#scaling

N-gated Hacker Newsngate
2025-05-04

AI graphs: the modern art of , where lines go up and IQs go down 📉🤯. Marcus and Davis remind us that just because something is trending, doesn't mean it computes. You'd have better luck finding meaning in a toddler’s scribbles 🖍️.
garymarcus.substack.com/p/the-

2025-05-01

Снижение затрат на разработку микросервисов

💻 Снижение затрат на разработку микросервисов Оптимизация микросервисов без DevOps: NGinx для паузы запросов при перезапуске backend, шина событий на Bun для бесшовного рестарта реплик через общий порт для разных процессов. Код и конфиги

habr.com/ru/articles/906204/

#typescript #javascript #bun #websocket #microservices #highload #scaling #nestjs #honojs

Scaling calculator

A calculator for language model scaling. How many FLOPs do you need to train a 13B model? How much does it cost to scale to 1T tokens? If the we cap training runs at one yetaflop, how many GPUs do you need to break that?

scaling.mishajw.com/

#AI #ML #LLM #scaling

Screenshot of scaling calculator for calculating the amount of compute needed by a language model.
2025-04-27

'Scaling ResNets in the Large-depth Regime', by Pierre Marion, Adeline Fermanian, Gérard Biau, Jean-Philippe Vert.

jmlr.org/papers/v26/22-0664.ht

#scaling #deep #gradients

JCONjcon
2025-04-26

hands-on? We got you.
Join Florian Habermann & Christian Kümmel for two —one for getting started, one for to the stars.
In-memory magic, ready.

🎟️ €19 each → 2025.europe.jcon.one/tickets

2025-04-25

'Scaling Data-Constrained Language Models', by Niklas Muennighoff et al.

jmlr.org/papers/v26/24-1000.ht

#datablations #epochs #scaling

Ecologia Digitaljosemurilo@mato.social
2025-04-23

"The vast investments in scaling... always seemed to me to be misplaced."

"The premise that #AI could be indefinitely improved by #scaling was always on shaky ground… In November last year, reports indicated that #OpenAI researchers discovered that the upcoming version of its GPT large language model displayed significantly less improvement, and in some cases, no improvements at all than previous versions did over their predecessors."

futurism.com/ai-researchers-te

Hacker Newsh4ckernews
2025-04-08

Can reinforcement learning for LLMs scale beyond math and coding tasks? Probably

arxiv.org/abs/2503.23829

Frontend Dogmafrontenddogma@mas.to
2025-04-02
Lena WestLenawest
2025-03-21

What’s harder: or starting? 🤥

Client Info

Server: https://mastodon.social
Version: 2025.04
Repository: https://github.com/cyevgeniy/lmst