Facebook'tan yeni AI özelliği: Fotoğraf ve videolarınızı otomatik düzenliyor, gizlilik kontrolü tamamen sizde. Akıllı düzenleme, tam kontrol ve kişiselleştirme! Bu yenilik hakkında ne düşünüyorsunuz?

🚩 #FacebookAI #YapayZeka #Gizlilik #SosyalMedya #Teknoloji

Zuckerberg’s AI hires disrupt Meta with swift exits and threats to leave https://arstechni.ca/EM3h #MarkZuckerberg #syndication #FacebookAI #Facebook #Biz&IT #meta #AI
Zuckerberg’s AI hires disrupt Meta with swift exits and threats to leave

Longtime acolytes are sidelined as CEO directs biggest leadership reorganization in two decades.

Ars Technica

Why Meta thinks AI users are better than real Facebook users is because...

Read more 👉 https://lttr.ai/AgAeq

#metaai #facebookai #CuratedAiTools

Facebook's new AI tool is seeking access to your unshared camera roll images to create AI-edited suggestions and stories. Privacy or advancement, what's your take on this? Share your thoughts. #AI #FacebookAI #TechNews #PrivacyConcerns
https://www.squaredtech.co/facebook-ai-photo-tool-now-asks-to-use-your-camera-roll-even-unshared-images?fsp_sid=2785
Facebook AI Photo Tool Now Asks To Use Your Camera Roll

Facebook AI is now asking users to allow Meta AI to access unshared camera roll photos to generate AI-edited suggestions and stories.

SquaredTech

Why Meta thinks AI users are better than real Facebook users is because...

Read more 👉 https://lttr.ai/AefFj

#metaai #facebookai #CuratedAiTools

Meta Unleashes FAKE AI Users

As if there weren't enough bots already...

Meet DeepSeek-V3 — the 671 billion parameter beast that’s making OpenAI and Anthropic nervous 👀

👀

🧠 It’s:
✔ Faster
✔ Cheaper ($5.6M training vs $60M+)
✔ More accurate on key tasks like coding, math, and comprehension
✔ Open-source + MIT licensed
✔ Deployable across NVIDIA, AMD & Huawei

📊 Performance Highlights:
🔹 MMLU: 88.5%
🔹 HumanEval: 82.6%
🔹 DROP: 91.6
🔹 MATH-500: 90.2%
🔹 Chinese C-Eval: 86.5%

But wait... ⚠️

🚨 Your data goes to Chinese servers.
🚨 It dodges politically sensitive questions.
🚨 It’s already being banned by gov agencies for “privacy risks.”

So is it the best LLM of 2025 or a privacy nightmare?

📥 Read the full analysis report here → https://deepseekagi.org/deepseek-v3-architecture/

💬 Drop your thoughts in the comments 👇
#DeepSeekV3 #AIRevolution #GPT4 #Claude3 #OpenSourceAI #AIComparison #MoE #FP8 #FutureTech #FacebookAI #LLMBattle

DeepSeek‑V3: Architecture, Performance, and Deployment - DeepSeek AGI

DeepSeek‑V3 is a Mixture-of-Experts (MoE) Transformer with 671 billion total parameters (only ~37B “active” per token)

DeepSeek AGI - DeepSeek AGI: AGI News, Tools & AI Model Comparisons

🚨 Meta has decided to start training it's AI-models on EU user data 🚨

To exempt yourself from this:
1. Open your email inbox
2. Search for an email including "AI at Meta"
3. Click object
4. Fill out the form (stating a reason for objecting is NOT mandatory)

5. Spread the warning (especially among your less tech-savvy connections)

#eu #privacymatters #PrivacyGuides #dataprivacy #datasecurity #MetaAI #AITraining #MetaPrivacy #gdpr #InstagramAI #FacebookAI #WhatsAppAI #meta

Meet your new AI FB Friend

Read more 👉 https://lttr.ai/Ach9j

#metaai #facebookai #CuratedAiTools

Meta Unleashes FAKE AI Users

As if there weren't enough bots already...

"Kill one man, and you are a murderer. Kill millions of men, and you are a conqueror. Kill them all, and you are a god." (c) Jean Rostand

З однієї сторони
Якщо ти вкрав книгу, то ти вор. Але якщо ти вкрав міліони книжок ... то ти геній, міліардер, філантроп і далі по списку.

Але з іншої сторони
Вони виклали ollama у відкритий доступ.

Тому поки що не знаю як до цього ставитись

#llama #Facebook #piracy #FacebookAI
https://arstechnica.com/tech-policy/2025/02/meta-torrented-over-81-7tb-of-pirated-books-to-train-ai-authors-say/

“Torrenting from a corporate laptop doesn’t feel right”: Meta emails unsealed

Meta's alleged torrenting and seeding of pirated books complicates copyright case.

Ars Technica