eager_eagle

No point in messing with swappiness settings, trust me. You want to address the reason this happens in the first place, not disable to reduce swap.

I've had a nearly identical scenario at work. The reason? A data verification process that went through 100s of thousands of files to run their checksums and check integrity. This was working as intended, but the system alerts were firing because of that high buffer/cache utilization. So check which processes have highest IO and you might find the answer. Swap is just a side effect.

yeah, that phrase grinds my gears. If you heard from a supposed professional that they only do the bare minimum to get a pass, you'd never hire their services. It's at best disrespectful with your colleagues who give a shit and have standards.

Yup, the longer you wait to update something, the higher the cost to eventually do that. I hate the "if ain't broke don't fix it" adage because of this. If I am the one who will eventually need to update it, you bet I will be "fixing" it more often than I have to.

I'd say that "struggles" is a bit too much. And it's still 5x to 10x cheaper than the models they compared to in benchmarks, let's see by how much they're increasing that cost.

This is so annoying and exhausting. Today I gave up linking a bank account with a broker via plaid because they wanted access to all my transactions.

I'm not in this area, but from what I remember from undergrad, real world systems have too many variables to simulate. It quickly becomes a combinatorial problem, not a reasoning one.

Spotify has too many users and control over the industry, nothing good can come from that. Do your part and stop using it.

for inference you're only counting active parameters towards VRAM, and some labs / models don't train at 32b precision, or even use the same precision for different parts of the network

to me and probably to over 2B people

free to make numbers up

no other platform even compares

if $1 is not installed after that, wipe your hard drive and start over

you can self host librechat, or use their hosted front of you don't want to bother but still prefer to use something open source

I still like doing cat file.txt | grep thing because I often re-run it to find something else like cat file.txt | grep thing2 - and if I do that with grep, the pattern sits awkwardly in the middle instead of at the end

it does, unless the reader has an OCR mode of sorts

they really cooked with that video, got me more excited than with most trailers

how many hours per day on avg? Just curious

any computer science course teaches those things

Eager Eagle

@ eager_eagle @lemmy.world

Posts

11
Comments

925
Joined

3 yr. ago

Eager Eagle

Why is my home server using so much RAM for cache + buffer?

Most of your tech debt is free

Trump says he will declare Strait of Hormuz a US territory

Most of your tech debt is free

DeepSeek’s updated V4 Pro AI model struggles on benchmarks, shines in cybersecurity: impresses in niche areas but leaves some developers disappointed in its overall capabilities.

Rust is adopting an LLM policy

Battle.net's discord permissions

AI and formal verification

Spotify threatens industry with ad-skipping tool: "Skip Ahead" is available for Premium users in selected markets, and enables them to skip entire ad breaks, intros, sponsorship messages.

Bytedance is training an AI model with up to ten trillion parameters: That's three times the size of Moonshot's Kimi K3, currently the largest Chinese model.

YouTube’s AI slop detector incorrectly targets Kurzgesagt as other creators fear same fate - Dexerto

Nah it wasn't him. He was with me studying about Jesus.

YouTube’s AI slop detector incorrectly targets Kurzgesagt as other creators fear same fate - Dexerto

finally something worse then curl | sh

Alibaba’s AI model Qwen3.8-Max made widely accessible ahead of open-weights release: Flagship 2.4 trillion parameter model highlights native multimodal capabilities and ... agentic tasks.

Linux Cat

This free font can trick AI scrapers into swallowing gibberish instead of your content

This free font can trick AI scrapers into swallowing gibberish instead of your content

I think that jobs are starting to pick up

Welp

Jellyfin to quadruple their prices following Plex's price increases

18-Year-Old NGINX Rewrite Module Flaw Enables Unauthenticated RCE

Dirty Frag: Universal Linux LPE - CVE similar to Copy Fail

"The Git Commands I Run Before Reading Any Code"

Self-hosted voice assistant with mobile app

VS Code Release 1.100

VS Code January 2025 (version 1.97)

VS Code Release 1.94

VS Code | January 2024 Release | 1.86

SSH protects the world’s most sensitive networks. It just got a lot weaker

"Slack’s biggest redesign ever tries to tame the chaos of your workday"