Podcast with Tamir Shklaz Part 3
In this thought-provoking podcast segment, Tamir Shklaz delvesinto the profound impact of AI on the education system, job landscape, and the
next generation. Discover how AI is reshaping traditional roles and creating
exciting opportunities like the "prompt engineer." While AI offers
boundless possibilities, Tamir also delves into the critical need to address
ethical concerns and ensure AI serves humanity's best interests. Join the
conversation as he unveils the indispensable role of coding as a powerful tool
for learning and problem-solving, propelling us towards a future of knowledge
In this podcast segment, Tamir Shklaz discusses the impact ofartificial intelligence (AI) on education and the future of work. He highlights
how AI accelerates job changes, creates new job roles like "prompt engineer,"
and empowers teachers by automating administrative tasks. He emphasizes the
importance of teaching kids how to use AI effectively as a powerful educational
tool. However, he acknowledges the cautious side of AI and the need for
collective intelligence to ensure AI is used ethically and safely. Looking at
the future of work, Tamir predicts that AI will be embedded in everyday tools
and that coding will remain a crucial skill for problem-solving and learning.
He also believes coding will become a powerful tool for constructing mental
models and understanding complex concepts in various subjects.
[00:00:00] Andrew Liew: I see. Now that's a verycompelling, interesting view out there for parents because I came from the
Singapore education system, but having worked overseas in Australia and America
and even China, I have seen the parents in different countries.
[00:00:16] Andrew Liew: It's up to them to decide. ForAsians no, it's up for the parents to decide. Coming back, what is your view on
the future of artificial intelligence with reference to educating the next
[00:00:28] Tamir Shklaz: kids? Yeah, so so AI is going tohave a monumental effect on education. The, first one is that it just
accelerates what I was talking about previously the acceleration of how quickly
jobs change, because we've seen now the what chat GPT able to do, and how it
[00:00:45] Tamir Shklaz: Not only replaced some jobsentirely things like customer service content creation. It's those a lot of
those jobs like BuzzFeed, for example, got rid of their entire team of content
creators or an entire [00:01:00] subset oftheir team because of, because it was just cheaper for them to have one person
and use AI to, to be able to create the, content that they needed, but then it
also created all the, these new jobs.
[00:01:10] Tamir Shklaz: Now, all of a sudden, thisprompt engineer is a job that didn't exist six months ago, but someone who is
able to communicate with these artificial intelligence systems to. Get the
output that they want to know how to tweak their words and ask for the right
things that is a new job that didn't exist 6 months ago.
[00:01:27] Tamir Shklaz: So that's 1 big effect. It'sgoing to have the next 1 is going to be on just the way in which education is
done. So 1 kids obviously going to need to learn how to use AI. So I think it's
going to be an incredibly powerful tool of. So previously, like when I needed
to learn something new or when people were learning from New York, Google it.
[00:01:45] Tamir Shklaz: If I wanted to learn a newprogramming language, I'll Google how to learn Java, how to learn C Yes, AI can
now create like a customized lesson plan for you. This is how you can learn
this thing and get very precise answers. So you need to learn how to actually
use those tools. It [00:02:00] will then alsoempower teachers.
[00:02:01] Tamir Shklaz: So a huge, time suck forteachers is just assessment and grading. And being able to create tests and
then grade those tests. AI has the potential to save teachers thousands of
hours. And most of their time, like a lot of their time in all honesty,
teachers are spent on admin, not on teaching. And AI has real potential to free
up that time.
[00:02:22] Tamir Shklaz: To focus teachers on what they reallyshould be doing and what they enjoy doing, which is interacting and inspiring
kids. And the last 1 is that you just always have access to a 24 7 private unit
to answer your questions. Like Khan Academy recently released a chat bot along
with their their math platform.
[00:02:39] Tamir Shklaz: We're planning on introducing aa coding chat bot for students to ask questions on coding at any point in time.
And so it will. really democratize access to high quality education because
everyone will have in a form a kind of tutor at the touch of their fingertips
all the time. Oh,
[00:02:56] Andrew Liew: it sounds very, exciting.
[00:02:58] Andrew Liew: Like you said now it creates [00:03:00] a new types of jobs called promptengineer. And as well as you also mentioned. Education can be more
personalized, but let's look at also the cautious side of things. Like these,
just I think one or two weeks ago, the entire Silicon Valley tech leaders
actually band together and say, hell, this thing is really getting crazy, like
getting do you see there will be one day that the robots will take over?
[00:03:22] Tamir Shklaz: There it's going to be a whole,it's a whole other can of worms that we're opening here, right? Look I think in
the short term, it's certainly incredibly exciting and the potential is,
incredible for what, it can do for teachers and students. Long term, it's so
hard to tell because of the nature of exponential change.
[00:03:41] Tamir Shklaz: And because of how little weactually know about these systems. I really don't know, is the answer on what
will happen in the long term. There is certainly a lot of concern, I think Elon
Musk, I think Steve Wozniak, co founder of Apple, signed the petition. I think
their particular one at [00:04:00] the moment,like the concern coming from them is that it's too quick, that the, transition
is going to be too violent and we, don't have enough time to adapt to the
[00:04:10] Tamir Shklaz: So we need to slow down therates of transition so that we get time to be like, okay, cool. So we're not
driving into 2000 degree water immediately. That's the one aspect. The other
aspect, which is the much more existential threat, which is is AI going to
replace all jobs? Is AI going to be misaligned and go terminate on us?
