🐙 Essential AI Skills For 2026

Hello friends, how are you doing?

I am doing good. Thank you.

Hey Richard, how's it going? Golden

Dumpling. Hello. Hello Anthony.

Yep, she fixed the email. Yep, I fixed

the email.

Hello.

Is she delayed again? I'm fashionably

late. I'm fashionably late. Not delayed.

But hello, Cypher. How are you doing?

Hello. Okay, I'm going to turn my audio

up a bit.

Good morning, good evening, good

afternoon. Hello everybody. How's

everyone? What can we do today? What can

we do today? We can do many things

today. The day is full of potential.

Where would you what can what are we

doing today? Well, we are going to be

covering essential AI skills for 2026.

Who Who here is like ready for 2026?

I think I am. I I think I'm like very

ready for 2026.

Um

yeah, I just am. Hello from Shanghai.

Oh, hello. Hello from Hong Kong.

Hello there. Me from Hong Kong. Me, too.

I'm in Hong Kong right now. It's 2 a.m.

in South Africa. Oh my god. Go to sleep.

Go to sleep. Come back. Um, come back

for the replays. But thank you so much

for being here. I do appreciate it a

lot. Morning from Nigeria. Good morning.

I want to be ready. Why are you not

ready for is anybody okay? Who is ready

for 2026? Put into the chat if you're

ready. And if you can also write either

put ready or not ready. I'm curious

about people's readiness level. Do that

while I open up the slides.

Hello from Canada. Hello. I'm still in

2023. Yeah, true. I would like to be on

2023. 2023 would be nice if I get two

years. Like three more years. Two more

years. That that would that would be

great. Ready. Ready. You ready? Not

ready. Not ready. What means ready? Kind

of ready. Just like ready for 2025 to be

to be finished like concluded. Yes.

Not ready.

Okay. I think that should be good.

Not ready. Ready for

I don't know how to pronounce that word.

Are the Italian dumplings? I don't know

how to pronounce. Not ready. Ready.

Ready. Okay. Okay. Well, you know what?

Whether you like it or not, it's

happening in 10 days. 11 days. 11 days

for me, 12 days for other people.

Anyways, I stop rambling. Well,

today we're going to be talking about

the essential AI skills for 2026. Okay?

Cuz going to 2026, it's actually really

interesting in my opinion. There's

there's um I feel like yeah like the AI

world for the end of 2025 has been

slowing down a little bit like we get

like Gemini 3 release, we get a few

other things. Uh but I think there are

trends that are coming out which I think

will be very interesting going into

2026. So that is what I'm going to be

covering today. All right, let's go. Let

us go. Okay,

where's my slides? Too many tabs. Okay,

perfect. Great. Okay. So, what we're

going to cover today is why AI skills

are non-negotiable these days. I don't

think I need to convince you guys you

guys are here. Um, so four essential AI

skill pillars, prompt engineering, AI

coding tools that you need, AI agents,

open-source.

So, this is this is a this is one that

I'm actually pretty excited to chat

about. Open source AI, critical

workplace skills, and then career impact

and next steps. And then you know

throughout the presentation

presentation talk I don't know what this

is live stream live stream please you

know don't leave me just just rambling

about stuff talk talk to me ask

questions please do okay so um the first

thing is that I I really hope you are

all like convinced at this point that AI

skills are really no longer optional in

my opinion if you're a modern human that

participates in society at this point if

you don't have AI skills you're going to

have problems s because most people

should be at this level right now. Um,

and I feel like there's like this

unspoken expectation at this level in a

lot of workplaces. Um, so if you don't,

I do think that

it will be hard for you to function um

well in society and then just like be

able to produce the things that you need

to produce uh without AI. So I really

think so. In fact, 81% of business

leaders expect AI to be deeply

integrated in their operations within 12

to 18 months. So that is fast. Three

times productivity increase reported by

developers using AI tools. 89%

organizations already leveraging AI in

some form. So really not surprising

anymore. um to incorporate AI into into

pretty much every part of life uh like

in in terms of your work and in terms of

and in terms of your own uh productivity

as well. So what this means for you is

that if you don't have AI skills, you

are falling behind peers and

productivity, missing out on career

opportunity, struggling with repetitive

tax and limited job market

competitiveness. With AI skills, we can

see that you'll be working three times

faster on key task wage increase

potential. We see a lot of job postings

now do have that requirement of certain

AI skills now. Um automating mundane

work and then future proofing career

trajectory as well. So um I do have some

additional resources here which you guys

can feel free to click in and just like

understand things better and oh I right

if you have not um signed up already

like for our newsletter you can you will

get this for free like um you I'll send

you guys I'll send you guys these these

things for free. So, please do so. I'mma

pin pin the comment here. Yeah, pin the

message. Um, so yeah, please sign up if

you do want to get these slides. Oh, I'm

going to like move my face because

that's pretty annoying that my face is

covering stuff

there. Maybe that's better. Okay, great.

Good.

Let's see if anybody has any comments.

Her live streams are always available.

That is true. I thought you were going

to give us an agent to track our

finances. Did I ever say that? You

gaslighting me? I don't think I ever

told said I was going to give you an

agent track your finances. [laughter]

Did I?

I I am into that right now though. Um I

actually redoing I kind of redid my

entire kind of investing strategy as

well. Um but anyways, we can talk about

that later. Uh right. Okay. So the four

essential pillars of AI skills what in

my humble opinion I believe that you

absolutely need to know at this point um

in 2026 would include um prompt

engineering. So just this is like the

defining career. This is the number one

skill um defining careers is like

literally the basics of everything. I'm

going to be covering all this in a lot

more detail as well. So [gasps] I'm

going to go over them briefly. So uh

prompt engineering so clear specific

instructions AI context and constraint

specifications iterative refinement and

building prompt libraries. So really

like the difference between good a good

AI results and really great AI results

do come down to the prompt these days.

It's not really about the tool that

you're using often times. It's just how

it is that you're using it. And

prompting is kind of like the language

of AI. Um yeah it's like it's like you

can have a lot of different tools

available to you but you don't know how

to prompt correctly. you don't have that

skill, then you're going to have um

you're not going to be able to get the

results that you want. So that's why

it's like literally the most important

skill. There's like a single skill that

you want to master um is is going to be

prompt engineering. And then there's AI

tool fluency. So mastering tools that

transform different industries. I do

think that there's like way too many

tools. And I actually tend to not

emphasize tools themselves that much,

but I think if you have like a

generalized

chatbot that you like like Chad Beauty,

Claw, Gemini, Perplexity, you know, like

general um tool that you like, it does

cover a lot of a lot of things that you

need that you would want to cover. But

um if you do want to really like use AI

um and really get the best results,

there are like certain tool kits. I kind

of call it like the toolkit of the

modern human as well that you can choose

to have like certain um AI tools that

you use like for me I have like a

certain stack stack of AI tools that I

use on a day-to-day basis but this would

be like AI writing content tools AI

writing assistance like if you are

someone who codes so cursor winer things

like that and industry specific AI apps

as well next is AI agents like this is

undeniable the future of work is agentic

AI in fact the now of work is also kind

of mostly agentic AI now I think I think

for like at this point if you haven't

encountered agentic AI in some form

already maybe you might not know it but

I I would be really surprised so it's

really important to understand like

agent capabilities building simple

automation workflows like this is where

um it it doesn't sound like super sexy

to be building automation workflows like

agentic workflows but there's so much

potential in this and there's so much

that's coming out of it right now we're

starting to like see the results of of

agentic AI

um integrated into the workplace. Um

agent orchestration basics and real

world agent applications. So this is a

shift from AI that suggests things to AI

that actually does things autonomously.

