part 3

May 12, 2026 14:09 · 19:31 · English · Whisper Turbo · 3 කථානායක
මේ පිටපත අද කල් ඉකුත් වෙනවා. ස්ථිර ගබඩාව සඳහා උසස් →
දේශන සූචිය
පෙන්වමින් පමණක්
0:10
S… Speaker 1 (part 3)
That is freemium.
0:15
S… Speaker 1 (part 3)
And that is what happens in all these apps,
0:18
S… Speaker 1 (part 3)
ChatGPT,
0:19
S… Speaker 1 (part 3)
Gemini, Claude,
0:20
S… Speaker 1 (part 3)
and all of that,
0:20
S… Speaker 3 (part 3)
where...
0:22
S… Speaker 1 (part 3)
95, 96,
0:23
S… Speaker 1 (part 3)
97 % of the people are free users.
0:24
S… Speaker 1 (part 3)
They use it,
0:26
S… Speaker 1 (part 3)
they use it, they use it.
0:27
S… Speaker 1 (part 3)
There will be a point of time where they will upgrade.
0:31
S… Speaker 1 (part 3)
So those three people,
0:33
S… Speaker 1 (part 3)
four people, five people who are paying for paid chat GPT account.
0:38
S… Speaker 1 (part 3)
are the ones who are paying for everyone else's free consumption.
0:42
S… Speaker 1 (part 3)
That's the freemium model.
0:43
S… Speaker 2 (part 3)
You'll be
0:48
S… Speaker 2 (part 3)
charged like consumption of your...
0:50
S… Speaker 2 (part 3)
I don't know how far is it...
0:53
S… Speaker 1 (part 3)
No, so it's blown out of context.
0:55
S… Speaker 1 (part 3)
So basically,
0:56
S… Speaker 1 (part 3)
when you think about costs,
0:58
S… Speaker 1 (part 3)
what are costs?
0:59
S… Speaker 1 (part 3)
What is costing you more?
1:01
S… Speaker 1 (part 3)
What is that?
1:03
S… Speaker 1 (part 3)
So, at economics of scale,
1:06
S… Speaker 1 (part 3)
everything gets cheaper.
1:07
S… Speaker 1 (part 3)
The same
1:17
S… Speaker 1 (part 3)
happens,
1:20
S… Speaker 1 (part 3)
this is scaling law.
1:21
S… Speaker 1 (part 3)
Simple scaling law.
1:22
S… Speaker 1 (part 3)
The same happens across the board.
1:24
S… Speaker 1 (part 3)
As long as you have a factory,
1:28
S… Speaker 1 (part 3)
which can manufacture as many mugs as you want.
1:30
S… Speaker 1 (part 3)
But let's say,
1:31
S… Speaker 1 (part 3)
after one lakh mugs,
1:33
S… Speaker 1 (part 3)
the factory capacity will be.
1:35
S… Speaker 1 (part 3)
If you have a factory, it will be a little bit.
1:38
S… Speaker 1 (part 3)
But over the course of time,
1:40
S… Speaker 1 (part 3)
it will be a little bit.
1:40
S… Speaker 1 (part 3)
Now,
1:41
S… Speaker 1 (part 3)
imagine a world where you want to make 100 billion mugs.
1:45
S… Speaker 3 (part 3)
Okay?
1:46
S… Speaker 1 (part 3)
You have a factory.
1:48
S… Speaker 1 (part 3)
The factory is not the hardware.
1:50
S… Speaker 1 (part 3)
Let's say you are able to build the factories also.
1:52
S… Speaker 1 (part 3)
Because if you have a factory,
1:54
S… Speaker 1 (part 3)
it will be energy.
1:55
S… Speaker 1 (part 3)
It will be electricity.
1:59
S… Speaker 1 (part 3)
If you have 100 billion mugs,
2:00
S… Speaker 1 (part 3)
you don't have the energy from the city.
2:03
S… Speaker 1 (part 3)
Right.
2:04
S… Speaker 1 (part 3)
So for the amount of compute or
2:09
S… Speaker 1 (part 3)
for the amount of usage the world potentially needs
2:13
S… Speaker 1 (part 3)
or the weights going up,
2:15
S… Speaker 1 (part 3)
the biggest bottleneck,
2:17
S… Speaker 1 (part 3)
the biggest reason why they're not able to scale as fast as possible is
2:22
S… Speaker 1 (part 3)
purely because there is an energy deficiency.
2:25
S… Speaker 1 (part 3)
Like electricity.
2:26
S… Speaker 1 (part 3)
Electricity.
2:29
S… Speaker 1 (part 3)
Because they are power hungry data centers.
