DSA resources and the relevance debate

Two posts that took off in the same week captured a real split in how students are preparing for software jobs in 2026: one shared GitHub repositories for data structures and algorithms practice, and the other asked whether that whole grind still makes sense now that artificial intelligence coding tools can write so much code. (threadreaderapp.com) (github.blog) Data structures and algorithms means learning the shapes code stores information in, like arrays, trees, and graphs, and the step-by-step methods code uses to search, sort, and update them. Companies still package that skill as timed interview questions on sites like LeetCode, which sells an “Interview Crash Course” built specifically around passing software engineering interviews. (leetcode.com) (support.leetcode.com) That is why curated GitHub lists keep spreading. Students are not just looking for theory notes; they are looking for ready-made maps like The Algorithms, which has more than 60,000 followers on GitHub, and structured paths like NeetCode’s roadmap, which organizes practice by patterns such as sliding window, graphs, and dynamic programming. (github.com) (neetcode.io) The appeal of those repositories is simple: they turn a giant syllabus into a checklist. A repo can bundle solved problems, topic order, code in multiple languages, and revision notes in one place, which is much cheaper than buying several separate courses. (github.com 1) (github.com 2) The other side of the debate exists because coding work changed fast in 2024 and 2025. GitHub’s 2024 Octoverse report said more than 97 percent of developers have used artificial intelligence coding tools at some point, and GitHub now documents features where Copilot can suggest code, review pull requests, and work with repository context. (github.blog) (docs.github.com 1) (docs.github.com 2) That makes the old question sharper: if a tool can generate a binary search or a breadth-first search in seconds, why spend 6 months memorizing patterns. Stack Overflow’s 2024 developer survey showed the tension clearly, with 76.61 percent of professional developers saying they already use or plan to use artificial intelligence tools, while 45 percent said those tools are bad or very bad at handling complex tasks. (survey.stackoverflow.co) (stackoverflow.co) So the argument in 2026 is not really “DSA or AI.” It is whether students should keep treating data structures and algorithms as the center of their training when the day-to-day job increasingly includes prompting, reviewing generated code, and catching subtle mistakes the model misses. (docs.github.com) (survey.stackoverflow.co) Interviews are the reason the old system survives. LeetCode still markets itself as a platform to “quickly land a job,” its help center says it has more than 2,000 company interview questions, and NeetCode still describes its roadmap as “a better way to prepare for coding interviews.” (leetcode.com) (support.leetcode.com) (neetcode.io) The practical outcome is that students now split their week in two. One block goes to classical drills like heaps, graphs, and dynamic programming because interviews still reward pattern recall, and another block goes to tool-assisted work like using Copilot, reading generated pull request comments, and debugging code they did not fully write by hand. (neetcode.io) (docs.github.com) That is why those two viral conversations fit together instead of canceling each other out. The GitHub repo thread is a map for clearing the hiring gate that still exists, and the “is DSA still relevant” thread is a warning that clearing the gate is no longer the same thing as being ready for the actual job. (leetcode.com) (github.blog)