That's what AI agents are. And finally,

responsible AI use critical thinking. So

if you guys have any have used um say

like Sora for example, right? like like

a lot of different um just this AI is

very powerful and it has like a lot of

things that will come out with but at

the same time there is that other side

of the other side of things where if you

are not responsible and is using you

can't understand like how to use AI

properly I think it would also be really

it would be really challenging for you

so that's why it's important to be able

to evaluate AI outputs for accuracy

understanding bias and limitations um

things like data privacy and security

and ethical AI decisionmaking

So the the truth is that AI does make

mistakes still and that is the case and

your job is to be able to catch these

mistakes and work with AI cuz in the end

like AI is still a tool. It's not like a

magical solution that's going to fix all

your problems unfortunately. Um yeah

unfortunately on that one. Yeah. So

these are the four essential pillars of

AI skills.

See if anybody has any comments

before I go into prompt engineering.

Um, hi for Uruguay. Urg, I would like to

know advanced rag blueprints and

techniques.

Abluma. Okay.

Uh, we do talk about that in our agents

sections. But if you have any like

specific questions you want to ask me um

about that, I can answer those questions

about about RAG specifically. But we do

cover them as part of like we don't have

like a rag specific thing but I do cover

them in in like our agents boot camp and

things like that. So I'm happy to answer

any specific questions that you might

have. Can you just ask them to prompt

themselves? No, you you actually cannot

ask them to prompt prompt themselves.

Like it's still at this point you still

need to give direction. And the problem

isn't because like it can't prompt

itself. The problem is that you can't if

you can't articulate what it is that you

want, how are you going to like that's

the problem, right? Like it's it's not

that like AI can't prompt itself. is

that you need to articulate what it is

that you wanted to prompt and that's

like communication skills. Um just like

how you communicate with humans, you

need to learn how to communicate with AI

as well. My company has taken THBT

license and ask everyone in company to

ask at least two questions every day to

it can be anything. Yeah, I mean I'm not

surprised. I I think there is like an

aggressive push towards that and it

makes sense. It genuinely does. Um yeah,

I think more and more companies are are

doing this and they probably will

continue to do this as well. Oops.

Okay, let me move my face a little bit.

I feel like it's a little bit annoying.

But there uh see where can we download

the material. So I have it pinned over

here. So if you click here, uh it's a

mailing list and then I will email you

all these resources like the Yeah, I

I'll email you the slides after the

workshop. Workshop. I keep like changing

the name. Live stream live stream. I've

been teaching too many workshops

recently. Yeah, but it uh just just sign

up there and you'll be good. Yep. The

fear is real. Tina, hope that you would

could explain more about AI agents. I

still don't know where to start. Could

you recommend? Yes. So, I will be

covering AI agents um as part of this uh

live stream in a bit. We'll be getting

there. Thank you for your contributions

to our learning and studies. Tina, thank

you so much. Appreciate it. Oh, I

thought you were on to context

engineering over prompt engineering. I

mean, prompt engineering context

engineering. It's prompt engineering is

kind of like context engineering is like

the evolution of prompt engineering but

in the end it's still about creating a

good prompt right it's still creating

either you're talking to AI directly

you're using it as part of a gentic

workflow it's still going to be about

the prompt itself so I'm just going to

call it prompt engineering just cuz that

covers all kinds of prompting but yeah

context engineering is just a form of

prompt engineering but specific for

building products and agents

all right

let us talk about mastering prompt

engineering this is like the fundamental

skill that you really really need to do

um is prompt engineering. So in 2026

prompt engineer is essential as writing

a good email. Yeah, I really do believe

that um you need to in order to be good

at prompting you need to be specific. So

vague prompts can get vague results. You

need to iterate because in the end like

you need to keep refining this output

over and over again to make it better.

And then building a library. So this one

uh just saving prompts that work. Don't

start from scratch each time. Like this

is personally I have like a few

templates like that I like to default

to. Um and over time you like you start

building up a good skill of how it is

that you can prompt. So somebody asked

earlier like can you get AI to prompt

itself. What I generally do is like if I

I can figure out what are the key things

I need in a prompt. I would actually put

that into the prompt and then ask like

something like Gemini chat beauty cloud

whatever in order to amplify that like

in order to um make it better as a

promp. So you can do that, but then the

beginning part of it, it's still really

important um for you to actually have

that base prompt. So this is the six

part prompting framework that I like to

use in terms of products. So um if

you're just talking to CHPT, for

example, like it's not you don't need to

be as specific um as as you do for this

one, but I want to give you like a

generalized prompting framework that

does cover um this this you can use for

like different products for if you're

building products. It can be like when

you're building agents as well. So this

is literally the six-part prompting

framework that I recommend for building

AI agents too. So I kind of want to give

you this structure whenever you're

trying to come up with a with a good

prompt. So the first one is a role. Um

you are describe the perspective

expertise needed. Task is your main job

is to whatever task it is that you want

to accomplish. The input is I will give

you blah like whatever it is you're

going to give it. Output is you should

respond with desired format style and

structure. Constraint is never do

certain things that you should not do.

and reminders always remember to etc etc

etc. So again this is kind of like over

overkill if you are um just kind of

prompting Chachi BT directly like this

but generally speaking like if you are

able to figure out what are the things

that go into this prompting framework uh

what like at least think through these

things then you would have a much better

result. So an example of here would be

like writing a business email. So you're

a professional business communication

specialist draft follow-up email to a

client after our first meeting input. I

will give you just key discussion points

for our meeting output. Respond with a

polish email uh with professional

greetings, meeting recap, blah blah

blah. Clear next steps, friendly closing

and constraint. Never use overly casual

language, jargon with explanation or

make promise I don't authorize.

Reminder, keep tone warm or

professional. Proof read for clarity and

ensure action items that are uh action

items are specific with deadlines. And I

would also recommend putting some

examples here as well if you it's if

it's something that is going to be like

writing a business email. So, uh I do

have like entire videos dedicated to

promp prompting like both for prompting

in general and then also for prompting

um for prompting for agents and

products. So, there's like more

frameworks that I do have. But I think

if you're going to like remember a

framework uh that would be pretty useful

and you can start when if you're

building your own products as well, it

would be this one. Recommend you

remember that. So, yeah. Um, so the pro

tip here is that, you know, something

like this framework template, you can

just save it as a note or a doc and

customize it for your different kinds of

task. And this is the one that I like to

personally use the most when I'm

building products.

Okay, so not going to go into too much

more detail about prompting. Um I can as

part of the resources I'll send you guys

uh after this live stream I'll also send

you guys like a couple videos that I

have both on like prompt engineering in

general as well as how to prompt like in

terms of in the context of building

products. How many of you guys are into

building AI products or have already

built an AI product? When I by AI

product, I mean um anything that you're

like anything related

that you're building whether that's an

agent or it's just like a AI product or

it's LM app something like that. How

many of you guys are building or

interested in building a product? I am

curious. Let me know in the comments in

the chat.

