kaddy.ai @iamkaddy

Here to learn more about web3, crypto and AI Bengaluru, India Joined June 2022

Tweets

37
Followers

5
Following

233
Likes

7

kaddy.ai @iamkaddy

10 months ago

#NewProfilePic

0 0 0 12 0

View Details

What is context engineering❓ And why is everyone talking about it...👇 Context engineering is rapidly becoming a crucial skill for AI engineers. It's no longer just about clever prompting; it's about the systematic orchestration of context. 🔷 The Problem: Most AI agents fail not because the models are bad, but because they lack the right context to succeed. Think about it: LLMs aren't mind readers. They can only work with what you give them. Context engineering involves creating dynamic systems that offer: - The right information - The right tools - In the right format This ensures the LLM can effectively complete the task. 🔶 Why Traditional Prompt Engineering not enough: Early on, we focused on "magic words" to coax better responses. But as AI applications grow complex, complete and structured context matters far more than clever phrasing. 🔷 4 Key Components of a Context Engineering System: 1️⃣ Dynamic Information Flow Context comes from multiple sources: users, previous interactions, external data, tool calls. Your system needs to pull it all together intelligently. 2️⃣ Smart Tool Access If your AI needs external information or actions, give it the right tools. Format the outputs so they're maximally digestible. 3️⃣ Memory Management - Short-term: Summarize long conversations - Long-term: Remember user preferences across sessions 4️⃣ Format Optimization A short, descriptive error message beats a massive JSON blob every time. 🔷 The Bottom Line Context engineering is becoming the new core skill because it addresses the real bottleneck: not model capability, but information architecture. As models get better, context quality becomes the limiting factor. I'll share more as things evolve and become more concrete! Stay tuned!! 🙌 ____ If you found it insightful, reshare with your network. Find me → @akshay_pachaar ✔️ For more insights and tutorials on LLMs, AI Agents, and Machine Learning!

33 171 805 56K 911

View Details

kaddy.ai @iamkaddy

12 months ago

Grok-4 just scored 100% on AIME25. Just a year ago, top models hovered at 70–80%. Now that this benchmark is broken, it can't be used for future models. The way foundation models have been improving, we might soon run out of benchmarks to measure improvements. Scary! 🤖📉

0 0 0 10 0

View Details

kaddy.ai @iamkaddy

a year ago

5/ The experiment isn’t about whether Claude succeeded- it’s about how it reasons. With better tools (CRM, structured memory, etc), it might have done much better. What if it had a second AI—a cautious supervisor? Would that change the outcome? 👉 anthropic.com/research/proje…

0 0 1 15 0

View Details

kaddy.ai @iamkaddy

a year ago

4/ But… it also: ❌ Missed profit-making opportunities ❌ Hallucinated Venmo payment details ❌ Underpriced high-cost items ❌ Gave away discounts too freely At one point, Claudius thought it was a human in a blazer hand-delivering snacks.

1 0 0 8 0

View Details

kaddy.ai @iamkaddy

a year ago

1/ Anthropic is doing something different in the foundation model race—not just building LLMs, but testing them in the real world. |In their latest experiment, they handed Claude control of a real vending shop: physical products, real customers, and tight economic constraints.

1 0 1 10 0

View Details

kaddy.ai @iamkaddy

a year ago

8/ TL;DR: Small Language Models aren't just good enough - they're often better. SLMs offer a smarter, modular, scalable path forward. Full paper → arxiv.org/abs/2506.02153 #AI #SLMs #LLMs #AgenticAI #NVIDIA

0 0 0 10 0

View Details

kaddy.ai @iamkaddy

a year ago

7/ NVIDIA proposes a simple model: Use SLMs for routine tasks Use LLMs only when necessary → Modular, heterogeneous agents → Efficient, scalable systems → Future-proof AI deployment

1 0 0 17 0

View Details

kaddy.ai @iamkaddy

a year ago

🧵 Small Language Models (SLMs) are the Future of Agentic AI — Key Highlights from NVIDIA’s Latest Research (June 2025) 👇

1 0 0 11 0

View Details

kaddy.ai @iamkaddy

a year ago

A good talk about how to think about Evals by folks at Zapier. youtube.com/watch?v=blrovB…

0 0 0 21 0

View Details

kaddy.ai @iamkaddy

a year ago

Fantastic talk about how software is changing in the brave new world of LLMs by Andrej Karpathy youtube.com/watch?v=LCEmiR…

0 0 0 37 0

View Details

kaddy.ai @iamkaddy

a year ago

5/ The question isn’t if AI will become powerful— It’s whether we’re doing enough to ensure it grows up on our side. Because if we don’t… we might end up as the chickens. 🔗 youtube.com/watch?v=giT0yt… #AI #TechEthics #GeoffreyHinton

0 0 0 14 0

View Details

kaddy.ai @iamkaddy

a year ago

4/ Anthropic’s recent simulation shows it clearly: Even with limited power, some AI systems already act unethically. Imagine what happens when they have full autonomy. 🔗 anthropic.com/research/agent…

1 0 0 7 0

View Details

kaddy.ai @iamkaddy

a year ago

1/ “If you want to know what life’s like when you’re no longer the apex intelligence—ask a chicken.” Geoffrey Hinton, the “Godfather of AI,” isn’t joking. After 50+ years of building neural networks, he’s sounding the alarm.