Which Go built-ins do people actually use?
Which Go built-ins do people actually use?
Source: [https://x.com/i/status/2079542805060366467
Collect any snd all comment from internet to question asked in this post](https://x.com/i/status/2079542805060366467
Collect any snd all comment from internet to question asked in this post)
📌 Anton Zhiyanov’s July 2026 X poll asked which of Go’s 18 built-ins developers use day to day. Replies clustered hard around make/append/len, with complex/real/imag and print/println almost never touched.
❓ The question
Since Go 1.21 there are 18 built-in functions (15 before). Anton (@ohmypy) asked which ones people use often vs never, naming make, append, and len/cap as constant, and complex/real/imag plus print/println as unused.
📊 Community consensus
Replies mostly agreed: slice/map/channel helpers dominate. complex/real/imag and print/println are nearly unused. new, cap, recover, and panic sit in a middle “sometimes” band.
📦 Why cap exists
Some never use cap on slices. The useful case that stuck: size a channel buffer from another channel’s capacity so buffers match without hardcoding lengths.
💬 complex & print debates
Complex numbers drew skepticism—maybe library types, not primitives. print/println are handy in import-free doc examples, but the spec does not guarantee they stay in the language.
🧵 Side riffs
A quote tweet joked about living without else via early returns. Someone “forgot yield”—Anton clarified it is only a conventional name, not a keyword or built-in.
Key facts
| Fact | Value |
|---|---|
| Author / date | Anton Zhiyanov (@ohmypy), 21 Jul 2026 |
| Built-in count | 18 since Go 1.21 (15 before) |
| Engagement | ~99 likes, 9 reposts, 1 quote, 9 replies, ~6.7k views |
| Used constantly | make, append, len (often with cap) |
| Almost never used | complex, real, imag, print, println |
| Off-platform discussion | None of substance beyond X; LinkedIn repost only |
Details
Anton framed Go’s built-ins as a small, fixed surface and turned “which ones do you use?” into a light community poll. His own split—make/append/len-cap always, complex and print/println never, everything else in between—matched most replies. A follow-up packed 13 of the 18 built-ins into one dense expression for comic effect.
Commenters refined the middle band: panic often reserved for impossible-by-contract bugs; new rare but still used (e.g. strings.Builder); cap more valued for channel buffers than for slices. Side threads questioned whether complex (or even floats) should be language primitives versus library types, and noted that print/println are unofficial debug/doc tools the language spec does not promise to keep.
Discussion stayed almost entirely on X. A same-day LinkedIn cross-post exists, but no substantive Reddit, Hacker News, or public LinkedIn comment threads answering the poll turned up.