Uh, okay. From Gabriel, should prompts

be tailored to each model to get the

best results? Great question. Great

question. I think when it comes to say

if you're just chatting with a chatbot,

I don't think it's really necessary to

do that. Um, however, if it comes to

building products, especially building

agents, then there are certain things

that that you would tweak um depending

on the model that you're using. Yes, I

don't consider that to be like the most

important thing. like the general prompt

itself it's good enough can get you most

of the results that you want. It's it's

just that if you really want to go from

like 80 like maybe like 90% to 95% or

95% to 98% then you might want to start

tweaking it a little bit. It's not like

the priority though

totally into building trying to build

building the infrastructure tried

haven't gotten anything to work properly

yet. Yes, that's why I'm here. I've

tried to build but I'm failing. I want

to build. Okay, great. Okay, let's and

then feel free to ask questions about

things that you're building

specifically. Um, I'm happy to talk

about that as well cuz I do think you

should for anybody that wants to be

building AI products at this point. You

can quote me on this one. You know what?

You can quote me on this one. If anybody

is interested in building things these

days, you absolutely can and you you

should not be limited by not knowing how

to code or or something like that. You

absolutely can at this point, which is

kind of magical.

Um, a integrated suspicious activity

monitor in Python to start by learning

the basics. Like what kind of product?

Um, never monetize build some agents for

myself. Maybe I should just follow a

template. Okay, cool. Yeah, we can talk

more about it, but I just think it's at

this point if you are interested, um, I

think there's a lot of things that you

can building be building and it's pretty

cool. Okay, so prompting. All right,

guys remember prompting. Prompting is

the number one fundamental skill. Okay,

now let's talk about vibe coding. So

building without code as I was saying

earlier I think anybody that wants to

build these days has the ability of

building. So what is vibe coding? So

this is a term coined by AI expert Andre

Kaparthy basically saying that if you

want to build things and you want to

code right now you can just describe a

natural language and AI can generate

working app. No coding is actually

required fully given to the vibes. He

says embrace exponentials and forget

that the code even exists. Um so in this

new paradigm shift in which you are able

to build things without code people who

are designers right like you're able to

become developers now like a designer

can describe their vision in natural

language and generate fully functional

website um or application and hand off

working code to developers. No back and

forth is really needed anymore. Um I'll

talk about like the role of developers

in just a little bit. I'm not saying

that it doesn't it isn't required but

really vibe coding democratizes the

ability to build and managers instead of

when if you're a manager you're able to

build prototypes now a product manager

for example can build a working

prototype to test ideas with users

before investing in development you can

validate concepts in hours not weeks

entre entrepreneurs can build their MVPs

now a founder with zero coding skills

can launch a functional SAS product and

I'm saying this like I literally know

people who have done this right test it

with your customers iterate based on

feedback and monetize

like single person like soloreneur being

able to build an entire company uh like

entire SAS product like this is very

reasonable like I actually know like a

lot of people when I say a lot like at

least it's like it's like I actually

know a lot of people who are

successfully doing this it's so possible

now which it would have been impossible

to do like no way you could have done

something like this even like last year

or the year before before vibe coding

tools and and the large language models

are are good enough but you can

Um and analysts you can do things like

building dashboards. So business

analysts can create custom data

visualization dashboards without waiting

for engineering resources. This is also

like completely doable now. So here are

like a couple different um vibe coding

tools for for nontechnical people that

you can check out if you like. Um

vibe coding isn't just about making

everyone a developer. It's about

democratizing creation. So nontechnical

people can build, test and iterate ideas

without depending on engineering teams.

And this is kind of huge. Um, I was

actually on a podcast yesterday. I don't

know when the episode's going to come

out and we kind of like talked about

this in previous times. Developers held

a lot of power uh in terms of just like

getting started. Like it was like not

knowing how to code was a massive

stumbling block. Like it's a massive

blocker to really building anything that

is useful. But these days like it is not

the case anymore. I think the role of

developers is so important still because

when you're using vibe coding tools, you

get you can get to like a certain point,

but if you really want to start scaling

it um and making it into something that

people can use, adding on custom

features, you still do need developers.

I'm not saying that you can just

completely not have developers. I think

you definitely still do need developers,

but to get to like the 0 to 0.5, 0 to

0.75, you can do now just like even if

you don't know how to code, it's pretty

crazy. Um, I'm gonna see if there's any

questions

about this.

Let's see. Prompt engineering sounds

like a great idea for browser extension

though. Yeah, it's the thing is like I

think there's so many different prompt

like prompt engineering tools that

people use and stuff like that. Why is

it that most people, you know, you can

use them? But I think in the end the

reason why prompt engineering is so

important is that you can't really

outsource that fundamental part because

it's about communication. It's like

having a human say you have like a human

relationship. Can you outsource

communication? No. You can outsource a

lot of things. You can outsource

implementation. Maybe you're like I want

to buy I don't know something like you

can't outsource the part in which you're

communicating what you want to buy. Like

that's what prompting is in the end.

You're communicating with the AI. So you

can make the communication better. you

can amplify all these things, but

without that base ability to

communicate, um, you're going to have

problems. So, like prompt engineering

isn't something that you can skip out

on, and I'm very adamant about that.

Like trying to get AI to prompt for you

without actually knowing how prompting

works itself, like you're not going to

be able to get the best results,

especially when you start building

products. Um, and then, you know, start

building products and having to test

things and evaluate things and then

you're like iteratively changing things

and tweaking things like with vibe

coding. It's really really hard for you

to do that without a good basis of

prompting.

Hi Tina, I just joined. Welcome. Is

there a website you're showing the main

resource we'll be using in the live

stream? Yeah. So if you look at the pin

comment, you can click over there. I'll

be sending you to slides afterwards.

That's what I'm going to be using. So

limits is equal to scalability and

safety. Uh what are the limits of VI

coding? Great question. So when it comes

to a non-developer

right I think you can get to like as I

was saying earlier 0 to 0.5 0.75 even

the limitations is when you want to

start scaling it is is where I would put

it like um let's say like scaling it to

more people like in the few thousands of

people you can get away with but like

perhaps like more than that um yeah I

would say like when you're getting to

actual having like users and trying to

scale it that's a limitation and another

limitation is going to be if you're

trying to add features So vibe coding um

you can get to a point in which you know

you have your core features and it you

know assuming they're not like very

obscure features you can probably get it

to work but if you're trying to add like

additional features to it that is more

advanced that's going to be more niche

that's going to be more custom then

you're going to start running into more

difficulty if you're vibe coding with um

with a non-developer like completely

with no code right I I think I think

that's kind of where the limitation is

going to lie and also most of these vibe

coding tools um these non-developer

centric vibe coding tools are going to

be web- based. So if you're going to I

think there's you know they're ruling on

the mobile side of things as well. But

if you're trying to um do like mobile

apps, if you're trying to vibe code

things that are not web- based, for

example, that's going to be very hard uh

to do with just completely no code

tools.

Uh I halfway through the presentation.

Can you share a copy? Yep. Check out the

link. Repl is amazing. I agree as well.

There's a lot of different vibe coding

tools and I think a lot of them there's

like different differences between them.

