf a rabbit hole to log everything that we did and learnt. It removes the core advantage of learning things by diging deep.
Source: techstructive-weekly-70
Shuffle has always been one of Spotify’s most-used features, and also one of the most misunderstood. For years, listeners have debated whether Shuffle is “truly random,” often noticing patterns or repeated tracks that didn’t feel random enough.
Link: https://engineering.atspotify.com/2025/11/shuffle-making-random-feel-more-human?ref=dailydev
This is a bit under-technical thing. I didn’t get a good flow of the algorithm. i want to understand with an example.
Source: techstructive-weekly-70
Link: https://youtu.be/Ctjiatnd6Xk?si=pznjX7kJd7fkChXx
This is so cool, What is the difference of AI and AGi. Think about it. AI was to humans, as AGI is to AI both are same things just the level of buzz in some of the terms is just hype. Make humans better, making it answer things possible.
Source: techstructive-weekly-70
Link: https://youtu.be/KJfz9ChQ0A0?si=x_9C3AzEQrk0Fy3p
Yeah, been there done that. Twch stack and programming languages are not one fit all shape. It depends. The typical reason in tech.
Source: techstructive-weekly-70
Current AIs struggle to create a whole that exceeds the sum of its parts
Link: https://secondthoughts.ai/p/a-project-is-not-a-bundle-of-tasks
This is one of the best and practically grounded takes on the progress on AI. If it can’t make a decision that lasts longer, it might collapse eventually. AI Coding has improved over the last year for sure, but is it there? Not really. A developer still has to understand the logic, think about the AI answers, there has to be a human intuition and judgement to make its way to a sustainable solution. As per the growth, developers might be overgrown by 2030, is that really that quick? Maybe the plateau is almost here. It can code, sure, but it doesn’t create anything. It can create something like they are created by humans, not something out of the blue.
How I build software in mid-November 2025
Link: https://behan.substack.com/p/alien-authors
Wow, this is just wow. The analogy of writers managing a bunch of written text from AI to create their own story is quite relatable to a developer managing a bunch of parallel agents to generate code and design a feature or come up with a solution. The whole story was really cleverly put and relatable. Though I would argue it’s quite a bad one for making one artist happy. I never want an artist to generate art with AI and tweak and tweak to please someone. Art is something that comes fully from the heart, otherwise, it’s not art. It’s pseudo art, artificial art, forced art.
Link: https://evanhahn.com/fizz-buzz-without-conditionals-or-booleans/
A nice pattern like loop for 3 and 5 divisibility, it won’t scale i think for other problems
Source: techstructive-weekly-69
Link: https://youtu.be/39PdgOYjBMg
Its a good model, it seems its quite heavy and removes the subtle mistakes and biases it has. I love hove Google makes a solid general purpose models. Unlike OpenAI, whose naming conventions are all over the place, its like a slop generation to me. But Gemini models hit different, they just solve what have been given to them. Quite a good upgrade from the 2.5 models. Waiting for Gemini 3 flash version
Link: https://blog.enamya.me/posts/bcrypt-limitation
Oh, that is wired, use Argon guys, if you aren’t just storing passwords. Nice to know that bcrypt is not safe for passwords greater than 72 characters, who would even store such a long password? But that is the thing, subtle decisions, like this is not a password, so we can use bcrypt, and bam, you would be wrong