GigaMemory на AI Journey Contest 2025: итоги

Приветствуем всех! С вами снова ML-команда RnD для B2C SberAI. Этой осенью в рамках AI Journey Contest 2025 мы представили задачу GigaMemory: global memory for LLM. Её цель — создание автономного модуля долговременной памяти для языковых моделей, способного накапливать и использовать знания о конкретном пользователе, по сути наделяя ИИ способностью «помнить» своего собеседника. Пришло время объявить результаты соревнования и разобрать лучшие решения участников!

https://habr.com/ru/companies/sberbank/articles/974310/

#artificial_intelligence #natural_language_processing #большие_языковые_модели #хакатон #соревнование #машинное_обучение #память #gigachat #хакатоны

GigaMemory на AI Journey Contest 2025: итоги

Приветствуем всех! С вами снова ML-команда RnD для B2C SberAI. Этой осенью в рамках AI Journey Contest 2025 мы представили задачу GigaMemory: global memory for LLM. Её цель — создание автономного...

Хабр

Over on youtube, @nate.b.jones on the channel "AI News and Strategy Daily" about the recent wave of firings at Amazon. Nate used to work there, I suspect in mid-level management from his presentation and dress, and was among those fired.

He has some interesting views on what this state of affairs has to say about the financial stability of AI financially in general, which I think points up the fact that when we speak of AI, we are often talking about technologies that are vaguely similar but applied in vastly different

The tl;DR is, Amazon didn't fire these 20,000 people because they've fully automated their operations at any level; they were fired because Amazon is actually in the weeds with AI, having rested on their AWS laurels while Microsoft Azure and Google Cloud have surged to the forefront delivering AI Services to the corporate workplace. Consequently, Amazon is now in a position where they must cede the space, or make massive investments in AI compute (read nVIDIA GPUs), that they simply could not afford without reducing fixed expenses (salaries, in this case).

He has a very interesting take on the whole thing. It's worth a watch.

https://www.youtube.com/watch?v=NP-qmffUNHQ

#AI #artificial_intelligence

The Dirty Secret Behind Amazon's 30,000 Cuts: Nvidia

YouTube
Poems Can Trick AI Into Helping You Make a Nuclear Weapon

It turns out all the guardrails in the world won’t protect a chatbot from meter and rhyme.

WIRED
#Servicetweet #News #Medizin #Medicine #KI #artificial_intelligence
➡️ Der "digitale Zwilling" eines vollständigen Klinikbetriebs, ein "Simulacrum" hat in China den "Betrieb" aufgenommen
➡️ Chancen u. Risiken eines gänzlich digitalen Klinikbetriebs
https://forschung-und-wissen.de/nachrichten/medizin/chinas-erstes-ki-krankenhaus-soll-10.000-patienten-pro-tag-diagnostizieren-133710762 1/x
Chinas erstes KI Krankenhaus soll 10.000 Patienten pro Tag diagnostizieren

Chinas KI Krankenhaus Agent Hospital prüft virtuelle Ärzte in simulierten Fällen und zeigt Chancen und Risiken medizinischer KI Diagnostik.

Forschung und Wissen

The #PC giant bets big on #artificial_intelligence to drive future growth—but rising chip costs and weaker earnings guidance unsettle investors.

#HP #IA #BigTech #TechInsider #JobsCut #MarketInsider

https://marketinsider.net/hp-to-cut-6000-jobs-as-ai-overhaul-reshapes-its-global-workforce/

HP to Cut 6,000 Jobs as AI Overhaul Reshapes Its Global Workforce - Market Insider

The PC giant bets big on artificial intelligence to drive future growth—but rising chip costs and weaker earnings guidance unsettle investors.

Market Insider

"For all the alleged complexity of generative AI, at their core they really are models of language."

A good read about #LLMs : Don't confuse them for #intelligence

#AI #artificial_intelligence #GenerativeAI

https://www.theverge.com/ai-artificial-intelligence/827820/large-language-models-ai-intelligence-neuroscience-problems

Large language mistake

Neuroscience indicates language is distinct from thought, raising questions about whether AI large language models are a viable path to artificial general intelligence.

The Verge

I was able to create a gem for Google in no time. I am attempting to do the same thing with Microsoft CoPilot, but I cannot access it with my personal account or go to the download studio because I do not have a work or school account.

How are people able to learn how to use the tech if they need an expensive subscription?

🤬

#artificial_intelligence #microsoft #developer

Sztuczna inteligencja jako Bóg. Dystopijna wizja twórcy ChatGPT Ryzyko narzucenia hegemonicznego światopoglądu jest realne AI, a moralność - kto zdecyduje o normach? #artificial_intelligence #technologie cynicy.pl/sztuczna-int...

Sztuczna inteligencja jako Bóg...
Sztuczna inteligencja jako Bóg. Dystopijna wizja twórcy ChatGPT | Cynicy.pl

Sam Altman, współtwórca i CEO OpenAI, jako jedna z najbardziej wpływowych postaci w branży technologicznej, nie unika kontrowersji. W wywiadzie dla Tucker Carlson Network poruszył tematy duchowości, moralności oraz potencjalnego wpływu sztucznej inteligencji (SI) na ludzkość. Jego refleksje odsłaniają zarówno nadzieje, jak i poważne obawy względem technologii, które szybko zyskuje na mocy i zasięgu. Tradycyjny […]

Cynicy.pl