Um so that's why like I said I tend to

not focus so much on the tools

themselves because I do think like there

are certain tools like for example like

right now my go-to tool has been using

Bolt. I think it's like one of the best

ones for just like getting everything

started. Um but you know you can make an

argument for a lot of the different

other vibe coding tools as well. I think

it's more like a skill thing like choose

a tool that you think vibes vibes with

you and makes sense for you. Um, and

it's more it's also specific to the

product you're trying to build, right?

Like certain tools have certain stacks

that is more useful for um that specific

thing that you're building versus

another. But really in the end, I think

it comes down to skill and what is like

with vibe coding, you're you're using

prompting, right? That's literally what

it is that you're doing using prompting.

You develop things like a PRP um which

is a prompt prompt requirements. Oh my

god. Product requirements prompt. So

it's like a uh like defining what the

product is. All of that is prompting in

the end.

Okay. So then for on the developer side,

so once you get to like 0.5, 0.75, it's

still really important for you to then

pass it on to a developer. And for

developers, you know, if you are a

developer already, if you're not using

AI assisted coding, um I

I'm not really sure what to say at this

point because like I do think there's so

much like coding tools like coding

agents that developers can use. I do

think this is one of the areas that has

seen the most amount of success when it

comes to using AI AI tools is is for

coding. So to developers, AI assistant.

So for developers, AI isn't replacing

coding, it's augmenting the workflow.

So, think of it as like an intelligent

pair programmer that understands your

codebase and guides you to solutions

faster. Like right now, like whenever I

want to go code something and I'm able

to do it so much faster, like crazy, so

much faster than when I was trying to

just code myself before. I I like I

don't know if I could even go back to

that at that point. It's just having

unlocked the ability to just code so

much faster now. Um, so yeah, like you

don't even do things like checking Stack

Overflow anymore. or like you literally

have an AI assistant that can help you

with coding. Previously, you got to do

things like searching Stack Overflow uh

for solutions, copy pasting code

snippets, adapting to your specific use

case, reading a bunch of documentations

um and then not knowing how to do that,

debugging, trial and error, repeating

for each problem. So, this is like super

timeconuming context switching heavy

situation. The old way like most of the

time when you're actually developing

like when you're like coding something,

you're not actually like writing code,

you're searching stuff or trying to fix

stuff. But now um using AI you can you

can describe what you need in cont in

context generating tailored code review

and guiding refinement AI running test

automatically and iterating until

perfect. Um it's really fast it's

contextual it's conversational and and

on top of that it's just like you still

have like a lot of control like you

still have a lot of control compared to

Oops. Uh I don't know what happened

there. Okay. you still have a lot of

control compared to just a pure vibe

coding tool where like you're going from

a from a no code perspective, right? But

with code with AI agents that help you

with code um as a developer AI assisted

coding, you have that control, but

you're just able to do so much more and

so much faster. So yeah, you can do a

lot of these things. Context aware code

generation, intelligent debugging,

automating tests, refactoring and

optimization, document generation,

multifile editing. Uh there's just so

much. And here's like a few examples

again like different ones. Yeah, not

going to go into too much detail about

the tools themselves. Um, but just

choosing one of these AI assisted coding

tools, they really would help so much

when you're coding. So the reality is

that there's a 300% productivity boost.

Developers using AI assistance report

completing task three times faster.

You're not learning to vibe, you're

learning to guide, refine, and validate

AI generated code through an iterative

conversational workflow. So bottom line

is that if you are a developer, AI

assisted coding doesn't actually replace

your skills. It really amplifies them.

You will still need to understand uh the

code architecture and debugging, but AIS

can handle the repetitive work for you

while you focus on solving the hard

problems. So if you are a developer,

this is absolutely a skill that you need

to know uh in my opinion, in my humble

opinion in 2026.

Any questions?

Can you put into the chat um you guys?

Are you a developer or are you a

builder? Let's say like builder is like

no code people. Um and developer is for

people who are technical developer or

builder.

Put it into chat. I'm curious.

So basically we got to invest in

self-articulating

thought through thought into

propositions. Yes. Reading skills right

now is low. Yes, I agree. um got to

invest in self-articulating thought into

propositions. Yes. Is like that's what

prompting is. You're like articulating

yourself. That is true.

What's your favorite one uh in terms of

AI assisted coding tool? So for again

not so much focus on the tools

themselves, but the one I'm currently

using most is is Warp. Um yeah, I am

using Warp.

The other there's like a lot of

different popular ones as well. It's

again like choosing one that you find

that works well for you and works well

for your codebase as well. So if you're

using something like Python, you would

be fine with most of these, right? But

if you're using something that's a

little bit more niche, um I don't know

like for example like goodo for example,

right? If you're doing like game

development, then you might want to

choose like a more specific one specific

for that. And then a lot of like people

also you can like switch out different

models as well um while you're using AI

assisted coding. So you can find one

that works well for your workflow.

Just got here. Hello, welcome. Welcome.

Okay, so vibe coding is cool. However,

you still need to understand the code

because sometimes AI does not get it

right. Exactly. So you get to that 0

75.5 to 0 75 of your product. Then you

do need to switch to coding. Like you

still need that that final step.

Developer builder I do both. Nice. Dev

dev builder dev builder builder dev. Oh,

okay. Good mixture that we have here.

Very nice.

Developer.

What is best for absolute beginners with

no code knowledge? So on the builder

side like previously over here uh these

are some of the ones that are the best

to begin with in my opinion. Um if you

want to start off yeah [snorts] there's

also like replet uh if you want to try

that. Firebase studio is the free

version. So for those of you who do want

to try something without paying right

now there's fire studio and then there's

also um like open source tools as well

like metagp for example. So I don't have

that here. I'll talk about open source a

little bit later but there is like

increase in open source tools and five

coding tools as well. So those are all

ones that you can try out a lot of

different ones that you can try out.

All right now let's talk about AI

agents. This is I would say okay like

any modern normal like person like needs

to know prompting and needs to know like

different tools and stuff and if you're

a builder um if you want to build stuff

where you're a developer I do think you

need to know like vibe coding for

builders AI assisted coding for

developers. Now for people who I think

agents is also really important for

people who do want to take that step um

and actually start building products

right like in the end like with large

language model like products the kind of

golden standard the thing that you're

aiming for is building something that's

a that's something that's an agent

something that's autonomous

um and this is also where there is a lot

of opportunity for impact in the

business world so what are agents let's

first define that so agents are software

systems that use AI to autonomously

pursue goals and complete multicept task

on behalf of users with minimal human

insight oversight. Sorry, not insight.

You didn't need human insight without

with minimal human oversight. The key

difference is that traditional AI, if

you're just talking to like a chatbot,

for example, is able to respond to the

prompts um and will like tell you stuff,

but agents can take action. It's able to

do stuff like make decisions, use tools,

and work independently to achieve

objectives. So, um we're moving from

chat bots to building autonomous agents.