2:33
S… Speaker 1 (part 3)
Right. Energy deficiency only.
2:36
S… Speaker 1 (part 3)
So what he's saying is the ultimate thing that you have to solve for.
2:39
S… Speaker 1 (part 3)
The ultimate thing that everybody is going to pay for.
2:42
S… Speaker 2 (part 3)
Is power.
2:43
S… Speaker 1 (part 3)
Is energy.
2:44
S… Speaker 1 (part 3)
Right.
2:45
S… Speaker 1 (part 3)
That is what eventually when all this will become cheap,
2:49
S… Speaker 1 (part 3)
all this robots are running,
2:51
S… Speaker 1 (part 3)
everything is happening.
2:52
S… Speaker 1 (part 3)
Then eventually you will basically be paying a electricity bill and a generalized
2:56
S… Speaker 1 (part 3)
statement.
2:56
S… Speaker 1 (part 3)
Got it, got it.
2:57
S… Speaker 2 (part 3)
On the last podcast,
2:59
S… Speaker 2 (part 3)
I was asking,
3:01
S… Speaker 2 (part 3)
how can a normal guy who never used AI become a generalist
3:05
S… Speaker 2 (part 3)
in the AI?
3:06
S… Speaker 2 (part 3)
Our topic is about it.
3:07
S… Speaker 2 (part 3)
Level 0 to 5.
3:09
S… Speaker 2 (part 3)
And people loved it.
3:13
S… Speaker 2 (part 3)
And even I liked it a lot.
3:14
S… Speaker 2 (part 3)
In case,
3:15
S… Speaker 2 (part 3)
these are the same questions.
3:16
S… Speaker 2 (part 3)
Because considering all these advancements,
3:18
S… Speaker 2 (part 3)
you put in a situation law.
3:19
S… Speaker 2 (part 3)
What do you advise?
3:20
S… Speaker 2 (part 3)
And do you still call that scope as generalist?
3:24
S… Speaker 2 (part 3)
What do you call it and what should people do right now?
3:28
S… Speaker 1 (part 3)
So here's a fun story,
3:29
S… Speaker 1 (part 3)
okay? Like you said,
3:30
S… Speaker 1 (part 3)
do you still call it AI generalist?
3:32
S… Speaker 1 (part 3)
Yes, we still call it AI generalist because we coined the word.
3:35
S… Speaker 1 (part 3)
I mean, oh yeah.
3:38
S… Speaker 1 (part 3)
So in a way,
3:39
S… Speaker 1 (part 3)
we put AI and generalist together.
3:41
S… Speaker 1 (part 3)
Generalist is a common word.
3:42
S… Speaker 1 (part 3)
I commonly started to use that in
3:46
S… Speaker 1 (part 3)
the office and then I started to speak about them in podcasts and all.
3:50
S… Speaker 1 (part 3)
At that point of time,
3:52
S… Speaker 1 (part 3)
nobody was talking in that lines.
3:53
S… Speaker 1 (part 3)
Now,
3:54
S… Speaker 1 (part 3)
there is this word that has become extremely popular and
3:59
S… Speaker 1 (part 3)
that is called as AI orchestrator.
4:02
S… Speaker 1 (part 3)
AI orchestrator is essentially AI generalist.
4:04
S… Speaker 1 (part 3)
It's the same.
4:05
S… Speaker 1 (part 3)
It's a different world.
4:06
S… Speaker 1 (part 3)
The world has,
4:07
S… Speaker 1 (part 3)
until the world has caught up to AI orchestrator,
4:10
S… Speaker 1 (part 3)
like a Sam Altman,
4:11
S… Speaker 1 (part 3)
like Elon Musk has spoken about it.
4:13
S… Speaker 1 (part 3)
Vinod Khosla has spoken about it.
4:15
S… Speaker 1 (part 3)
Everybody has spoken about AI orchestrator or AI generalist in some form
4:19
S… Speaker 1 (part 3)
or fashion.
4:20
S… Speaker 1 (part 3)
So one year ago,
4:21
S… Speaker 1 (part 3)
the acknowledgement was not loud yet.
4:26
S… Speaker 1 (part 3)
But one year or two,
4:28
S… Speaker 1 (part 3)
it's actually one and a half years old.
4:30
S… Speaker 1 (part 3)
The world is one and a half years old.
4:31
S… Speaker 1 (part 3)
When we designed it,
4:33
S… Speaker 1 (part 3)
it was designed or the way,
4:36
S… Speaker 1 (part 3)
because it was made for my team.
4:37
S… Speaker 1 (part 3)
That became the roadmap that I shared here.
4:41
S… Speaker 3 (part 3)
Right.
