they also did some interesting new experiments
– system-level runtime measurements
– pareto analysis
– quantization as a defense
– password-cracking test overall pretty nice work! thorough, detail-oriented, and fair glad to see my results held up http://
arxiv.org/abs/2507.07700
LLMS
-
New AI Security Research: Quantization Defense and Runtime Analysis
By
–
-
OpenAI Embeddings API Changes Impact Model Performance Distribution
By
–
the other results, which use the OpenAI embeddings API got slightly worse. this also makes sense to me the API has likely changed a bit over the last two years they serve the models differently, use different quantization techniques etc. so there is a small distribution shift
-
Context Drives Generative AI Revolution Through RAG and Knowledge Graphs
By
–
#Context is fueling and boosting the current Generative #AI revolution! See these books:
> https://
amzn.to/34utTz7
> https://
amzn.to/317XV9L
————
#RAG #LLMs #GenAI #DataScience #MachineLearning #Semantics #KnowledgeGraphs #IoT #InternetOfContext -
Kimi K2: New Capabilities This Week
By
–
What specifically does Kimi K2 allow me to do this week that I couldn't already do last week?
-
Building Neo4j-Powered Applications with LLMs and Haystack
By
–
New book from @PacktDataML >> "Building Neo4j-Powered Applications with LLMs: Create LLM-driven search and recommendations applications with Haystack, LangChain4j, and Spring AI" Available at https://
amzn.to/4l9lLKO -
Grok 4 Token Efficiency Compared to o3 Model
By
–
Grok 4 is really expensive – not just because its price per token is higher (it's priced the same as Sonnet or around 2x o3), but also because it needs to generate about 2x more tokens than o3 to complete the same tasks. We can see this clearly from data provided by
-
Kimi K2 Launch Video Highlights Open Source Leap and API Availability
By
–
Kimi K2 launch video! Quite a good leap for open source, according to my tests.
— 🚨 AI News | TestingCatalog (@testingcatalog) 11 juillet 2025
K2 is also available via an API 👀 https://t.co/lE2FqXwvuq pic.twitter.com/8qfoQEXTDYKimi K2 launch video! Quite a good leap for open source, according to my tests. K2 is also available via an API
-
Out-of-Distribution Data Problem in AI Models
By
–
but they're not out-of-distribution, that's exactly the issue
-
@testingcatalog — 2025-07-11
By
–
BREAKING 🚨: Kimi K2 is a new open-source Agentic Model from Moonshot which outperforms Claude 4 on certain benchmarks!
— 🚨 AI News | TestingCatalog (@testingcatalog) 11 juillet 2025
Kimi moment! 👀 https://t.co/kTdkAY7B8S pic.twitter.com/VMzEKyvEmiBREAKING : Kimi K2 is a new open-source Agentic Model from Moonshot which outperforms Claude 4 on certain benchmarks! Kimi moment!
-
From Language Models to Real-World AI Materialization
By
–
La bataille des modèles de langage du type ChatGPT est déjà (presque) une bataille du passé. Le vrai enjeu des prochaines années sera la manière dont les IA se matérialiseront dans le monde réel.
— Rafik Smati (@RafikSmati) 11 juillet 2025
Pour preuve, une info largement passée sous les radars en France : Open AI vient de… pic.twitter.com/HzkFADoULWLa bataille des modèles de langage du type ChatGPT est déjà (presque) une bataille du passé. Le vrai enjeu des prochaines années sera la manière dont les IA se matérialiseront dans le monde réel. Pour preuve, une info largement passée sous les radars en France : Open AI vient de