So traditional AI like chatbt, you're

just kind of talking to it. You're

waiting for a prompt. You're responding

to a question. There's no like actual

follow-up action per se, and you're not

using tools independent. You you can't

use the tools independently. Um, it also

requires a human for each step. For

example, if you're talking to chat to

BT, you would just do something like,

oh, write me an email, generate this

text, and then you like copy paste it

into your into your Gmail, which is

great. You know, that's that's very

useful. But when it comes to agents,

agents are able to initiate tasks

autonomously. You can complete

multi-step workflows, use tools and

APIs, make decisions independently, and

it will report back to you when it's

complete. An example of an agentic

workflow would be you can ask a question

like send a follow-up to John. Um your

agent will figure out like goes through

your emails, figure out what it is that

you're talking about, who is John,

right? Which followup. It will find the

email, draft the response, schedule the

thing, and then also confirms it after

being sent. So I hope you can see the

difference between the difference

between just using a chatbot um versus

building an auton autonomous agent to do

it. So here are some examples of real

world agent applications. Customer ser

customer support agents. So these are

agents that can handle tickets and to

end um reads inquiry searches knowledge

bases draft response and escalates if

needed and follows up. The result is

that you're going to get 65% reduction

in support tickets is based on chatbase.

Um there's also like sales research

agents. These are agents that are able

to research prospects, find contact

info, analyze company data, draft

personalized outreach, and schedule

follow-up. So, you can see like the

results are are here, right? You can

literally see the results in the

business world now. And it's impactful.

For example, we can see from Lumen that

there's four hours saved per seller

weekly, which is huge. 4 hours per

seller. Um,

data analysis agents are able to pull

data from multiple sources, clean it,

run analysis, generate visualizations,

create reports, and distribute to

stakeholders. So result is that your

weekly reports are going to be

previously which was taking hours, it's

only going to be taking minutes. Now

another example is H is HR onboarding

agent. So it's able to do stuff like

create an account, send a welcome email,

schedule meetings, assign training,

track completion and follow up of

missing items. So this will result in

much more much more consistent

onboarding with zero manual work. Um so

these are all agents that people have

built and they're seeing real results in

the in the workplace already. So what I

think you need to know about agents if

you're interested in building them is

first of all understanding how they work

uh when to use them including workflow

design like how it is that you can break

them into different tasks tool

integration integrating and testing and

validation this is really important and

then also monitoring so I'm going over

this really really fast right now I also

know that but I just kind of want to

give you guys kind of the basics of

here's the things that you do need to

know when we're building agents like we

have an entire 28day boot camp where we

go into this in a lot more detail but so

I'd definitely don't have time to uh

talk through all of it right now, but I

do want to tell you like these are the

things that you need to learn if you are

if you do want to go explore this

yourself. So with agents still it is

still important to understand that it's

not a magic solution. Agents still need

clear instructions and guardrails. So

back to the prompting situation, right?

And guardrails uh knowing what it is

that it should or shouldn't do, error

handling, so planning for failure and

edge cases. It still needs human

oversight. Um, and the way that you want

to do that, you want to start simple and

you still need iteration over time. So,

I really think that agents represent

such a massive opportunity. Um, because

there's actually a lot of like AI

solutions that people would build right

now, but like where like for example, we

do B2B consulting and we build B2B like

AI solutions for companies, right? I

actually think where a lot of value lies

going into 2026 is not necessarily like

building like an entire AI agent as a

full solution these days. It's more

about helping companies integrate

agentic workflows like more custom

agentic workflows into their current

workflows. Like for example, if you're

building if you want to like get an

agent to help you, I don't know, uh,

report generation, right? Let's just say

like data analysis agent, like you want

to do that. You want to automate some of

that very manual process. You can't

really just tell the company, hey, like

change your entire workflow. That's

probably not going to happen. Um, that's

why like out of the box solutions also

aren't the best. So generally what they

need to do and this is what we get hired

to do is that we would help them figure

out what like building a custom agentic

solution that's able to fit into their

current data pipelines their their

current like data analysis pipelines to

be able to come up with that final

report um and distribute that. So

there's a lot of value there's a lot of

impact that lies in this area. It's like

almost like that glue like you're

connecting together agentic solutions

and building custom solutions for

existing companies. Um so this is where

like I also really recommend people who

are interested in building agents to

start like thinking about freelancing. U

if it's for your own company like think

about how it is that you can integrate

agents into existing workflows as well.

Anyways, I'm going to stop rambling. So

I do think there's a lot of potential

here like massive massive potential

here. Um potentially for people who want

to be like freelancers, start agencies

and stuff. Uh a lot of companies are

looking for this kind of work and I know

this because that's what we do. Um okay

so getting started well so you don't

actually need to build agents from

scratch which is wonderful. You can

start by understanding agent workflows

and you can explore no code platforms

like NA10 make zap year to create simple

agents. Isn't that pretty cool? Like you

can literally create simple agents now

using no code tools as well. Um yeah we

have a boot camp that covers agent

development in depth from planning to

production to deployment. So even like

within our boot camp like we offer two

tracks, right? One of them is a no code

and a code track. Um and the reason why

we do this is again like I don't like to

actually focus so much on the tools

themselves because they keep changing

and getting better over time. But if you

understand how agents work, what what's

the structure of agent, what's the

infrastructure, what are the actual

basics required, you can actually use

different tools to build the agent. The

tool itself is is simply just a tool.

Okay, enough rambling about agents

before I go into open source. Any

questions about agents?

I hope you guys have questions about

agents. Let me drink some water first.

Do you think Gemini has better data

analysis capabilities in chatbt?

Um, in some ways, yes, in some things

that you do. Me personally, I actually

still much prefer Claude for like data

specifically. Anything that's more

technical that cries analysis, I tend to

go with Claude, but Gemini is is really

good. So is Chachet as well. I think

Gemini might be a bit better than

Chacht. Don't quote me on that one cuz I

think it, you know, they're similar, but

I personally think Claude is the best

right now in my opinion. Um, let's see.

Ra like rather than the sort of

step-to-step execution like any intent

sequences or even something hyper

sophisticated like do a thing and it

does all the things to do the thing.

Okay, I think that was a response to

somebody else. Cool. Um

what's the cost of your boot camp

please? Our boot camp is 997.

Yes, 997. If you go to attend some of

our if you have gone to some workshops

um then you do get a discount as well

but the base price is 997.

We do not currently have a boot camp

available. Um we're I think we're

launching again early next year. So you

can also sign up for the weight list if

you want if you want uh to get more

information. So all of our like

workshops and things like that because I

don't like spamming YouTube um about

this. So that's why we have like a

mailing list where if you're interested,

we send out information about um when

that we open enrollment and things like

that. Are you drinking directly out of a

vase? Excellent question. Yeah, I am.

Not a vase. Is it is it is a water jug?

Is that better? [laughter] I know,

right?

Um

talk about AI job marketing for people

who get certifications. I'm not sure if

I'll earn a living or if it's just for

fun. You know, hold that question. I do

want to talk about that. I do think

careers uh in the job market is very

interesting to me. If you ask me that,

can you ask me that question a little

bit later? I want to get through the

slides first. Um

I'm actually Yeah, I think that one's

interesting.

Boot camp with them worth it? I hope. Is

that a statement? If it is, thank you

very much. Oh, hey. Yes, it is a

statement. You went through a boot camp.

Hello. Um thank you. I'm very happy to

hear that.

Okay, I agree. I love cloud

data analysis agents are sus. Bad data

is worse than no data. And most

companies set up semantic models and

write reports and dashboards unless

something like analyzing unstructured

data. I think that's the thing. There's

actually a lot of um a lot of it is

unstructured data. And by unstructured

data, what we're referring to here is

stuff that's like, you know, people

maybe going on interviews and saying

things, uh sending emails like kind of

like written. Oh my god. natural

language like unstructured data as

opposed to numerical data. I think

agents are amazing for that and a lot of

companies can get so much insight by

incorporating unstructured data into

their complete analysis. Even with

structured data, uh I think that

combination you can get really really

good results from it. Even like with our

own company like when we do our analysis

with with you know our internal products

and like whatever things like that um

yeah like when we're we you know

whenever we do like an audit um and look

at the data and then also analyze the

data yeah we get a lot of really amazing

insights um from that.