4:41
S… Speaker 3 (part 3)
Right.
4:42
S… Speaker 1 (part 3)
But still,
4:43
S… Speaker 1 (part 3)
yes, AI generalist is still very,
4:45
S… Speaker 1 (part 3)
very relevant.
4:46
S… Speaker 1 (part 3)
That is it.
4:46
S… Speaker 1 (part 3)
That is actually,
4:47
S… Speaker 1 (part 3)
the notion that everybody has to become a generalist.
4:51
S… Speaker 1 (part 3)
What is a generalist?
4:52
S… Speaker 1 (part 3)
You are an AI first problem solver.
4:55
S… Speaker 1 (part 3)
How do you solve problems using AI?
4:57
S… Speaker 1 (part 3)
By knowing which AI tool to use where.
5:00
S… Speaker 2 (part 3)
That is what an AI orchestrator is.
5:01
S… Speaker 2 (part 3)
So LHS is good to RHS.
5:03
S… Speaker 2 (part 3)
But I think there is a new framework
5:08
S… Speaker 2 (part 3)
that we designed called as ADAPT.
5:10
S… Speaker 1 (part 3)
What is ADAPT?
5:12
S… Speaker 2 (part 3)
A is acknowledgement.
5:14
S… Speaker 2 (part 3)
That is step number one.
5:16
S… Speaker 2 (part 3)
The biggest problem is you have to acknowledge that
5:21
S… Speaker 2 (part 3)
AI is here.
5:21
S… Speaker 2 (part 3)
Our lives
5:26
S… Speaker 2 (part 3)
will revolve around AI.
5:28
S… Speaker 2 (part 3)
So we don't have an option.
5:30
S… Speaker 2 (part 3)
Acknowledge it.
5:31
S… Speaker 1 (part 3)
Or accept it.
5:31
S… Speaker 2 (part 3)
A lot of jobs are already going because of AI.
5:35
S… Speaker 2 (part 3)
Acknowledge it.
5:37
S… Speaker 2 (part 3)
The job that you are doing,
5:39
S… Speaker 2 (part 3)
acknowledge it.
5:43
S… Speaker 2 (part 3)
But also acknowledge the fact that the job that
5:47
S… Speaker 2 (part 3)
you are doing today as is,
5:48
S… Speaker 2 (part 3)
is going away.
5:50
S… Speaker 2 (part 3)
But that job will transition into a new job.
5:53
S… Speaker 1 (part 3)
Correct.
5:53
S… Speaker 2 (part 3)
So you have to level up for this.
5:55
S… Speaker 2 (part 3)
Acknowledge that also.
5:56
S… Speaker 2 (part 3)
Acknowledge the fact that AI has the powerful thing.
6:00
S… Speaker 2 (part 3)
We all have to use it,
6:02
S… Speaker 2 (part 3)
whether we like it,
6:03
S… Speaker 2 (part 3)
don't like it,
6:04
S… Speaker 2 (part 3)
respect it,
6:06
S… Speaker 2 (part 3)
don't respect it.
6:07
S… Speaker 2 (part 3)
I say it,
6:07
S… Speaker 2 (part 3)
you say it, you have to use it.
6:09
S… Speaker 2 (part 3)
That is the first layer that you have to do.
6:11
S… Speaker 1 (part 3)
Agreed.
6:11
S… Speaker 2 (part 3)
Don't fight with it.
6:12
S… Speaker 2 (part 3)
There is no way you can win.
6:14
S… Speaker 1 (part 3)
Yeah.
6:15
S… Speaker 2 (part 3)
Once you acknowledge,
6:17
S… Speaker 1 (part 3)
right,
6:18
S… Speaker 2 (part 3)
you come to D.
6:20
S… Speaker 2 (part 3)
which is dabble.
6:21
S… Speaker 1 (part 3)
Okay.
6:22
S… Speaker 2 (part 3)
What does dabbling mean?
6:23
S… Speaker 2 (part 3)
You are dabbling with multiple
6:27
S… Speaker 1 (part 3)
things.
6:28
S… Speaker 1 (part 3)
Right?
6:28
S… Speaker 2 (part 3)
Dabbling means you can use your content.
6:31
S… Speaker 2 (part 3)
So you can use your content.
6:33
S… Speaker 1 (part 3)
You can use a small tool.
6:59
S… Speaker 2 (part 3)
You will never realize the true potential of AI.
7:03
S… Speaker 2 (part 3)
But dabble chele dan kondi,
7:05
S… Speaker 2 (part 3)
you will never go to the next phase.
7:06
S… Speaker 2 (part 3)
So what is the job of a dabbler?
