Sau

I wonder how much time will it take to reach 100 looking to connect to more people in tech space, I recently started posting more often on X

@mdev_1 @geminicli thanks I will check it out

@cachesaur @geminicli their codebase also makes for a great study github.com/google-gemini/…

tbh @geminicli really cooked with model routing. ~50m tokens burnt, and I've only reached like 40% quota? long live these limits

i genuinely think most people just wanna make noise nowadays, keeping your head down and working hard until you have something to share has got more signal. desperation for attention only makes your PSNR go down.

@Hi_Mrinal looks a lot like Zed icons too

REALLY LOVING THE ASTHETICS OF ATHAS EDITOR

@ItsRoboki not good enough it should be 98% rust

@cachesaur rust is already mostly in rust

"rewrite docker in rust. make no mistakes"

@Devbosetwt @kunalvg taste comes from trying various things, even the badly made content/tutorial

@kunalvg Everything boils down to lack of taste and having no voice I’d say cinema plays a huge role in shaping ideas and beliefs

Every other tweet right now: “Instead of watching a movie, watch this AI tutorial.” Go watch the movie. You can come back and still outwork the tutorial crowd. Most are just indulging in bookmark vanity. Very few are taking action.

@bit2swaz hi just came across this I have been working on something similar context-span a memory layer to share the context globally in multi-agent system

Engram is finally done it's a Rust memory backend for LLM agents that sits between your agent and its conversation history made up of three memory layers: - short-term (Redis) - long-term semantic search (LanceDB) - core pinned facts full REST API, token budgeting per request, pair preserving trim, background workers, Prometheus, and Swagger raw numbers from my laptop: - 100 msgs context assembly: 280 µs in-memory, 22 ms real stores - e2e throughput: 64k msg/s - token savings vs full dump: 40% repo link in comments

@forgebitz we should redesign lsps for the agentic era coding

the real bottleneck of multi-agent code maxxing is not tokens it's the typescript language server

the first time I read it I almost went into a depressive state, everything felt so numbing and my desire to do anything was zilch It was silent feeling of accepting defeat without even trying to fight my out of it like how do you fight against the inevitable, the time itself.

@_chenglou forever thankful to learning web-dev fundamentals from the odin project, they had a section about tdd and its importance

Test-driven development is cool again

@code_kartik twitter is way past its peak which was 2020-2023 just a shell of its former self (just like me)

It’s really depressing to scroll x nowadays rarely I see a genuine post and like it. The golden days of this platform have already been passed.

@Hi_Mrinal let it happen by Tame Impala

If I comment something similar God will set me as an example for "why not to"🥀🥀

@Hi_Mrinal bookmarked it for night reading

SHAMELESS REPOSTING ^^

@skylane0_0 Me too. 🤭

aur kisiko chahiye toh bolo😋im feeling unemployed today

@mdev_1 any resource on this? I have 15tb of college provided gdrive space

inspired by this, I converted my 1 TB GDrive into swap space now my machine has theoretically infinite RAM

@vedantReddy45 okay +1 and you'll get there sooner or later

No way instead of increasing it decreased 😭😭 People aren't taking me seriously 😒 Time to be serious boyiii 😸

@_wedan does the work it was made for.

@cachesaur cool stuff 🔥

the little helper cli agent I built using only 200 python loc and gpt-oss-120b is actually helping me in my current project, organizing files, suggesting naming conventions and random doubts i + zed + grunty + claude + chatgpt + deepseek make a good team

@theo I have to ask why we just dont do +1 to the previous uuid? incremental ids don't sound too bad compared to this

Do I have to do another UUID video?

@paulbohm talk about the overhead of it first generate an uuid, then test it against the stored uuids in db and then based on that proceed its like 3 extra db call for every single request i wonder what was the latency of a simple signup

If your startup does not have a UUID microservice you’re ngmi