All right talk about open source now.

[sighs and gasps]

Okay, this one is kind of like kind of

it kind of came out of left field for

me. Maybe some of you guys in the chat

like if you can if you saw this coming

like please do put in the chat because

I'm first to admit here that I kind of

saw like oh open source is really cool

and stuff but I did not expect how

quickly open source is rising. I'm

actually kind of shocked that nobody's

really talking about it yet on the

internet. So I don't know maybe we're

going to get like more internet people

talking about it um soon. But yeah, this

one kind of like came out of left field

for me. I just didn't see it being so

fast. We can see that with open-source

AI. Um the performance gap is very much

closing now. So gap is narrow from 15 to

20 points to 7 to n points parody

expected in Q2 2026. This is from a

benchmark analysis. Oh um yeah. So I'm

before like I agree with that. So with

open source like what I mean by that um

is is models that you're able to use and

tweak and download and do stuff with

without having to go through like a

company. So whenever we think about very

popular ones like uh let's see like Chad

GPT cloud Gemini like these are all

closed source AI. So in order to access

these models, you have to go through

their platform. You have to use their

APIs and like build on their

infrastructure. And it's also more of a

black box because we don't really know

what exactly it is that they're doing to

their models, right? In in in in the

back end. Um but with open- source, this

is in contrast to that. These are models

that you can actually use. They're

available. Often times they're really

cheap or free completely and you can use

them build on top of them. There's a

community on top of them. You can

develop them, put them into different

products, fine-tune them as you like. Um

but these are called like open source

models and the most popular open source

model these days is deepseek for

example. So that's what I mean by open

source. Yeah. And then another like

super interesting thing is is that open

the open source AI movement is being led

by China. Like I I think that's super

interesting because I think western AI

has been very much like concentrated um

has al has like very much been closed

source AI, right? well from China is

like just kind of just like rising up

and I didn't really expect this at all

but Chinese open source like starting

with Deep Seek like Gwen like a lot of

these um Chinese models are open source

they're really really cheap were free um

and that's why like tools that are built

on top of them is also free were a lot

cheaper and uh I read the stat recently

I think it was A16Z that said that um

now 80% of the companies that are

pitching to them are building products

using open-source. Like they're building

products on top of open source AI.

That's super interesting. This is like a

really big difference than just like a

few months ago when the majority of

people were still using closed source

AI. And that's because open source AI

has gotten so much better and you still

get the benefits of open source of it

being cheap, you know, being able to

tweak it, do a lot more things to it. So

people have just really like switched

over to using open source um when

they're building stuff in in particular.

Yeah. Yeah. So you get like 3.5 times

cost savings compared to proprietary

models. Um 11 times year-over-year

growth in tech industry AI adoption. So

industry adoption exposure there's a lot

of it in technology, healthcare,

manufacturing. It's also really useful

for things like healthcare for example

because um you generally like you I

don't know like most cases you don't

want to be uploading patient data into

like a closed source AI like Chacht for

example directly because you don't know

what's going to happen to it, right? But

with open source like you have that

model so you know where that data is

going build an infrastructure and

privacy around it. So this will allow

you to still abide by healthcare

regulations and building healthcare

products. So this is like a huge

innovation that is opening up a lot of

these um industries that previously were

constrained because of privacy and

regulation concerns. Yeah. Uh and

manufacturing as well. So I'm going to

stop here for now. So yeah, like I think

there's a lot more about open source I'm

really interested in covering. And also

starting in in 2026, our AI agent boot

camp, we're going to be covering open

source AI deployment, finetuning, and

production as well because I think this

is absolutely crucial. It's like a huge

thing. It's just going to start

exploding more in in um 2026. So

definitely check that out if you want.

Let me see if anybody has any questions

about open source.

Whoopsies.

Oh no. How do I get myself?

Whoops.

Why can I not see comments?

Too many tabs open. Okay, cool. Let's

see. Comments. Should one be concerned

with security issues using China origin

AI etc or open source? So that's that's

why it's open source, right? Because um

I think China knows that a lot of people

are not going to be cool with using

their Chinese servers. That's why

they're developing on the open source

side. You're not using Chinese servers,

right? You're you're running these

Chinese models on your own servers. You

can do it locally. You can do it on the

web um on the cloud or whatever. So

you're you're just simply using the

models that are being developed by

China. So that's why like no, you should

not you don't need to be concerned about

security issues because you actually

have way more control of the security

with open source than you do with closed

source because you don't really know

what open AI is doing, right? Um

what does China do with your prompts

though? Security risk enterprises.

They're not doing anything with your

prompts because again you're taking the

models and doing the stuff that you want

to do yourself. So you're not like

sending information back to the

companies themselves. You're just using

the models and tweaking it and

developing it.

Um,

oh, thank you so much, Rex. Yes, if you

are interested in the boot camp, um, if

you want, Rex has put the link on there.

We do have a boot camp that should be

coming up early in 2026. So, last time,

and thank you guys so much for this,

like we sold out under like a couple

hours, and prior to that, we sold out

under 1 hour, I think. Um, so we only we

do limit it to 100 spots because we want

to make sure everybody gets the best

experience possible and we do sell to

the weight list first and we've

consistently like sold out through the

weight list. So if you are interested

sign up, please sign up for the weight

list so you can get more information

about that and you that's also how we're

going to be sending you out information

about uh shorter workshops that we do as

well. Like the app sprint workshop is

when we built applications in an hour

and a half. We had an agent

breakthrough. like starting it's like a

mini agent workshop that we did a couple

weeks ago. We had a freelancing workshop

as well. Probably going to have an open

source workshop um for next year too. So

that's where we're going to communicate

these these workshop things.

Um okay let's see open source sounds

like how Python was able to become so

widespread. People can build commercial

products on top of open source. Exactly.

This is exactly what it is. There's like

the parallel is very very clear here. Um

it's like there was a time in which you

know people were using proprietary

coding languages right like what was

popular was not like JavaScript or like

or like Python it was like propriety

languages that are being developed

within companies um but then the open

source movement came and then there was

a lot more um interest in stuff like

Python because you can it's a community-

based thing and everybody had access to

it and then it was it was free to use

and then um yeah over time now the

common thing that people do is using

open source developing like coding

coding languages, not like proprietary

ones anymore. Like when you're going to

learn coding, you're learning like

Python or JavaScript or something like

that. You're not learning like I don't

know I don't even know what the names

are. I can't even think of the names

because they're like proprietary. It

completely just like died off. Um and I

feel like I feel like this is seems to

be a trend that may be repeating on the

AI side of things as well. Just the open

source side. um building things out in

the open, having a much cheaper, having

a lot more control, community

involvement, fine-tuning. That would be

really cool. That seems to be the trend

um that we're moving towards now.

Um essential AI skills as a servant.

Well, essential AI skills in my opinion,

like let me put that as a qualifier. In

my opinion, these are the things that

you should be learning. Yes. Um

yes. Yes, agent breakthrough was great.

