Hacker News: ganiszulfa

New comment by ganiszulfa in "Show HN: I built a toy TPU that can do inference and training on the XOR problem"

ganiszulfa — Tue, 19 Aug 2025 04:13:47 +0000

Amazing project, and amazing write-up, I especially like the animations. What's the end goal here? Putting these TPUs in the consumer hands or edge devices?

New comment by ganiszulfa in "LeetCode for System Design Interviews?"

ganiszulfa — Fri, 20 Jun 2025 06:51:24 +0000

It looks really good. Great Job!

I thought it was somehow related to LeetCode due to its name. I wonder if it's okay? I read that it's powered with AI yet it's still free, you must have burning a lot of money in this product. What's your monetization strategy in the future?

Building the World's Most Over-Engineered Personal Blog

ganiszulfa — Tue, 17 Jun 2025 10:47:06 +0000

Article URL: https://ganis.net/okblog

Comments URL: https://news.ycombinator.com/item?id=44297602

Points: 1

# Comments: 0

New comment by ganiszulfa in "How to post when no one is reading"

ganiszulfa — Mon, 02 Jun 2025 05:47:31 +0000

If it's not tied to your income, I agree. But I can't imagine the stress when your readership numbers determine whether there's food on the table or not.

New comment by ganiszulfa in "Show HN: My New Startup – Minimus"

ganiszulfa — Sun, 04 May 2025 12:21:31 +0000

Interestingly, the posted link doesn't mention anything and only displays the login box. I'm not the owner, but here's the main website: https://www.minimus.io

New comment by ganiszulfa in "Liverpool's title win has completed a mysterious Fibonacci sequence"

ganiszulfa — Fri, 02 May 2025 04:10:59 +0000

In this case, it's a complete coincidence. I did this for UEFA Champions League winners over the last 33 years and found no Fibonacci sequence at all. I'm pretty sure it won't be found easily in other sports or competitions either.

New comment by ganiszulfa in "Can LLMs do randomness?"

ganiszulfa — Wed, 30 Apr 2025 12:23:41 +0000

LLMs are acting like humans, I believe humans will have biases if you ask them to do random things :)

On a more serious note, you could always adjust the temperature so they behave more randomly.

New comment by ganiszulfa in "The Hallucinating ChatGPT Presidency"

ganiszulfa — Wed, 30 Apr 2025 05:07:39 +0000

I always find it helpful to tell the model to not hallucinate and just say it if it doesn't know. If the model supports configuration rules, put it there.