Lmst

AHAAHAHAHAHAAAA !! LOL !

Adversarial Poetry as a Universal Single-Turn Jailbreak Mechanism in Large Language Models

https://arxiv.org/html/2511.15304v1

#ai #AiResearch

A new paper arrives on Nov 24, 09:00 JST.
It’s time to move one step beyond what we thought we understood.

#AI #Emergence #AGI #AIREsearch #MachineLearning #AITheory

https://winbuzzer.com/2025/11/21/ai2-releases-olmo-3-a-fully-open-model-flow-to-challenge-black-box-ai-xcxwbn

AI2 Releases OLMo 3: A Fully Open 'Model Flow' to Challenge Black Box AI Paradigm

#AI #OpenSource #OLMo3 #LLM #AI2 #GenAI #MachineLearning #AIResearch #OpenScience #DeepLearning #AIModels #EthicalAI

Olmo 3: Charting a path through the model flow to lead open-source AI

https://allenai.org/blog/olmo3

#HackerNews #Olmo3 #OpenSourceAI #ModelFlow #AIResearch #MachineLearning

Are We Building Intelligence—or Copying It?

#AIResearch #TechEthics #NeuroTech #FutureOfAI #BioDigital #FediverseTech

🚀 Wow, Kagi Assistants have arrived to revolutionize...research assistance? 🤔 After a "happy accident" of topping a benchmark, we've learned that Kagi is all about putting humans at the center, which is why you need two AIs to assist you in being human. 😂 Keep those benchmarks accidental, Kagi!
https://blog.kagi.com/kagi-assistants #KagiAssistants #AIResearch #HumanCentered #TechInnovation #HappyAccident #HackerNews #ngated

Stanford research reveals AI's critical challenge: Distinguishing beliefs from facts. Significant cognitive limitations identified in language models across professional sectors. #AIResearch #Technology

Adversarial Poetry as a Universal Single-Turn Jailbreak Mechanism in LLMs

https://arxiv.org/abs/2511.15304

#HackerNews #AdversarialPoetry #LLMs #Jailbreak #AIResearch #NaturalLanguageProcessing

Have you seen this plot? Before panicing and claiming “half the Internet is AI-generated,” let's look at the data: current detectors often mislabel human text as #AI (≈1 /3 of cases) and their accuracy can swing from 100% to 0%. Until detectors actually work, we can’t estimate anything reliably.

#AIResearch #generativeAI #LLM

DOIs: 10.1080/08989621.2024.2331757 and 10.1177/02537176241247934

This website illustrates nicely how the US lost the competition–at least for now–in open(ish) LLM models: https://www.atomproject.ai/.

What it fails to mention is that the new open(ish) models are, more or less, on par in their capabilities with closed models as well.

It'll be interesting to see what the consequences are, but my money is on the open models fostering innovation, and not on closed models winning the race for AGI. #AIResearch #AGI_hype

/via Wired

Huyền thoại AI, "bố già AI" Yann LeCun, tuyên bố sẽ rời Meta vào cuối năm nay để thành lập một startup mới tập trung vào trí tuệ máy nâng cao (Advanced Machine Intelligence). Dù rời công ty, ông khẳng định Meta sẽ tiếp tục là đối tác chiến lược trong hành trình mới.

#AI #ArtificialIntelligence #YannLeCun #Meta #Startup #TríTuệNhânTạo #CôngNghệ #MachineIntelligence #AIResearch #NghiênCứuAI

https://vietnamnet.vn/bo-gia-ai-yann-lecun-roi-meta-2464590.html

#LeCun “We’re missing something big.”
My current work points to slow, phase-coherent transitions inside representation dynamics—far beyond scale.
#AIResearch #EmergentIntelligence #PhaseTransitio