Did you know that we are already done?
it is on default free gpt model, same on claude, and gemini.

As my friend suggested, most probably they are generating deterministic code and run it to get the answer.

Daily (308/373 streak until LC cap)

Today is easy one, Happy Saturday. It is a bit bullshit tbh.

https://leetcode.com/problems/weighted-word-mapping/description/?envType=daily-question&envId=2026-06-13

#daily #easy #nopattern

I found what I am gonna be doing when AI finally replaces me

Daily (309/373 streak until LC cap)

Amazing problem, no jokes. It is non-bullshit, probably from the list, and uses 2 most popular linked list patterns.

Highly recommended. Note though, you should solve it in O(1) space complexity.

https://leetcode.com/problems/maximum-twin-sum-of-a-linked-list/description/?envType=daily-question&envId=2026-06-14

#daily #medium #recommend #pattern

I have already mentioned that I'm now fully vibecoding everything. I don't event bother to open PR - AI does it.
Do I enjoy it - yes, but a bit less.

Let's separate work related and personal projects.

For personal projects, I'm enjoying a bit less in a sense that I don't feel "ownership" and I don't have much "understanding", and even sometimes ideas are not mine. This way I'm a bit more upset. But, on the other hand, you have so much opportunities to get "your idea working" fast.

Example1: I was deving symptom diary, as I'm hypochondriac, and got tired doing that in google docs. In comments I will show you the initial design of it (I'm backend guy pretty much). It was awful. I was struggling a lot with frontend, as I don't like it and don't see the reason to learn but still have to make it work as I want it to. So Initial, very ugly, half working version appeared in few months. Now with ai, I did the same progress, but more features, less bugs and amazing design in few weeks(it was with shitty copilot gpt model/free)

Example2: I needed my dev tools, for ip lookup, regex, encoding/decoding, bits conversion etc. I made it in few days. Back then it would be multi months effort. I'm happy that I can deliver something so fast, it feels like cheating in games, and it still brings this enjoyment "I created this product, and use it, nice, cool", however not from code perspective. Now I have more time to spend on learning and improving some low level stuff that I genuinely interested to learn and improve, instead looking up how to fu cking center the div.

Example3: in my symptom diary, I wanted to switch color from blue to green recently. Yea, AI didn't create some "config" or something so that I have one place of changing it. But it is not a problem, I asked opus to do that and it did it in 1 shot in 10 minutes zero problems, zero errors.

For the work related things, AI is again, pros and cons. I personally started working more, as there are higher expectations from developers, besides that I need to spend time on learning these tools. Usually the code that is generated is shit, let's be honest, it is less readable. Now writing to reading code ratio went down from 10/90 to 0/100, so mentally I'm getting tired much more often: more PRs to review + my own PRs.
However, it again gives you speed and a lot of the time "shortcuts" so that you might skip fixing some boring bug you are not interested in

Example1: my project's build somehow broke env for all the people due to mismatch of cpu architecture or something like that. I didn't even created these pipelines, didn't know how they work, BUT, I prompted for 10-15 mins, opened PR - the guy switched to my branch, confirmed it is working - we moved on. I would spend around a day to figure it out on my own.

Example2: the "figuring out" how some project/code works boosted 2-3 times. Because now I don't have to search and guess where some specific behavior lives and how it works. This is the top1 usage in my case, and it helped me many times to spend much less time to understand something - figure out how it should be done and fix (ask AI to fix). On my own I would do it in 1 months, with AI I did in 1-2 weeks. I feel I more and more learning stuff that are critical and interesting for me, rather then figuring out how exactly same shit is done in this specific company and all this boilerplate. Now it is for AI, and I can be busy learning TLS, networking, POCs etc.

Exmaple3: POCs, omg, this is game changer for sure. It is SO OP to be able to confirm your hypothesis in a matter of hours, rather than weeks-months.

Daily (310/373 streak until LC cap)

Amazing problem again, no jokes. Easier than yesterday's one, again uses the well known pattern.

Highly recommended

https://leetcode.com/problems/delete-the-middle-node-of-a-linked-list/description/?envType=daily-question&envId=2026-06-15

#daily #medium #pattern #recommend

Daily (311/373 streak until LC cap)

Should be marked as easy on these constraints.

https://leetcode.com/problems/process-string-with-special-operations-i/description/?envType=daily-question&envId=2026-06-16

#daily #medium #nopattern

Daily (312/373 streak until LC cap)

As always, second version of yesterday's problem, pretty hard, no real reusable pattern - so up to you I wouldn't spend time on solving tbh.

https://leetcode.com/problems/process-string-with-special-operations-ii/description/?envType=daily-question&envId=2026-06-17

#daily #hard #nopattern

I'll be honest, I really don't think you need to learn that much design, patterns etc, not now, maybe 5 years ago. I already learnt it many years ago, so this is good literature in case you want to learn.

Distributed Systems, storages, networks, queues, OS -> it is a must. This is your expertise that will make less replaceable, as to make decisions for AI -> you need this expertise. Go as deep as you can here.

DevOps / SRE -> for highly critical components, probably will never be fully handed over to AI, so oncall people and SREs will be with us, so worth learning along with docker, kubernetes and clouds.

Security - a must, but my advice, as you fill the gaps in above's you will know enough to spot non-secure stuff straight away.

P.S. I didn't mention algos, because let's be honest in backend I rarely use anything else than dictionary/list, but would be great to cover basic leetcode algos, especially graphs, bitwise/binary operations, and bunch of data structures: trie, list, priority queue, etc.

For gaps that I had: https://t.me/tenfoundation/896

Daily (313/373 streak until LC cap)

This is fucking amazing problem, for real, I'm back to school, when I need to solve math puzzle that is also about something in real life, incredible feeling.


https://leetcode.com/problems/angle-between-hands-of-a-clock/description/?envType=daily-question&envId=2026-06-18

#daily #medium #nopattern

Daily (314/373 streak until LC cap)

We have met this q as daily some time ago, very easy, 1 min coding.

https://leetcode.com/problems/find-the-highest-altitude/description/?envType=daily-question&envId=2026-06-19

#daily #easy #nopattern

Sorry for the silence, everyone - I know some of you were worrying. I'm currently going through a tough personal time right now. Overall, everything is going to be okay, but I just need a little time to recover.

Please keep doing your daily leetcode. I promise I will catch up with all of the problems and post here soon. In the meantime, I’m already planning the next article: Concurrency and Race Conditions for System Design (including real, practical applications).

Soon we are gonna return the regular pace: at least 1 post a day.

Hello all new people, probably from Anna's meet, if you are from somewhere else - pls do let me know, maybe I should say kudos to somebody else.

Quick intro about the channel, and what we are doing here is in "replied".

Btw, Anna is writing stuff here: https://nammemememe.substack.com

you might get a lot of interesting things about FAANG interviews, offers, how to prep, etc.