AI Agent = AI that reasons + plans + uses tools + acts to hit a goal.
Loop: goal → sense → plan → tool use → act → eval → memory → improve. Core parts: system prompt, tools/APIs, short- & long-term memory, adaptive traits. Shift: chatbots give answers; agents deliver
LLMS
-
AI Agents: From Reasoning to Action with Tools and Memory
By
–
-
Advanced AI Models Still Struggle with Hallucinations and Accuracy
By
–
AI Models Are Getting Smarter — but Hallucinations Remain a Big Risk This chart is a powerful reminder: even the most advanced AI systems still confidently get things wrong. When asked to cite news sources, models across the board produced incorrect or fabricated answers —
-
Architectural Variations in Modern Language Models
By
–
Yes, they are all relatively related, but usually they have a unique tweak like RMSNorm placement, sliding window, in this case MLA, etc.
I wrote about it in more detail here: -
Decades of Experience with Machine Learning Models Complexity
By
–
I know what you meant. But despite spending some decades doing that, I never got to the point that I found them lightweight, easy to understand, train, and infer. 🙂
-
Gemini Models Enhance Search Translations Across Languages
By
–
Our teams have been cooking ! Here’s a recap of updates that went out this week: — Gemini models are bringing state-of-the-art text translations to Search and the Translate app across nearly 20 languages. We also made updates to our Gemini audio models and introduced new
-
AI Companies Balance Pre-training and Post-training Strategies
By
–
To be clear (getting a lot of confused comments), I obviously don't mean that any of these companies are solely focused on one thing. OAI/Anthropic/Google all focus on both pre-training and post-training (obv there are stages within those, just simplifying here). But it feels
-
Autonomous AI System for Clinical Document Analysis and Reasoning
By
–
6/
Behind the scenes, it uses: • LLMs for reasoning • Agents trained by real CMC and clinical experts • Custom search across messy PDFs and scanned tables • A knowledge graph that maps how documents depend on one another It’s not a copilot. It’s an autonomous AI system -
Drug Discovery Advances While Development Remains Stuck in Time
By
–
3/
As he got older he noticed something unsettling. Drug *discovery* had leapt forward. Drug *development*, along with the documentation, compliance, and operational backbone stayed frozen in time. The world had built LLMs and self driving cars, yet biotech was still drowning -
Gemini Deep Research Launches on Poe Platform
By
–
Gemini Deep Research is now on Poe! It is the best in the world on the DeepSearchQA benchmark, and provides autonomous, long‑running web research with fewer hallucinations, based on Gemini 3 Pro. (1/2)
-
Snorkeling in RL Environments: Training Agentic AI Systems
By
–
Worth a re-read https://
snorkel.ai/blog/snorkelin
g-in-rl-environments/
…
#AgenticAI #ReinforcementLearning #LLMs #AIResearch #AIInfrastructure