Two lists dominate how software engineers prep for coding interviews. The Blind 75, first assembled on the Blind forum around 2018 by Yangshun Tay, and the NeetCode 150, which expanded on that foundation with 75 more problems and added video walkthroughs. Most engineers I know have an opinion about which is better. A lot of those opinions are based on vibes, not a real comparison.
Here’s what I actually think after looking at both carefully.
what the Blind 75 actually covers
75 problems across about 10 pattern categories. Arrays, binary search, trees, graphs, dynamic programming, the usual. The list is lean. You can finish it in 4 to 6 weeks if you’re putting in an honest hour or two a day.
The appeal is the constraint. You’re not trying to memorize 400 solutions. You’re learning to recognize 10 or so patterns and apply them. For engineers interviewing at FAANG or similar companies on a tight timeline, this is genuinely the faster path to being interview-ready, not perfect, but functional.
One real limitation: the Blind 75 doesn’t cover heaps and priority queues well, barely touches tries, and skips a few advanced graph problems that show up in harder rounds. If you hit a “find the k-th largest element” question in a round and you never practiced heaps, that’s a gap.
what NeetCode 150 adds
The NeetCode 150 is the Blind 75 plus 75 more problems. It fills most of the coverage gaps: more heap problems, more backtracking, advanced DP (2D, unbounded knapsack), tries, and bit manipulation edge cases.
The other thing NeetCode adds is video explanations for every problem. This matters more than people admit. Reading a solution and watching someone explain why they made each decision are different cognitive experiences. The video forces you to follow a reasoning chain, not just pattern-match the code.
The cost is time. 8 to 12 weeks at a similar daily pace. That’s fine if you have it. A lot of people who start NeetCode 150 don’t have 10 weeks and end up half-finishing it, which is worse than finishing the Blind 75.
how to decide (honestly)
There’s a fairly simple heuristic. If your real interview is in 6 weeks or less, do the Blind 75. If you have 8+ weeks and you’re targeting senior or staff roles, do NeetCode 150. If you’re targeting Google L5+ or Meta E5+, do NeetCode 150 regardless of timeline and also practice hard problems on top of it, because those lists weren’t designed for the hardest rounds.
One thing I’d push back on: the idea that NeetCode 150 is strictly better. It’s not. For someone with 4 weeks and a FAANG phone screen, doing a focused 75 problems deeply is better than skimming 150 shallowly. More coverage doesn’t help if you never got the first set of patterns into muscle memory.
The Stack Overflow Developer Survey 2024 showed that among developers actively job-searching, time-to-interview-ready is one of the factors they most consistently underestimate. The lists aren’t the bottleneck. Depth of practice on each problem is.
the categories that trip people up in both lists
There are a few areas where candidates consistently underprepare even after completing one of these lists:
- Sliding window edge cases. The pattern looks simple until you hit a problem with multiple constraints simultaneously.
- Graph traversal in grids. Both DFS and BFS on a 2D grid feel different than on an adjacency list, and the code looks different too.
- DP with state compression. This barely appears in Blind 75 and shows up in NeetCode 150 but not enough times for most people to get comfortable.
If you’re finishing either list and have time, I’d add 10 to 15 targeted problems in whichever of these three areas felt weakest. Don’t add more lists. Drill the gaps.
a note on problem-solving vs. answer-memorizing
This is where people waste the most time. Going through 150 problems by reading the solution when you’re stuck, then moving on, gives you pattern exposure but not problem-solving ability. Interviewers don’t ask exactly the problems on these lists. They ask problems that require applying those patterns to new situations.
The only way to build that transfer skill is to actually be stuck. Set a hard 25-minute limit before looking at solutions. If you’re solving every problem in 10 minutes, you’re probably looking at hints too fast. The discomfort of being stuck is part of what makes the preparation work.
LeetCode’s Blind 75 discussion thread has had this debate for years. The consensus, such as it is, lands consistently: it’s not which list, it’s how hard you think during each problem.
Pick one list. Finish it. That matters more than which one you chose.