Skip to main content
HashnodeShinaBR2

Command Palette

Search for a command to run...

Til 2025-04-04

Updated
•1 min read
S
ShinaBR2

Enthusiasm for building products with less code but more reliable.

Part of seriesTIL

  • After increasing the CPU to 4CPU and RAM to 8GiB, processing 1080p quality video with size 700MB took 1795496 seconds (~30 minutes), using 8 CPU will reduce to ~17 minutes

  • There is a new way to run Python in the

  • Another library for building AI Agents: Agno

  • Even in the same data center, network glitches can still occur. I didn't know this before.

  • Devin is another AI stuff, I'm tired of hearing AI agent, agentic AI, etc. People are always try to run faster on the same race crazily, FOMO. Genspark is another tool like Manus but they say it's better. The pattern nowadays is: something + is the first/the best + AI + fancy noun. That's crazy. Like stupid cryptocurrency FOMO, nothing different.

Today read:

  • https://towardsdatascience.com/agentic-ai-single-vs-multi-agent-systems/

  • https://blog.algomaster.io/p/handling-failures-in-distributed-systems

#til#ai

Comments

Join the discussion

No comments yet. Be the first to comment.

TIL

Part 18 of 31

Today I learn

Up next

Til 2025-04-03

The best way to get the answer is not to ask questions, but to put the wrong answer. This is fun, but true indeed. Google’s NotebookLLM doesn’t support some kind of website, like this https://bbycroft.net/llm There is a tool for monitoring the brow...

More from this blog

ShinaBR2

36 posts