7:08
S… Speaker 2 (part 3)
When you come to dabbling,
7:09
S… Speaker 2 (part 3)
your job is to play with as many tools as possible.
7:14
S… Speaker 2 (part 3)
Try, try,
7:15
S… Speaker 1 (part 3)
try.
7:16
S… Speaker 2 (part 3)
Play with a lot of tools.
7:17
S… Speaker 2 (part 3)
When you play with a lot of AI tools that are out there,
7:20
S… Speaker 2 (part 3)
they're free.
7:21
S… Speaker 2 (part 3)
Out there.
7:22
S… Speaker 2 (part 3)
I have an understanding.
7:23
S… Speaker 2 (part 3)
There's a lot of tools.
7:25
S… Speaker 2 (part 3)
Oh, I need it.
7:26
S… Speaker 2 (part 3)
Oh,
7:26
S… Speaker 1 (part 3)
I need it. Oh, I need it.
7:27
S… Speaker 1 (part 3)
I need it.
7:27
S… Speaker 1 (part 3)
I need it.
7:28
S… Speaker 2 (part 3)
I need it.
7:28
S… Speaker 2 (part 3)
I need it. I need it.
7:29
S… Speaker 2 (part 3)
I need it.
7:30
S… Speaker 2 (part 3)
I need it.
7:31
S… Speaker 2 (part 3)
I need it. I need it.
7:32
S… Speaker 2 (part 3)
I need it.
7:33
S… Speaker 2 (part 3)
I need it.
7:34
S… Speaker 2 (part 3)
I need it.
7:34
S… Speaker 2 (part 3)
I need it. I need it.
7:39
S… Speaker 2 (part 3)
So as you dabble with a lot of AI tools,
7:42
S… Speaker 2 (part 3)
you'll get the full spectrum of things.
7:44
S… Speaker 1 (part 3)
So do you advise, if you have a problem statement,
7:45
S… Speaker 1 (part 3)
if you have a problem statement,
7:46
S… Speaker 1 (part 3)
if you have a problem
7:51
S… Speaker 1 (part 3)
statement, what do you advise?
7:54
S… Speaker 2 (part 3)
So when you don't know anything,
7:57
S… Speaker 2 (part 3)
and you go directly with a big problem statement,
7:59
S… Speaker 1 (part 3)
maybe a presentation or pitch deck.
8:02
S… Speaker 2 (part 3)
If you have any problems,
8:03
S… Speaker 2 (part 3)
you will try 10,
8:07
S… Speaker 2 (part 3)
15, 20 different tools.
8:08
S… Speaker 2 (part 3)
You will figure out what is good with what.
8:10
S… Speaker 2 (part 3)
For example,
8:14
S… Speaker 2 (part 3)
I don't know.
8:14
S… Speaker 2 (part 3)
Let's say Chronicle.
8:15
S… Speaker 1 (part 3)
You try.
8:18
S… Speaker 2 (part 3)
you like something about it and then you try one more software,
8:21
S… Speaker 2 (part 3)
let's say Gama.
8:22
S… Speaker 2 (part 3)
You realize,
8:23
S… Speaker 2 (part 3)
I really like Gama for presentation.
8:25
S… Speaker 2 (part 3)
But you know what?
8:26
S… Speaker 2 (part 3)
Chronicle is created in a presentation.
8:28
S… Speaker 2 (part 3)
Gandhi is in social media.
8:29
S… Speaker 2 (part 3)
So you explore your new ways.
8:32
S… Speaker 1 (part 3)
Correct.
8:34
S… Speaker 2 (part 3)
Yeah. So that is very important.
8:36
S… Speaker 2 (part 3)
You have acknowledged.
8:37
S… Speaker 2 (part 3)
Then you have dabbled.
8:39
S… Speaker 2 (part 3)
You played with like 10,
8:40
S… Speaker 2 (part 3)
20, 50, 100 tools.
8:41
S… Speaker 2 (part 3)
Then your job is to amplify.
8:43
S… Speaker 2 (part 3)
What is amplify?
8:45
S… Speaker 2 (part 3)
50, 60,
8:46
S… Speaker 2 (part 3)
70 tools to play.
8:47
S… Speaker 2 (part 3)
Free trials.
8:48
S… Speaker 2 (part 3)
You understood the power of AI.
8:51
S… Speaker 2 (part 3)
You understood what works well for what.
8:52
S… Speaker 2 (part 3)
Eventually you will realize.
8:54
S… Speaker 2 (part 3)
If I don't have tools.
8:55
S… Speaker 2 (part 3)
This is what
9:01
S… Speaker 2 (part 3)
I'm getting to.
9:01
S… Speaker 1 (part 3)
Eventually.