Thank you. Thank you very much. I'm

really really really glad that that you

thought it was great. It's helpful.

Um

thank you guys for our past students.

Thank you so much for sharing your

experiences. Um we're really happy about

that. We spent a lot of effort like

genuinely we spent a lot of effort in

making the workshops and boot camps and

things like that. So that it is making

me really happy right now. Anyways,

okay. So um does it matter which model

is better? Does the model

[clears throat] you use do what you want

matters? That is true.

So questions, [gasps] do we know the

start date for the early 2026 boot camp?

We do not currently know. We know it's

going to be in Q1, but we don't exactly

know the the date yet. So I don't want

to like say something incorrect.

[laughter]

Yes, but we will announce it as soon as

we do know the date. Um

what open source was open source boosted

by deepseek or were there already talks

in industry about building things out in

the open? Yeah. So there was definitely

talks and I think again like I don't

know what was going on the mind of these

model developers. Um I think probably

perhaps what happened was uh there was

like open source that was already

brewing but it was just sort of you know

it it was just like nowhere near as good

as a closed source model. So people

weren't really paying attention to it as

much. But when DeepC came out, that was

the first time when people saw like, oh

my god, like a open- source model that

is way cheaper, has all this control, no

like none of those privacy concerns and

whatnot, is able to perform on par close

source AI. I think that was like the

breaking point. And then after that, it

was a flood of open source that started

coming out um primarily led by Chinese

AI um Chinese AI companies for these

open source models. And then people

started developing tools on top of the

open source models. So they were able to

make it so much cheaper than the closed

source equivalent, right? Because closed

source if you're if you're developing on

closed source models then you have to

pay the premium for using those closed

source prices. So then you know when

you're selling these products you have

to also mark it up higher. But for

people who are using open source models

on top of the other benefits that we

talked about um you're you're also just

able to make it a lot cheaper for people

um because you're paying less for the

models themselves. So the cost overall

is also a lot cheaper. So yeah, that

kind of opened up the floodgates uh in

terms of the open source movement and it

happened really really quickly just like

within the past few months.

Yep.

Have how many of you guys have tried out

open source? Can you put into the

comments if you have tried out open

source models

before

or open source products?

Put in the chat. Okay. Oh no. Okay, let

me go to through the size a little bit

faster. I got excited. Okay, so let's

talk about critical workplace AI skills.

So again, in my opinion, these are the

things that you do need to know outside

of just the technical side of things,

right? Data literacy, I think, is super

important. Still understanding data

itself is non-negotiable. Being able to

read and interpret data insights,

identify trends and anomalies, check for

errors and biases, making data informed

decisions because with good data is

really the basics, the basis of how

you're developing products and on AI as

well. AI is built on top of data. So

that's why it's important to have data

literacy, critical thinking, don't have

AI outputs blindly. I'm just going to

put that here. Uh because so many people

are using AI these days and there's

always like that other side of the other

side of the equation where people might

be using AI in ways that maybe are not

the best ways. And it's important for

you to be able to distinguish between

what is a good way of using AI, what is

not and looking at the outputs like is

this AI, is it not AI generated, is this

actually real? You know, all of these

questions. Um AI is powerful but not

infalluable. So is quality control,

continuous learning. So if there's like

a singular skill like meta skill that I

think has been the most beneficial to my

life by far is just the ability to

learn. Especially now with the AI

landscape, it's changing so quickly.

There's so much development. I was

talking about the open source thing.

This is happening like just a few months

ago. you know when it meteoric rise and

80% apparently of companies that are

like startups are building off open

source now like this literally happened

within the span of like months and the

AI landscape itself is changing so

quickly as well so being able to learn

is is like that superpower like that you

the modern superpower is is your ability

to learn and workflow integration so

knowing when and how to use AI this is

so important like when people are

building products like we've found that

people who go through our boot camps for

example, right? Like there's people who

go through our boot camps, build agents,

um, and then like sell them and things

like that. And what's interesting is

like the people who are able to build

like the most impactful products, like

the ones that they can sell, the really

great ones, they're actually ones who

really understand when and how to use

the AI. like they're actually people who

have understanding of um enterprises

like understanding of the business side

of things and what they're building and

then combining that with knowing what AI

is good at and what it's not good at to

build really good solutions. So that

combination is is where it's at. So you

need to be able to identify automation

opportunities, integrate AI into daily

tasks, balancing AI uses human expertise

and optimizing processes continuously.

So understanding that you can use AI to

eliminate grunt work, not to replace

thinking. So yeah, like this is also a

really big one. So in my opinion, these

are the critical things that are not

technical in nature, but they are really

important as we're moving in 2026. So

technical skills can get you into door.

These workplace skills will let you be

able to lead, innovate, and advance your

career faster than those who only focus

on tools themselves.

Okay, so career impact, talk a little

bit about this. So AR skills are the new

career currency. Like I don't know if

anybody wants to argue with me about

that one, but we can do that if you

want. I I think is quite clear in every

industry professionals with AI skills

are becoming indispensable. They're not

just keeping up. They're leading

transformation of how things are going

out. So if you you need AI skills in

order to stand out like while others are

waiting around, you're the one that you

can be mastering the tools that defining

the future of work. Um using AI can just

make you so much more productive. So it

will lead to faster promotions, better

opportunities, higher earning potential,

all within reach. Especially if you're

working at a company, if you can

integrate AI solutions into their

company, there's like so many hang

lowhanging fruit now. Um, building

something like that, you can like

significantly save the money or increase

output and just do things that you

weren't able to do before. Um, yeah,

like stuff like customer service stuff,

uh, analysis, automation, like report

generation. These are all like very

common like lowhanging fruit that a lot

of companies deal with. So, there's

actually like solutions that you can

build and people know about these

solutions. It's not like it's rocket

science here, but if you build it and

custom for specific types of companies,

then they would be able to reap those

benefits. Um, creating like being able

to create build ideas that are

impossible for launching projects that

you just simply could not have at all.

Like if you want to be an entrepreneur,

a solarreneur, there is no better time

to do this right now. You can literally

build products from scratch, like

whatever it is that you want to build so

quickly and at such low cost. Um there

has never been a time that you are able

to do this. The number of like

soloreneurs that I know like people who

are lifestyle um soloreneurs you know

like freelancers uh who have agencies.

Yeah. It's just massive. Like there's so

many people doing this and they're

they're like doing really well doing

this because they understand how to use

these um these skills and future

proofing as well. As AI is continuing to

reshape industries, you're going to be

ready like you're not just scrambling to

catch up. it is very much the way where

things are headed. Um, yeah, AI is is is

the future and it's already here. So, if

you don't know these skills, it's going

to be difficult for you. I would me

personally like you there's like pros

and cons to everything, right? Me

personally with the technology this

powerful that's here, I would much

rather be the one that is defining how

it's being used as opposed to scrambling

to catch up to it. Okay. So, I think the

most costly mistake is not starting now.

Um, so every day you delay the gap

widens. The professionals learning I

today will be the leaders of tomorrow.

And I do not doubt that at all. I will I

will defend this one.

