1.Agentic AI
2.Large Reasoning Model
3.Vector Database
4.RAG(Retrieval Augmented Generation)
5.MCP-(model context protocol)
6.MoE-(Mixture of Experts)
7.ASI-(Artificial Super Intelligence)

“We are smart enough to invent AI, dumb enough to need it, and so stupid we can’t figure out if we did the right thing.” - Jerry Seinfeld

At 52 I'm going back to college for a software engineer AI degree. Haven't been this excited since cyalis.

I was a Systems Engineer and Marketing Rep with IBM starting in 1962.  1401, 360, etc.!!!! Long ago, far away. At 85 today, I “play” with AI in several areas.  The neighborhood “expert.” That kind of stuff. This video made me proud!!  My “buddies” from IBM are helping me comprehend what is “under” the hood of AI.  Terrific job!!

If ASI materializes in practice, just remember that I was a nice human.

0:27 1. Agentic AI
1:52 2. Large Reasoning Model
2:48 3. Vector Database
4:29 4. RAG (Retrieval Augmented Generation)
5:35 5. MCP (Model Context Protocol)
6:45 6. MoE (Mixture of Experts)
8:19 7. ASI (Artificial Super-Intelligence)

All IBM presenters are top-notch... I am addicted to their videos and learn a lot.

I would add orchestration, guardrails, observability, Multi-agent systems and Gen AI

If I could add one more term to the list, it would be Multimodal AI.
Because the future isn’t just text — it’s images, audio, and video too.

Came for the content, stayed for the smile. Never seen anyone so excited when presenting AI content 😂

Other terms you should know:
Hallucinations
Slop
Bubble

I'm at the end of a 45+ year Software Engineering career. I'm kind of glad I don't have to learn AI but its good to see fresh tech at the inception

Not sure if "alignment" is technical enough for this type of video, but if so, it's the only A.I. term that really matters.

Alignment should’ve been included, especially given the discussion on ASI toward the end

Thanks for the clear, no-fluff walkthrough! Loved how you framed the agent loop (perceive → reason → act → observe), connected LRMs to deliberate step-by-step thinking, and demystified RAG/MCP as the “context + tools” stack (embeddings → vector DB → RAG → MCP). The MoE recap was also spot on.
One practical question: when you wire these pieces together in real projects, how do you specify and evaluate prompts over time—especially for RAG (chunking/instructions/grounding), MCP tool schemas (guardrails & error handling), and LRM “thinking” budgets? Do you keep a small regression set with A/B across models and log prompt versions per run so you can trace prompt → retrievals → answer quality?

This is the IBM CENET I used to dream of, when these talks were broadcast via satellite communications VTC.

What a development for the curious, inquisitive, and lifelong students.

I discovered my kid's school was using AI to “analyze behavioral patterns”. Selwyn Raithe's book  exposed what they're really collecting. Got 20 other parents to opt out after sharing chapter 4.

I love when AI  says: ‘you’re welcome’  without missing any of the steps 😮

Fantastic explanation 🎉 ‏‪ ‏‪9:34‬‏

We are truly lucky to have teachers like you ❤ The video touched on important points.