9:02
S… Speaker 2 (part 3)
So you amplify them.
9:04
S… Speaker 1 (part 3)
Now.
9:04
S… Speaker 2 (part 3)
It's about depth.
9:06
S… Speaker 2 (part 3)
Not width anymore.
9:08
S… Speaker 2 (part 3)
So, okay,
9:08
S… Speaker 2 (part 3)
Claude became your answer for,
9:12
S… Speaker 2 (part 3)
let's say, assistance and everything.
9:13
S… Speaker 2 (part 3)
So,
9:14
S… Speaker 2 (part 3)
Claude has this possibility.
9:15
S… Speaker 2 (part 3)
What is memory?
9:17
S… Speaker 2 (part 3)
Memory needs to manage something.
9:19
S… Speaker 2 (part 3)
Prompting Claude needs to be done.
9:21
S… Speaker 2 (part 3)
How can I go deeper into automating things?
9:24
S… Speaker 1 (part 3)
Connectors.
9:25
S… Speaker 2 (part 3)
How do you
9:30
S… Speaker 2 (part 3)
play with co -work?
9:31
S… Speaker 1 (part 3)
How do you play with Claude?
9:33
S… Speaker 2 (part 3)
I will go deep into it.
9:34
S… Speaker 2 (part 3)
Then you go to console .antropic .com.
9:36
S… Speaker 2 (part 3)
There isn't a playground.
9:37
S… Speaker 2 (part 3)
There isn't a lot of parameters.
9:39
S… Speaker 2 (part 3)
There isn't a lot of parameters.
9:43
S… Speaker 2 (part 3)
Then you learn a playground.
9:44
S… Speaker 2 (part 3)
There's also a lot of managed agents.
9:46
S… Speaker 2 (part 3)
You go deep.
9:49
S… Speaker 2 (part 3)
You become an expert in a bunch of tools.
9:53
S… Speaker 1 (part 3)
That is the phase here.
9:55
S… Speaker 2 (part 3)
When you do this,
9:56
S… Speaker 2 (part 3)
you will...
10:00
S… Speaker 1 (part 3)
actually end up building your toolkit.
10:03
S… Speaker 1 (part 3)
Because you've
10:07
S… Speaker 1 (part 3)
gone through the whole dabbling at this point of time,
10:10
S… Speaker 1 (part 3)
you have amplified,
10:11
S… Speaker 1 (part 3)
let's say tool A to be a presentation tool.
10:14
S… Speaker 1 (part 3)
So, you go deep into it.
10:16
S… Speaker 1 (part 3)
You understand what you can do.
10:17
S… Speaker 1 (part 3)
How well can you perform with it?
10:18
S… Speaker 1 (part 3)
That is what amplifying will allow you.
10:21
S… Speaker 1 (part 3)
So, you build your toolkit.
10:22
S… Speaker 2 (part 3)
Now,
10:22
S… Speaker 1 (part 3)
amplification,
10:23
S… Speaker 1 (part 3)
when you're playing with it,
10:25
S… Speaker 1 (part 3)
automatically,
10:26
S… Speaker 1 (part 3)
there's a characteristic that kicks in.
10:27
S… Speaker 1 (part 3)
A characteristic is,
10:29
S… Speaker 1 (part 3)
you expect more from AI.
10:31
S… Speaker 2 (part 3)
Correct.
10:32
S… Speaker 1 (part 3)
Then,
10:33
S… Speaker 1 (part 3)
that will open up a new evolution for you.
10:35
S… Speaker 1 (part 3)
No, agents don't forget all of them.
10:37
S… Speaker 1 (part 3)
A new thing starts
10:41
S… Speaker 1 (part 3)
running in your head.
10:44
S… Speaker 1 (part 3)
Right.
10:44
S… Speaker 1 (part 3)
What is that new thing?
10:45
S… Speaker 1 (part 3)
How can I do more?
10:47
S… Speaker 1 (part 3)
How can I do more?
10:49
S… Speaker 1 (part 3)
How can I solve more problems,
10:50
S… Speaker 1 (part 3)
more problems,
10:51
S… Speaker 1 (part 3)
more problems?
10:52
S… Speaker 1 (part 3)
That is when you navigate to a problem solver.
10:54
S… Speaker 1 (part 3)
In ADAPT,
10:56
S… Speaker 1 (part 3)
P is problem solving.
10:57
S… Speaker 1 (part 3)
You come to problem solving and this is a very important phase
11:01
S… Speaker 1 (part 3)
in everyone.
11:02
S… Speaker 1 (part 3)
Right.
