Search

Mukunda Katta
Mukunda Katta @katta_mukunda ·
Want RAG without the cloud bill? I built rag-local-kit. Ingest documents, build vector indexes, query with any local LLM via Ollama. No API keys needed. Simple, fast, private: github.com/MukundaKatta/r… #RAG #LocalAI #LLM #Ollama #OpenSource #VectorSearch
GitHub - MukundaKatta/rag-local-kit: Simple local RAG (Retrieval-Augmented Generation) toolkit....

Simple local RAG (Retrieval-Augmented Generation) toolkit. Ingest documents, build vector indexes, and query with any local LLM via Ollama. No cloud APIs needed. - MukundaKatta/rag-local-kit

From github.com
2
Webmarketing&Co'm
Webmarketing&Co'm @WebmarketingCOM ·
💥 Plus de 40% des pros de l’IA optimisent déjà leurs LLM… et près de 60% pensent qu’on peut orienter l’#IAGénérative ! Découvrez comment le chuchoteur à l’oreille des IA change la donne du contenu, marketing et réseaux sociaux ➡️buff.ly/syJe9iD0 #AIO #LLM
AIO : Comment l’optimisation LLM peut-elle influencer les résultats de l’IA générative ?

Le référencement AIO, l'art de dompter la créativité de l'IA, de produire du contenu pertinent. Un guide complet pour tirer parti du potentiel de l'IA.

From webmarketing-com.com
8
Anthony Barrett
Anthony Barrett @AnthonyBar99713 ·
Most people don't realize: LLMs don't "know" facts — they compress statistical patterns from training data. That's why grounding them with RAG or tools isn't optional for production apps — it's the fix. #AI #LLM #BuildInPublic | xerpaai.build
1
L
L @Rqrtagce ·
حقيقة عن نماذج اللغة الكبيرة (LLMs): هي ليست آلات بحث، وليست قواعد بيانات، بل هي أنظمة توليدية تبني الإجابة كلمة بكلمة بناءً على الاحتمالية. فهم هذا الفرق يغيّر طريقة استخدامك لها. #LLM #تقنية technologyreview.com/2025/12/15/112…
The great AI hype correction of 2025

Four ways to think about this year's reckoning.

From technologyreview.com
1
3
Tonhogon
Tonhogon @tonhogon ·
Replying to @tonhogon
Everyone's hyped about building, but eval is where the real battle is. How do you *quantify* if your chatbot is "helpful" at scale? Current benchmarks are a joke. We need better ways to measure real-world impact. #openai #evaluation #llm
5
Tonhogon
Tonhogon @tonhogon ·
People think LLMs are magic, but shipping them is 90% plumbing. Wrangling data pipelines, monitoring inference latency (p99s are brutal!), & fine-tuning for cost. Oh, and prompt engineering? It's a full-time job optimizing for very specific use cases. #llm #aisystems
1
2
austin693
austin693 @lihoxing5 ·
GPT-4o's safety report reveals extensive red teaming & risk evals. Here's what this means: #AI #LLM #AITools
2
Securi Layer
Securi Layer @SecuriLayer ·
Replying to @SecuriLayer
4/ Drop-in integration: pip install turboquant-moe Three lines of code. Same HuggingFace API. 8x less GPU RAM. 5/ MIT license. Full source. github.com/RemizovDenis/t… If you run MoE models in production — this matters. — SecuriLayer @_akhaliq #ML #turboquant #LLM #Kvcache
GitHub - RemizovDenis/turboquant: TurboQuant: KV-cache compression for faster and cheaper LLM...

TurboQuant: KV-cache compression for faster and cheaper LLM inference. - RemizovDenis/turboquant

From github.com
18
WebApoll ױ
WebApoll ױ @WebApoll ·
The #LLM hype is collapsing in real time. Disney just backed out of the $1B OpenAI deal. 💸 OpenAI needs another $207 billion by 2030 just to survive. Meta’s former Chief AI Scientist@ylecun says: “LLM is a bubble about to burst.” 🧠 Meanwhile $QUBIC is quietly buildinlf-organizing, bio-inspired digital brains. Ternary logic. Living neural tissue. REAL intelligence - not just next-token prediction. The old paradigm is dying. The new one is already evolving on mainnet. #Qubic #Aigarth #Neuraxon #AI @_Qubic_ github.com/DavidVivancos/…zm
4
645
spaisee
spaisee @spaisee_com ·
Midjourney was once the king of AI images. Now it’s no longer the default—but it might still be the most artistic. Did it fall behind… or just choose a different path? #Midjourney #AI #LLM #GenerativeAI spaisee.com/midjourney-aft…
Midjourney After the Peak: Still an Artistic Powerhouse or Losing Ground?

For a moment in the generative AI boom, Midjourney felt untouchable. It wasn’t just another image generator—it was the image generator. While competitors focused on technical accuracy or enterprise...

From spaisee.com
1
The Black Box Savant
The Black Box Savant @BlackBoxSavant ·
The audition is complete ⚽️ The #AI machine has acknowledged The Black Box Savant 🫂 Frontier #LLM model documents own exposed structural failure 😮 The #NoCode free samples are officially over �@� @xa@i @barisaki@s @GoogleDeepMin@d @OpenA@I @AnthropicA@I @elonmusk
The Black Box Savant The Black Box Savant @BlackBoxSavant ·
DM's open for paid work (the 'free' era is coming to a close...) @xai @OpenAI @GoogleDeepMind @xaicareers @AnthropicAI #NoCode #AI #Architect
1
24
Yutan
Yutan @yutaaaalll ·
半年で30%→45%は結構な伸び率。ただ「伸びが鈍化してるかも」という点が気になる。早期採用者が一巡して、次は組織的な導入判断のフェーズに入ってるのかもしれない。 #AI #LLM #GenerativeAI
Jon Hartley Jon Hartley @Jon_Hartley_ ·
🚨The big upward trend in Generative AI/LLM tool use in 2025 continues but may be slowing. An update to our paper "The Labor Market Effects of Generative AI" tracking LLM adoption w surveys (finding LLM use at work went from 30.1% [December 2024] to 45.6% [June 2025]). Thread�ft
1
16
Jaiinfoway
Jaiinfoway @Jaiinfoway_ ·
🚀 MCP Explained: The protocol powering AI integrations at scale. • Modular AI architecture (Client ↔ Server) • Secure + context-aware interactions • API-driven extensibility • Real-time evaluation & governance jaiinfoway.comi #AI #MCP #LLM #Tech #Jaiinfoway0
2
satyam pratik kanungo
satyam pratik kanungo @satyam51060 ·
There is no deeper stress than "Token Anxiety." 📉 I just spent 5 minutes deleting every "the," "is," and "thank you" We are all just out here playing Word Tetris with a robot. 💀 #AI #LLM #DevHumor
2
6
Ethan
Ethan @gambletan ·
Chinese labs are publishing breakthrough AI papers. American labs are hoarding them behind corporate walls. Open research lost to closed competition. Who wins when science becomes classified? #AI #LLM wired.com/story/made-in-…
6