/images/avatar.png

The Go init Function

The init function in Go is a special function that is called when a program starts. It is used to initialize the program and set up any necessary state.

Pointers vs Values in Go Structs

In Go, structs are a common data type used to represent complex data structures. When working with structs, it’s important to understand the difference between struct pointers and struct values.

Stack vs Heap in Go

When it comes to programming in Go, understanding the differences between stack and heap is crucial for efficient memory management.

Go Modules

Package management in Go has been a hotly debated topic in the Go community. Before Go 1.11, there was no official package management tool for Go, with the most popular third-party options being dep and glide. However, with the release of Go 1.11, the Go team introduced their official package management tool Go Modules which became the officially recommended package management tool as of Go 1.13.

Human-Interface Design Best Practices

Human Interface Guidelines do not stop interface design from dropping into two pits: treating it as “making things look good,” or as assembling components from a platform spec. The former is pretty and unusable; the latter is a compliance collage.

In real products, interface design isn’t a style competition—it’s about guiding users to correct outcomes under time constraints, complex conditions, and incomplete information. Apple HIG and Material Design are worth studying, but button radii and shadows are not the lesson.

GPT Best Practices

Most teams fail with GPT not because the model isn’t powerful enough, but because they’re still treating it like a search box or an infallible consultant.

The reality is simpler and more honest: models are probabilistic assistants. They’re fast, capable, and prone to confident mistakes. What you put in—task boundaries, input materials, verification mechanisms—determines the ceiling of what comes out.

This isn’t about the latest buttons on any particular product or chasing model naming trends. It’s about one thing that matters more: how to make GPT-style models controllable, verifiable, and actually useful in real work.