11:03
S… Speaker 1 (part 3)
Now, if you have to solve this phase,
11:03
S… Speaker 1 (part 3)
the thing that you said,
11:06
S… Speaker 1 (part 3)
you will think about how do I solve a problem.
11:08
S… Speaker 1 (part 3)
Now, if you have a problem,
11:10
S… Speaker 1 (part 3)
The beautiful thing is,
11:12
S… Speaker 1 (part 3)
you know, already amplified something.
11:13
S… Speaker 1 (part 3)
Bunch of tools.
11:14
S… Speaker 1 (part 3)
On top of that,
11:15
S… Speaker 1 (part 3)
you have dabbled with so many more tools.
11:17
S… Speaker 1 (part 3)
Which is in your loose memory.
11:18
S… Speaker 1 (part 3)
Now the problem is,
11:20
S… Speaker 1 (part 3)
by the time you get to this point,
11:22
S… Speaker 1 (part 3)
you will be able to connect dots.
11:23
S… Speaker 1 (part 3)
Hey, this tool,
11:25
S… Speaker 1 (part 3)
this tool, this tool,
11:25
S… Speaker 1 (part 3)
this mood workflow.
11:26
S… Speaker 1 (part 3)
Presentation is an example.
11:28
S… Speaker 1 (part 3)
Because we're talking about that example,
11:30
S… Speaker 1 (part 3)
right? Presentation is a design.
11:33
S… Speaker 1 (part 3)
There's so many things.
11:34
S… Speaker 1 (part 3)
Information.
11:35
S… Speaker 1 (part 3)
A content in the structure,
11:38
S… Speaker 1 (part 3)
then design in the presentation,
11:42
S… Speaker 1 (part 3)
eventually presentation.
11:43
S… Speaker 1 (part 3)
So now you have four things.
11:44
S… Speaker 1 (part 3)
Oh, deep research on this topic.
11:46
S… Speaker 1 (part 3)
I want to start this mug selling business.
11:49
S… Speaker 1 (part 3)
In amplification,
11:52
S… Speaker 1 (part 3)
in dabbling,
11:53
S… Speaker 1 (part 3)
I tried 50 tools.
11:54
S… Speaker 1 (part 3)
In that I amplified,
11:55
S… Speaker 1 (part 3)
let's say perplexity,
11:56
S… Speaker 1 (part 3)
deep research.
11:56
S… Speaker 1 (part 3)
Perplexity,
11:58
S… Speaker 1 (part 3)
deep research.
11:58
S… Speaker 1 (part 3)
So model counts in the option because you have.
12:01
S… Speaker 1 (part 3)
Amplified them.
12:02
S… Speaker 1 (part 3)
You've gone deep into them.
12:03
S… Speaker 1 (part 3)
So you don't only know deep research.
12:05
S… Speaker 1 (part 3)
You also know something called as model council exists inside of perplexity.
12:09
S… Speaker 1 (part 3)
That's what it is.
12:10
S… Speaker 1 (part 3)
Your question is,
12:12
S… Speaker 1 (part 3)
if you come to the model,
12:12
S… Speaker 1 (part 3)
any model called any model,
12:14
S… Speaker 1 (part 3)
you have done it first.
12:15
S… Speaker 1 (part 3)
That is the model council.
12:19
S… Speaker 1 (part 3)
So it debates with each other models to give you an answer.
12:22
S… Speaker 1 (part 3)
Like five people are working from five different companies to give you an answer.
12:26
S… Speaker 1 (part 3)
I will tell you that I have a
12:30
S… Speaker 1 (part 3)
project inside of my chat GPT only,
12:34
S… Speaker 1 (part 3)
which is called chat GPT project,
12:35
S… Speaker 1 (part 3)
where I also have all the transcript meetings of
12:39
S… Speaker 1 (part 3)
all my investors.
12:40
S… Speaker 2 (part 3)
Okay.
12:41
S… Speaker 1 (part 3)
Every meeting that has happened with my investors,
12:43
S… Speaker 1 (part 3)
which are online,
12:44
S… Speaker 1 (part 3)
I'm meeting the transcripts.
12:45
S… Speaker 1 (part 3)
So they know how,
12:47
S… Speaker 1 (part 3)
what questions they ask when I say something,
12:49
S… Speaker 1 (part 3)
a presentation,
12:50
S… Speaker 1 (part 3)
let's say this is for my investor.
12:51
S… Speaker 1 (part 3)
If presentation,
12:53
S… Speaker 1 (part 3)
this is the presentation I'm planning to present to my investors.
12:57
S… Speaker 1 (part 3)
What do you think they will ask?
12:59
S… Speaker 1 (part 3)
And I can vet that also.
