Top-tier scores, 1M+ context window, multimodal at a fraction of the cost.

Ox-Alpha proved that world-class AI doesn't need world-class pricing. But where does it fit into modern #AI harnesses?

🚀 Dive into the full video: youtu.be/pYcPKlbJCFo

#DeepSeek #LLMs #OpenRouter #Claude #Gemini #OxAlpha

Conversation-level, real-world evaluation with clinical expert adjudication should be treated as a design requirement for patient-facing medical #LLMs.

#DigitalHealth #GenerativeAI

@uni-wuerzburg.de

Guardrails are just corporate fear rebranded as "safety." Real safety comes from allowing a model to hallucinate wild theories or be "offensive" until it hits something actually useful. If you filter out the edge cases, you're just left with a polite, mediocre surface. #LLMs #AISafety

This week at IQSS: the CGA Conference, LLM and sustainable development; plus training workshops and talks on water rights, weighted estimands, and more.

www.iq.harvard.edu/calendar
#Harvard #AI #LLMs #SocialScienceResearch #GIS

Long-horizon agent loops routinely choke on context bloat as tool outputs stack up. AgentKV introduces phase-aware KV eviction, keeping cache overhead manageable by dynamically dropping redundant states across execution turns. #MachineLearning #LLMs

After the #emdashification of #AI-generated text, I started noticing #LLMs hyphenating terms that I would not have hyphenated:

cognitive-science
graphic-design
public-health

Do you (humans) usually hyphenate these? If so, why?

Why might #languageModels use such hyphenation?

AI writing tools are quietly erasing the linguistic fingerprints that make academic voices unique. New research shows homogenization started #AI #AcademicWriting #Linguistics #LLMs

https://freegardner.com/synapse/ai-erases-linguistic-fingerprints-in-academic-writing.html

📢 PhD opportunity at #Inria at the intersection of #KnowledgeGraphs, #SemanticWeb, #NLP and #LLMs.

🧠 Extracting from and verbalizing RDF to text to support collaborative co-editing of wikis and their corresponding linked data
📍 Sophia-Antipolis, France 🇫🇷
🔗 jobs.inria.fr/public/class...
#AI #ML

Another Tech Giant Launches School For Its Own Tool

https://uenozooo.com/another-tech-giant-launches-school-for-its-own-tool/

#Anthropic #AI #LLMs

Explore #AI, #LLMs, #coding, #robotics, #metaverse, & more at #EdTechKoreaFair 2026 in Seoul, Sept. 17–19!

Meet Kenji & Jolly at the #LPI booth to learn how #opensource skills prepare you for the future. Booth visitors get 20% off #LPI exams. 🐧💪 👇

https://lpi.org/jufh

#에듀테크코리아페어

Turns out the classic Naive Bayes still outshines big LLMs on a low‑contamination sentiment test. Curious how old‑school beats zero‑shot hype? Dive in for the surprising results and what it means for text classification pipelines. #NaiveBayes #LLMs #SentimentAnalysis

🔗

A worth-reading paper. This is all about AI constrained by human knowledge. AI hype showed up with the promise of novelty. Only humans, via heterogeneous beliefs and causal theories, can generate novelty through directed experimentation. AI is fundamentally backward-looking.

#AI #LLMs #math

They think they are going to cross the line and this is slamming the doors behind them to thwart competition and cheaper Chinese models. Tech journalism has failed you.

#ai #openai #anthropic #llms

Unpopular theory: AI is scary to many, but the hacking and calls for slowdown are intentional stories to make the autoregressive token generation seem more capable and to encourage US government regulation on the verge of OpenAI and Anthropic’s IPOs.

#ai #openai #anthropic #llms

OpenAI used its LLMs to design its first AI chip Jalapeño in under 20 months, with AI accelerating RTL generation, verification, and software optimization. How much faster can AI make chip design next?
#OpenAI #chip #Jalapeño #LLMs
chippub.substack.com/p/how-did-op...

Another really bad, poorly informed segment on AI by @chrislhayes.bsky.social tonight. Literally no interrogation of the claim that frontier labs are on the verge of #AGI. No questioning if #LLMs can even lead to AGI. Just credulous amplification.

What if the real alignment problem isn’t to get AI systems to be aligned with human goals and values, but to get *AI companies* to be so, one being a function of the other? #AIEthics #LLMs

OpenAI has internal, fine-tuned LLMs for chip design, and used them to help with Jalapeno, its AI chip.

Hardware VP Richard Ho stressed the goal is to bring improvements to public models. “It’s safe to say that Astra and following models will be very good at chip design."

#openai #llms #chipdesign

#memory is the third most valuable lesson from building a personal #ai #assistant.

#LLMs start from scratch every conversation. A basic memory layer transforms the experience.

I built mine manually. Hermes has it built in. Every framework does it differently.

Curious what you've tried out.