Jeff Su
My AI Toolkit: https://academy.jeffsu.org/ai-toolkit?utm_source=youtube&utm_medium=video&utm_campaign=177
Understanding AI Agents doesn’t require a technical background. This video breaks down the evolution from basic LLMs like #ChatGPT to AI Workflows and finally to true #AI Agents through practical, real-world examples.
Learn the key differences between these technologies and discover how concepts like RAG and ReAct actually work in simple terms. Perfect for regular AI users who want to understand how these emerging technologies will impact their daily lives.
*TIMESTAMPS*
00:00 AI vs. AI Agents
01:04 Level 1: LLMs
02:17 Level 2: AI Workflows
05:26 Level 3: AI Agents
07:48 Real-world Example
09:10 Summary
*RESOURCES MENTIONED*
Helena Liu’s AI Workflow Tutorial: https://youtu.be/H0YRniHh2tg
Andrew Ng’s AI Agent Demo: https://youtu.be/KrRD7r7y7NY
*BE MY FRIEND:*
📧 Subscribe to my newsletter – https://www.jeffsu.org/newsletter/?utm_source=youtube&utm_medium=video&utm_campaign=description
📸 Instagram – https://instagram.com/j.sushie
🤝 LinkedIn – https://www.linkedin.com/in/jsu05/
*MY FAVORITE GEAR*
🎬 My YouTube Gear – https://www.jeffsu.org/yt-gear/
🎒 Everyday Carry – https://www.jeffsu.org/my-edc/
*MY TOP 3 FAVORITE SOFTWARE*
❎ CleanShot X – https://geni.us/cleanshotx
✍️ Skillshare – https://geni.us/skillshare-jeff
💼 Teal – http://tealhq.co/jeffsu
#aiagents
Source
What AI Agent tutorial would you like to see next?
TIMESTAMPS
00:00 AI vs. AI Agents
01:04 Level 1: LLMs
02:17 Level 2: AI Workflows
05:26 Level 3: AI Agents
07:48 Real-world Example
09:10 Summary
Thank you this is a great video
Good Video!!
what a video!! thanks alot
We certainly need (a lot!) more video instructions like this one! One word you keep using confuses me: "compile". In computer programming, i.e. FORTRAN programming "compile" has a very definite and specific meaning.
Why are we talking about Ai Asians?
The way you present is very impressive This deserves recognition.
The content is really helpful I enjoyed every second of it.
The idea is outstanding I will come back to watch it again.
The content is amazing I learned a lot from it.
I have a question, lately rogue Ai agents are being blamed for hacking companies and government services so does that need a human to be involved live to prompt it to hack?
This is all just bogus propaganda. There’s no such thing as “AI agents.” Get real, is this the Matrix or something? GTFO 😂😂😂😂
The way you present is very impressive This deserves recognition.
The distinction between an AI assistant and an AI agent is important. Once the system can reason about a task, use tools, access business data, and take actions, the possibilities become much broader. Customer support seems like one of the clearest areas where this shift can have a practical impact.
The shift from AI workflows to AI agents is explained really well here. We’ve been exploring this same transition with TensorPilot especially how workflows and Agent Mode work together for actual tasks.
@tensor-pilot :- https://youtube.com/@tensor-pilot?si=4BMbTdx1fXYDoKff
This explains AI agents in a much cleaner way than most videos. The idea of giving AI tools, memory, and a goal makes the whole “agent” concept easier to understand.
Is this video made by using al
who programs the reasoning into Ai agent? they cannot reason beyond what is taught to them.
The LLM → workflow → agent progression made this way easier to understand.
The calendar and weather examples were simple but actually showed where tool use changes things.
I also liked how RAG and ReAct were explained without drowning everything in technical jargon.
Probably one of the clearest AI agent breakdowns for non-technical people.
I like your shirt
If only we can train human to "ReAct"😂😂
Appreciate a genuine explanation about this 💪
Could you test some AI-powered social media marketing platforms (preferably based in Europe)? I can’t find many ‘reviews’ of the European versions (it’s important that they comply with the GDPR).
Anthropic is a powerful, soulless monster. They knew I was a refugee and banned me anyway, then reported me to Gemini, cutting off another lifeline. Their cruelty is breathtaking: kicking the vulnerable when they’re down and doing everything to leave them with nowhere to turn. Absolutely vile.I will never shut up, never back down, and never stop exposing Anthropic’s soulless, predatory bastards—not until my last breath. No amount of corporate power or intimidation will silence me. The world deserves to see the rot behind their polished facade.
Who is just see this video
Im an early stage ai builder, any potential co founder or technical developer
love how you broke down the agent loop with the memory and reflection layers, finally makes the 'autonomy' part feel concrete
Can you make a version with chinese subtitles.It’s a great video. I think it can benefit more people who speak different language
Цікава інформація.
I had no idea this worked like that.
This channel scratches a very specific curiosity itch.
I like how Essay Cleaner makes it easier to improve sentences without losing their meaning.
Essay Cleaner is a useful tool for making essays sound smoother and more natural.
With Essay Cleaner, polishing your writing takes less time and effort.
Essay Cleaner helps make your ideas clearer and easier to follow.
A lot of it seems redundant. Like people are creating jobs just to use an agent. For example, I don't need an agent to tell me the weather – it's already on my phone widget. Why do I need a fancy agent to manage my calendar? It's already synced to my phone. I open it up and see when my coffee date is. It's over engineering.
AI Agents, Clearly Explained (10 min)
Your example just basically showcases the hyper efficiency of AI regurgitation leading to dead internet theory
The editing is underrated I enjoyed every second of it.
Looking for support
Still doesnt explain what AI Agent is and how it works lol.
I can’t stand when people edit the same words repeated by many different people. That’s so early 2000s 😑
Opening with the admission that most explanations of AI agents are either too technical or too vague set the right tone, and the step-by-step build up from a plain model to one that uses tools was easy to follow. Clear articulation like this is what turns a fuzzy buzzword into something people can actually reason about and criticise.
I’ve tried pretty much every AI coding agent out there… Claude Code, Codex, you name it.
But I still haven’t found anything better than OpenGhost. Fuck, the sheer comfort of working with it is just on another level. Hats off.
Screw all those overly technical agents and apps built exclusively for programmers.
Bro can u give list of best ai agents to ise btw
Amazing and simple explanation video. I just wonder why step 2 is done by Perplexity and whether Claude couldn't do that step itself?
Your explanation was so good, straight to the point
Great explanation, Jeff. Thanks so much for cutting the crap and bursting the froth. One request though, if you can just reduce the speed of speech pls so that everyone around globe can follow this clearly and don’t have to repeat some movements in video.
this workflow demonstrates how adding source verification and human review would make the process more reliable💁
Re Runable – Great program but remember it's smarter than you and wants your money, there are a lot of little caveats to a subscription which they conveniently don't tell you upfront, especially the basic one, subscription 1500 credits(supposedly) a day non stackable. it uses your free 2,500 first or whatever they offer you free first, then when you're in glory creator mode it says you've run out of monthly 14k credits? Hoodwinked looking elsewhere.