13:01
S… Speaker 1 (part 3)
This is problem solving.
13:02
S… Speaker 1 (part 3)
Problem solving is not a tool in most of the cases.
13:06
S… Speaker 1 (part 3)
It's very unlikely.
13:07
S… Speaker 1 (part 3)
You need a sequence of tools.
13:09
S… Speaker 1 (part 3)
And you have to connect with each one of them.
13:11
S… Speaker 1 (part 3)
That in a way becomes a workflow.
13:13
S… Speaker 1 (part 3)
Sometimes it becomes a workflow every day.
13:16
S… Speaker 1 (part 3)
That becomes an AI agent.
13:19
S… Speaker 1 (part 3)
When you become
13:23
S… Speaker 1 (part 3)
a very strong problem solver.
13:26
S… Speaker 1 (part 3)
You by default automatically move to the next phase of
13:30
S… Speaker 1 (part 3)
T, of ADAPT,
13:31
S… Speaker 1 (part 3)
which is tying it all together.
13:33
S… Speaker 1 (part 3)
You put different workflows.
13:36
S… Speaker 1 (part 3)
An AI agent,
13:37
S… Speaker 1 (part 3)
a wipe -coded product,
13:38
S… Speaker 1 (part 3)
a cloud project,
13:41
S… Speaker 1 (part 3)
using co -work and you stitch everything together for it to work in symphony.
13:45
S… Speaker 1 (part 3)
You explore things like paperclip.
13:47
S… Speaker 1 (part 3)
What is paperclip?
13:49
S… Speaker 1 (part 3)
Paperclip is like one AI agent.
13:52
S… Speaker 1 (part 3)
Right.
13:52
S… Speaker 1 (part 3)
Think about 15 different AI agents working like your employees.
13:56
S… Speaker 1 (part 3)
Paperclip is the CEO,
13:58
S… Speaker 1 (part 3)
which is also an AI where you,
14:01
S… Speaker 1 (part 3)
you just tell the CEO,
14:02
S… Speaker 1 (part 3)
you're the investor.
14:02
S… Speaker 1 (part 3)
You just tell the CEO,
14:03
S… Speaker 1 (part 3)
this is what needs to be executed.
14:04
S… Speaker 1 (part 3)
The CEO figures out marketing angel,
14:06
S… Speaker 1 (part 3)
product angel,
14:07
S… Speaker 1 (part 3)
design angel,
14:08
S… Speaker 1 (part 3)
splits up all the work and gets it all executed.
14:11
S… Speaker 1 (part 3)
Eventually, you've got task list,
14:13
S… Speaker 1 (part 3)
which is all AI.
14:16
S… Speaker 1 (part 3)
Everything is an AI agent inside of it.
14:18
S… Speaker 1 (part 3)
That's paperclip.
14:19
S… Speaker 2 (part 3)
Now,
14:19
S… Speaker 1 (part 3)
the reason I'm giving an example,
14:21
S… Speaker 1 (part 3)
these are not perfect yet.
14:22
S… Speaker 1 (part 3)
And one year later,
14:23
S… Speaker 1 (part 3)
when we are talking about it,
14:24
S… Speaker 1 (part 3)
I'll probably show you on my phone that this is happening right now.
14:27
S… Speaker 1 (part 3)
Because that's how tech is moving.
14:29
S… Speaker 1 (part 3)
But when you're all tied together,
14:30
S… Speaker 1 (part 3)
that's when magic happens.
14:32
S… Speaker 1 (part 3)
People who figure out how to
14:36
S… Speaker 1 (part 3)
tie things together are basically orchestrators.
14:39
S… Speaker 1 (part 3)
And this is how you become a great AI orchestrator.
14:43
S… Speaker 1 (part 3)
This is how.
14:45
S… Speaker 1 (part 3)
And if you have a problem on your side,
14:47
S… Speaker 1 (part 3)
you're thinking AI first.
14:49
S… Speaker 1 (part 3)
Yeah.
14:49
S… Speaker 2 (part 3)
That's it.
14:50
S… Speaker 1 (part 3)
You can connect with them.
14:53
S… Speaker 1 (part 3)
You do AI first problem solving.
14:56
S… Speaker 1 (part 3)
When you do AI first problem solving,
14:58
S… Speaker 1 (part 3)
the kind of solutions you can build.
15:01
S… Speaker 1 (part 3)
a human can't do.
15:02
S… Speaker 1 (part 3)
It's not practically not possible.
15:06
S… Speaker 1 (part 3)
We've spoken about examples.
15:07
S… Speaker 1 (part 3)
That is what an AI orchestrator is.
