Pixel Publisher

Dedicated to spreading all latest news on Global Technology | Our industry does not respect tradition – it only respects innovation – Satya Nadella

[META PCs @METAPCs]

JUST ADDED! New ready-to-ship gaming PCs optimized for the best performance ✅

If you've been looking for a sign it's time to get a new PC, here it is. Check em' out 👇

[The Hacker News @TheHackersNews]

Critical flaw found in #Docker Engine allows attackers to bypass authorization plugins (AuthZ) - CVE-2024-41110, CVSS score 10.0.

This #vulnerability can lead to severe privilege escalation, affecting numerous Docker versions.

Details here: https://t.co/n04jOEAvEf

#DevOps

[NVIDIA Omniverse @nvidiaomniverse]

Headed to #SIGGRAPH2024?

👇Check out these events you won't want to miss:

[NVIDIA @nvidia]

Land an internship at NVIDIA: Linh Nguyen, who leads our campus recruiting teams, shares inside tips about how to stand out from the crowd. #NVIDIAlife https://t.co/8Fy5YieJfQ

[Pixel Publisher @pixelpublisher]

Google Chrome introduces new security warnings for suspicious downloads. Users can now send encrypted files with passwords for deep scans, ensuring comprehensive threat detection.

[The Hacker News @TheHackersNews]

Researchers warn of a campaign exploiting Selenium Grid services for #cryptocurency mining.

Over 30,000 instances are exposed to remote command execution, necessitating immediate security measures.

Learn more: https://t.co/QaHflqNAJm

#infosec #technews #cybersecurity

[Maxime Labonne @maximelabonne]

Here are the juicy lessons from the 1T model Tele-FLM-1T:

- Trained progressively in 3 stages: 52B -> 102B -> 1T parameters on ~2T tokens

- Depth growth technique: selects layers to duplicate based on their input-output distance metrics, prioritizing middle layers with smaller distances

- µP-based hyperparameter search with a small model (Tele-FLMµP) to efficiently search for optimal hyperparameters, which can then be scaled up for the larger model

🤗 Model: https://t.co/R6L1N4zIn1
📝 Paper: https://t.co/KMEo4V5sb8

[DeepLearning.AI @DeepLearningAI]

OpenAI unveiled GPT-4o Mini, a compact yet powerful multimodal generative model designed to outperform its peers at a lower cost.

Learn more in #TheBatch: https://t.co/cNVA1P7MHr

[Awni Hannun @awnihannun]

MLX best practice of the day: don't evaluate graphs too frequently.

[The Hacker News @TheHackersNews]

⚠️ Cyber Alert: #CrowdStrike warns of a new phishing campaign exploiting the Falcon Sensor update mishap.

This sophisticated attack targets German customers with fake installers, aiming to steal sensitive data.

Read here: https://t.co/kxvN8HWbol

#Cybersecurity

[Bloomberg Technology @technology]

Apple's iPhone has been shunted into sixth place in the Chinese smartphone market, according to market tracker IDC https://t.co/LpRG7GWM8T

[NVIDIA AI Developer @NVIDIAAIDev]

Today, @AIatMeta released Llama 3.1 405B 🦙, their most powerful open #LLM that can be used for synthetic data generation, improving model accuracy.

➡️ https://t.co/OI71dyibKq

Technical deep dive on how to use synthetic data to improve LLMs and evaluate RAG 👇

[Maxime Labonne @maximelabonne]

I'm releasing the first uncensored Llama 3.1 Instruct model with GGUF quants

This is an abliterated model that works well in my tests. You can learn more about it in my article.

🤗 Model: https://t.co/mQwn7w5nBS
📝 Article: https://t.co/fkBA1EYWNH

[NVIDIA AI Developer @NVIDIAAIDev]

Explore how @UneeQAI integrated NVIDIA ACE into their proprietary #AI animation system, Synanim, to power the creation of lifelike experiences with digital human avatars. https://t.co/P1D93Y27bs 👇

[Awni Hannun @awnihannun]

Latest MLX has einsum

pip install -U mlx

Attention in 3 lines:

[DeepLearning.AI @DeepLearningAI]

Enroll in Intro to Federated Learning and Federated Fine-tuning of LLMs with Private Data today! ➡️ https://t.co/026gzebJRJ

[NVIDIA @nvidia]

Join us at #SIGGRAPH2024 for two keynote presentations from our CEO and Founder Jensen Huang on Monday, July 29 at 2:30pm MT.

Don't miss #OpenUSD day, #GenerativeAI day, research panels, our Startup Innovation Zone, training labs, and more: https://t.co/k4kqrRN6G8 @Meta @WIRED

[U.S. Federal Housing (FHFA) @FHFA]

Team Open Angels was recognized for best Consumer Experience idea: standardizing the rental application process through tenant screening and approval. #fhfatechsprint #ai #generativeAI #innovation #fhfa #techsprint

[The Hacker News @TheHackersNews]

North Korean cyber espionage group APT45 is now deploying #ransomware.

This marks a significant shift from traditional espionage to financially-motivated attacks, affecting critical infrastructure.

Read more: https://t.co/WgfOO2h5bW

#CyberAwareness #ThreatIntelligence #Tech