All right. Open floor. Okay. Sorry, I

kind of went over a little bit. Feel

free to drop anybody that needs to. Um,

but I'll stay around for a little bit

longer. We can chat about the things I

just talked about in the slides. And

also if you have any additional

questions, we ask me anything when we

talk about like learning questions,

career questions, tool questions,

whatever you want to ask. Please feel

free to do so. Um, let me go back to

the chat.

Does this this page designed by AI? Yes.

So, all these slides are also designed

by by AI. They're designed by AI and

they're executed with AI as well. Yes.

[laughter]

Nice icon graded. Oh, thank you very

much. [gasps]

Thank you to our wonderful team. Uh,

yeah. So, Ibrahim from our team, which

he's also a um I don't know if he's here

right now, but shout out to Ibrahim. He

is our instructor. Uh he made a

he he made a internal app that's able to

produce these these um slides. Well,

previously it would have taken forever

to make these slides. [sighs]

Yep.

To use open source, do you need a lot of

personal comput? Nope, you don't need to

do that. You can also run things on the

cloud as well. So you don't need a lot

of personal compute.

Uh do we need a good understanding of

maths and Python to learn AI? Nope, you

do not.

Do you still offer the vault to the

folks in the wait list? That's what I

purchased and was just as good as live

stream. There are follow-up meetings to

ask questions worth every worth every

dollar. Do we still offer the vault? I

do not believe we've been offering a

vault for the past few cohorts.

No, I do not believe we are going to be

we had offered them. That's actually

really good feedback though. We I think

we actually stopped offering devolve

because we thought that people were not

getting as much out of it um as we

wanted to. But thank you for that. We

can maybe re-evaluate that.

Um let's see.

How often do you live stream? I just

randomly found the stream in real time

after your video is recommended by

YouTube. Oh, that's a good question. I

stream

usually like once a month, I would say.

Um, perhaps more. If there's like

something really exciting I want to talk

about, then I will stream more often.

Sometimes a little bit less, but I feel

like kind of now once a month. Yeah,

plus or minus one or two. Um,

ah, so many questions. Thank you. What's

your favorite color?

Wow, I'm stumped. I don't know.

What's my favorite color? Yellow.

Maybe it's cuz my last name is yellow,

but favorite colors. Yeah, I guess I

like yellow.

Um

would love to see it. The slide the the

slide app. Yeah, maybe I'll show we'll

we can like I will demo it at some

point. Not now, right now, but because

there's like proprietary information in

it. Yeah, maybe we'll demo. I think

you're referring to like the slide to

make how like how it is to make the

slides, right? How we make the slides.

Yes. Maybe we'll demo it sometime.

That's that's an example, right, of like

an app, an internal app that we use that

dramatically changes like our um it

increases output so much. Like our team

is actually really small. Uh the fact

that we're able to like do these live

streams, like create content, that we're

able to like make um run these boot

camps and like workshops and stuff.

Yeah, it's pretty crazy.

Purple. That's true. Well, purple is one

of my favorite colors, too. Purple is

complimentary to yellow, but I do I'm

surprised it's not purple. [laughter]

Purple is like octopus purple. Yes,

octopus purple. My favorite color is

yellow. Octopus blue. Um

okay let me see if you use open source

where's the data stored compared to

non-opensource if that makes sense. Yes.

Yes. So you can store your data in

databases that you're controlling. So

you can store it you can have it locally

but you can have it on the cloud. So you

can store it store it like you know

wherever it is that you want to store

it. You have like your own servers like

on prem. Um if you're if you're a

company you care a lot about about

having a lot of privacy control. Um so

it yeah you can store in a lot of

different places and then you can just

connect your data storage to the

application that you're building with

the open source large language model for

example with like closed source um you

can it's usually the data that's stored

there there are ways of doing it outside

of the ecosystem that you're building on

top of but generally speaking like with

for example if you're going to use like

chat GPT right like open AI stuff and

you're using like GPT models then the

data you they do have like databases

that you can connect to as So they have

like a full ecosystem of this, but you

can also have your date your data stored

in different places too. So it's not

necessarily that um you need to like

store your data in a specific place. You

can just store it where you like and

then have access to that with your

model. I hope that makes sense. But with

open source, you generally have more

flexibility on how it is that you're

storing it and where it is that you're

storing it as well.

Um,

do you have a guide on how to build a

platform on top of an open source model?

That is a workshop idea that I think

we're going to do beginning of next

year. Yes. Not yet, but yes, I think so.

What do you think about the field of AI

safety? I'm so glad that you said that.

I did a video about AI safety and it was

like the worst performing video I had

done in like 3 years. I was kind of sad

about that cuz I thought it was a really

important video. So, I was like, darn

it. Um, I'm really glad that you asked

that. I think AI safety is one of those

things that's like super slept on and

then something's gonna happen big at

some point. Everybody's gonna go, "Oh my

god, AI safety." And then we're going to

have like a huge influx of um, attention

on AI safety. But it's a field that's

like inevitably going to grow even more.

Um, knock on nothing like terrible

happens, but it's just like we're like

kind of just like waiting for a

disaster. Knock on wood, but like I were

kind of like waiting for a disaster to

happen right now for people to really

pay attention to it. While really like

it's something that you should really be

paying attention to. Uh for people who

are interested in AI safety, I think

that's a field that is very ripe for a

lot of opportunities.

Uh [sighs and gasps]

let's see. Let's see.

Have you ever had a shirt moment during

a live what? I don't understand the

question. I'm learning AI by myself, but

I'm not sure in which situation to use

it because I'm not sure whether

potential customers even know what they

that they need it. Take into account

take that into uh take into account that

I'll be freelancing. So, okay, that's a

great question. You should not be

thinking about whether your potential

customers think that they should or

should not know whether you want to use

AI in a situation. That's not their job.

That's your job to figure out where it

is that you should be using AI. What it

is that you're freelancing for is

solving a customer's problem. Like

you're not being like, hey, you should

use you should incorporate AI into

company. What you're selling, you know,

selling if you're freelancing would be

like, I can solve your problem. Let me

solve your problem. And then the way

that you solve the problem is going to

be through AI. Does that make sense?

You're not a like you're not advocating

for the use of AI purely because you

think a company should use AI. you're

just offering a solution in general, but

your solution happens to be AI.

Um, how do you sell an AI workflow? I

guess that's kind of related question,

kind of similar to what I talked about.

There's different ways of doing it. You

can build like a product like a AI

solution and try to sell that. Um, like

you can build like a vibe coding tool

that's like a AI solution for example,

right? Like a a full AI product and sell

it. You can also do freelancing. Um, and

you can also where you're like working

with a company and building up custom

workflows for them. So that's what we do

as a company. We can also do consulting

as well, which is something that we also

do. We don't build the products

themselves, but we kind of like help

companies incorporate AI into their

existing workflows.

Yes. Makes sense. Wonderful. Great. All

right. I'm going to leave it as that

because we are a little bit over

already. Thank you so much for joining

this this live stream. I really hope

this was helpful for you. Um yeah, I

really really hope this was helpful for

you moving into 2026. So you guys got a

few days to plan out what it is that

you're going to be learning for the um

for 2026. And I hope this is these are

the skills that you're going to be

learning then. And again, if you want to

get the resources like the the slides

and stuff, just you can sign up on the

pinned comment. Uh it's free. Like I

we'll send you an email with all the

resources and slides and stuff like

that. All right. Thank you all so much

for joining and have a wonderful rest of

your day or evening.