15:10
S… Speaker 1 (part 3)
That leads,
15:11
S… Speaker 1 (part 3)
a good AI orchestrator is obviously a very good AI first problem
15:15
S… Speaker 1 (part 3)
solver or an AI generalist.
15:17
S… Speaker 2 (part 3)
Got it.
15:18
S… Speaker 1 (part 3)
So that is your ADAPT framework.
15:22
S… Speaker 1 (part 3)
Simple, but works really,
15:24
S… Speaker 1 (part 3)
really well.
15:25
S… Speaker 1 (part 3)
In fact,
15:27
S… Speaker 1 (part 3)
In my internal training,
15:29
S… Speaker 1 (part 3)
this ADAPT framework,
15:31
S… Speaker 1 (part 3)
me and my team and Dilip and everybody worked on
15:35
S… Speaker 1 (part 3)
this presentation of ADAPT.
15:37
S… Speaker 1 (part 3)
I think it's a 2025 page presentation of this ADAPT
15:41
S… Speaker 1 (part 3)
framework. I also have a link on the page.
15:42
S… Speaker 1 (part 3)
Super.
15:43
S… Speaker 1 (part 3)
For people who are genuinely curious,
15:45
S… Speaker 1 (part 3)
they can click and read it.
15:47
S… Speaker 1 (part 3)
It's free.
15:48
S… Speaker 1 (part 3)
Let them use it.
15:49
S… Speaker 1 (part 3)
Many people have a dabbling and they think
15:53
S… Speaker 1 (part 3)
that is what AI is.
15:56
S… Speaker 2 (part 3)
Cool, cool.
15:57
S… Speaker 1 (part 3)
You get my
16:01
S… Speaker 1 (part 3)
point, right?
16:02
S… Speaker 1 (part 3)
Okay,
16:03
S… Speaker 1 (part 3)
presentation,
16:03
S… Speaker 1 (part 3)
data,
16:05
S… Speaker 2 (part 3)
data.
16:05
S… Speaker 2 (part 3)
So,
16:07
S… Speaker 2 (part 3)
I get what
16:12
S… Speaker 1 (part 3)
you mean. And they think that is AI.
16:15
S… Speaker 1 (part 3)
That's a problem.
16:17
S… Speaker 1 (part 3)
Because if you go further out,
16:19
S… Speaker 1 (part 3)
you will not grow as fast.
16:21
S… Speaker 1 (part 3)
But here is the beautiful thing,
16:24
S… Speaker 1 (part 3)
no? The narrative...
16:26
S… Speaker 1 (part 3)
And there's a little negative effect.
16:28
S… Speaker 1 (part 3)
Right?
16:29
S… Speaker 1 (part 3)
I think AI works for jobs.
16:32
S… Speaker 1 (part 3)
Jobs works for jobs.
16:34
S… Speaker 1 (part 3)
That is not true.
16:36
S… Speaker 1 (part 3)
That is not untrue.
16:37
S… Speaker 1 (part 3)
But jobs have evolved.
16:41
S… Speaker 1 (part 3)
You have to adapt to life.
16:43
S… Speaker 1 (part 3)
You have to adapt to life.
16:45
S… Speaker 1 (part 3)
Because a new job will evolve and
16:49
S… Speaker 1 (part 3)
you will be the best fit for that.
16:50
S… Speaker 2 (part 3)
What kind of jobs can come?
16:53
S… Speaker 2 (part 3)
So,
16:55
S… Speaker 1 (part 3)
See, there are two,
16:56
S… Speaker 1 (part 3)
what kind of jobs can come?
16:57
S… Speaker 1 (part 3)
There are two line of thoughts here.
16:59
S… Speaker 1 (part 3)
Okay,
17:00
S… Speaker 1 (part 3)
line of thought, the jobs that will come that never existed before.

මෙම පරිවර්තනය AI විසින් ජනනය කරන ලදී (ස්වයංක්රීය කථාව හඳුනා). දෝෂ අඩංගු විය හැක - විවේචනාත්මක භාවිතය සඳහා මුල් ශ්රව්ය එරෙහිව තහවුරු. AI ප්‍රතිපත්තිය

❤️ ආදරය STT.ai? ඔබේ මිතුරන්ට කියන්න!
සාරාංශය
මෙම පරිවර්තනය AI සාරාංශයක් ජනනය කිරීමට සාරාංශ ක්ලික් කරන්න.
සාරාංශගත කිරීම...
මෙම පරිවර්තනය ගැන AI අහන්න
මෙම පරිවර්තනය ගැන යමක් අහන්න - AI අදාළ කොටස් සොයා සහ පිළිතුරු දෙනු ඇත.
වෙනත් පරිවර්තනය කරන්න