Grainme

indie dev, making 95 helping devs go deeper.

[Grainme @Grainmee]

i'm building a free coding platform where you start at E-rank and grind your way to S-rank

by implementing systems from scratch, research papers and more.

solo leveling, but for developers.

https://t.co/5maqXMOWSy

[Grainme @Grainmee]

@DevKokooo I highly recommend this gitbook brother 😁

https://t.co/v4c3SWxEYy

[AI Notkilleveryoneism Memes ⏸️ @AISafetyMemes]

@hilbertspaess

[Antonio Sarosi @antoniosarosi]

When I was in uni after they taught me "Design Patterns" in Java I tried applying them to various projects and every single time it made the code more complicated.

Many times you don't even need an abstraction to begin with. And if you need it, you do after clearly seeing why, not before.

Some abstractions are similar to others but I've never really seen a consistent "pattern" in them.

OOP folks will simply say "well that's the point you don't apply it exactly as written in the book you adapt it to your use case". If so, it's more like a design "blueprint", or "idea", not a pattern. I think that's what the teaching materials failed to communicate to CS students in general.

Another thing is that modern languages are much more versatile than Java in the 90s so there are a lot of "patterns" that can be reduced to taking a function as a parameter or stuff like that, which wasn't as common in heavy OOP languages in the past.

Overall though it's better to wait until you see clear pattern to abstract it. And even then be cautious, one thing I've always hated is code that looks duplicated but it's not really duplicated cause there's a few things that are different and if you try to make an abstraction the abstraction is so complicated that's it's just better to have almost duplicated code.

[Mitchell Hashimoto @mitchellh]

Quick Superlogical demo to end the week. This demo is just of the basic functionality, but I want to highlight just how FAST it is. There's a WHOLE lot more functionality already present and coming but I wanted to keep the demo short. But seriously, check out that speed.

I'm showing the macOS app here and I get asked a lot so let me just say up front: its not macOS only. We're still figuring out exactly what platforms will be stable enough for the initial public release, but we're working on others. In particular, the web interface is very functional but I'll demo that another time (or someone else will).

This is also all showing off the architecture I previously talked about for terminal multiplexing which is significantly different from other mainstream multiplexers: this is all a custom binary protocol where the server is maintaining N replica distributed terminal state machines. It works!

Tons of hard work going into this by the team @almonk @dizzyup @pearkes. Hope we can bring you something you can use soon. ❤️ Happy to answer questions in the replies.

[CJ (Coding Garden) @CodingGarden]

We're at a point where coding without an isolation mechanism or sandbox is way too risky.

Use containers at a minimum, VMs if you have enough resources or a remote machine with either of these if you can.

Watch 🍿: https://t.co/X39CSZNG7M

[Grainme @Grainmee]

@rnmp deep updates 🧙‍♂️

[Grainme @Grainmee]

@DevKokooo

[Grainme @Grainmee]

@OFettal it could be semantic search, mal9awch lik fold 7 👀

[Grainme @Grainmee]

be persistent, be patient and be humble.

[achal @achalllll]

Just wrote a beginner-friendly breakdown of HNSW: The algorithm powering fast vector search

https://t.co/D1f7WZYkIT

[Grainme @Grainmee]

@achalllll i am going to be adding HNSW to 95 soon as a gate 👀

https://t.co/5maqXMOWSy

[Grainme @Grainmee]

95, your escape from vibe-coding.

[Grainme @Grainmee]

"contribute to open source" is the shallowest advice you can ever get.

[Koko @DevKokooo]

Treat programming less like a craft and more like math that you can formally prove before you run it

First principles thinking > reasoning by analogy

An analogy is a transfer of a model from one context to another, with hidden assumptions, and without full validation of the conditions that made the original model work in the first place

"Google interviewed candidates like this, so we're going to do the same too" has created a whole industry of leetcode-ification that didn't prove actual skills on the job

"Forbes 100 company has these best practices" led to cargo cult programming mentality, assuming that copying those same practices would bring the same success

"We're building Uber for dogs and AI for your washing machine" -- wtf does this even mean?

It's the same trap I fell into years ago with Clean Code architecture

We have to remember that "best practices" is not law, is not religion, and everything depends on the context of your project / company

[Grainme @Grainmee]

@PawelJLisowski @athasdev If you rely heavily on extensions, you might want to see if they are available on Zed

[Grainme @Grainmee]

@viii_fn @sudobunni you can use https://t.co/5maqXMOWSy

it’s a free platform and we’re adding challenges actively on the platform.

[Grainme @Grainmee]

you'll build a small langchain-style document chunker, starting with basic json and adding:

- sentence splitting & multi-document handling.
- metadata propagation & stable chunk IDs.
- source indexes, chunk_size, chunk_overlap.
- recursive separators & markdown header metadata.

[Grainme @Grainmee]

new side quest: "build your own chunking", the first part of the mini Langchain arc is live.

this teaching-sized version covers the core ideas behind text splitters and retrieval pipelines.

[Grainme @Grainmee]

dropping a new side quest on 95 soon: mini Langchain textsplitter.

this is floor 1 of an upcoming gate where you build a complete mini Langchain from scratch.

language agnostic. no fluff. 100% free.

[Grainme @Grainmee]

@N_and_ni @Pinterest ?