[00:04:31] Tamir Shklaz: Those a different breed ofquestions, which I don't, I think is more common amongst actually, no, that's
not necessarily not necessarily true. I recently listened to, I forgot his full
name, but it was on the Lex Friedman podcast, a very well established AI
researcher talking about the dangers of this, but that just paints a picture of
the landscape of, how society is reacting to it.
[00:04:55] Tamir Shklaz: Yeah, I
[00:04:56] Andrew Liew: mean like the one of the reasonwas about with as a [00:05:00] parent or as aneducator I think we are always wanting the best for our kids and one of the
concern is Misinformation because right now AI or chat GPT doesn't have a moral
conscious of his own so whatever we feed it's there's always a story you feed
the, bad wolf, you become the bad wolf, you feed the good wolf, you become the
[00:05:21] Andrew Liew: Like, how do we actually ashumanity actually put some safe real guts to ensure that the technology or AI
is being served, used to serve human serve kids, especially kids because they
are the next generation. What
[00:05:36] Tamir Shklaz: is your view on it? Look, Ithink this, points even more so to the, Fundamental importance of educating
kids on this, right?
[00:05:43] Tamir Shklaz: Like the answers. I don't knowthe that I don't have the answer, but where there is the answer is in the
collective intelligence of the human species and as we learn and grapple with
this problem and the more people that are intimately. involved and understand
these [00:06:00] problems, the more likely weare to come to a correct solution.
[00:06:03] Tamir Shklaz: So if we're able to get kids tostart just being aware of the stuff and understand it, by the time they get
older, they'll be more able to participate in the politics, in the policy
design and the engineering that would be able to solve. Solve these problems.
Having said that, I do think there are a few principles that could be applied
to try and safeguard against this.
[00:06:22] Tamir Shklaz: I think OpenAI, from what I'veseen, is doing a good job on this, which is to make the, algorithm and what's
happening as open source as possible. Be as transparent as possible in regards
to this. Try and get as many different perspectives from different cultures and
societies as possible. I think we're in a world where the...
[00:06:42] Tamir Shklaz: Access to this is moredemocratized, more people contribute to it, as opposed to a world where the
power is highly centralized, controlled by a few people. That world, the former
world of one where it's more democratized, is far more likely to come up with a
balanced AI, an AI which is more accurate to [00:07:00]reality.
[00:07:00] Tamir Shklaz: Yeah, but it is a, it'scertainly a huge challenge.
[00:07:03] Andrew Liew: Yeah it's a challenge that wewon't know, but like you said, I think I'm in line with you as well, man. If it
is very democratized technology, human beings are able to have collective
intelligence and ethical grounds to train the next generation.
[00:07:16] Andrew Liew: Now let's talk about anotherinteresting view because AI has now, especially generative, has created so many
new jobs and of course it has a creative disruption and does eliminate some
jobs because a lot of repetitive jobs is replaced by automation. And now apart
from training, educating people, the future of work is gonna change.
[00:07:35] Andrew Liew: What is your view on the futureof work, tech people? With people using tech in the next 10 years, we reference
to preparing for the next generation.
[00:07:45] Tamir Shklaz: Yeah. So think similar things ofwhat we discussed, of just how rapidly it's gonna change. I, like, I really
like this, one analogy of because 'cause this change is exponential, it is so
difficult to predict what's gonna happen in 10 years because, [00:08:00] Human beings have really bad intuitionswhen it comes to exponential growth.
[00:08:05] Tamir Shklaz: We're, very used to seeingthings on like a linear scale. To give a, an example of this is imagine you had
a lily pad on a pond and every day. This lily pad doubles in size. So on, let's
say on day one, it's one centimeter across day two, it's two centimeters across
day three, four centimeters, eight et cetera, and keeps, doubling like that.
[00:08:26] Tamir Shklaz: I, you probably know the answerto this. So, it'd be unfair to ask this question, but if you ask people this
question, okay, on which day is this lily pad likely to cover half? of the
lake, right? So it doubles every day. And so let's assume it takes 30 days for
it to cover the whole lake. On which day would it cover half of the lake?
[00:08:45] Tamir Shklaz: Most people by intuition aregoing to say, okay, if it takes 30 days to cover the full lake, it will take 15
days to cover half the lake. But that is an exponential growth. When it will
cover half the lake is on day 29, the day before. And that's [00:09:00] Is usually a very mind blowing realizationto people. And so if we take that analogy to, to work and the future of
workers, It just shows how crazy it's going to become in regards to What the
landscape of work is, going to change like having said that I can speak to a
few trends that I do think we can predict with reasonable confidence.
[00:09:24] Tamir Shklaz: 1 is that is going to beembedded in everyday tools. So we started seeing this already. It's in Microsoft
office suites and Google Docs, but everything is just going to have an AI
partner. You'll be using it on any tool from Google meets to zoom to. Your
presentations, et cetera. You're always going to have some form of AI partner
to hold your hand or assist you in the work that you're doing.
[00:09:48] Tamir Shklaz: I think no code or low codesystems are going to become more and more mainstream, particularly like AI is
another way. You're starting to see and there's a large[00:10:00]group of people being like, is AI going to replace software engineers and I've
started to see how AI can really empower low code systems, which is how you
build a software system in that a layman could do without necessarily needing
to know of code, because now you can just communicate with the machine and In a
very human way, do this, do that, do this. So those are going to just become
more powerful and become more and more mainstream. Yeah, I think those are the
two ways that AI is really going, or at least the two trends that I've, been
seeing when it comes to AI.
[00:10:33] Andrew Liew: So talking along that line it'sinteresting, like you mentioned no code and low code system will is gaining
[00:10:40] Andrew Liew: And we are seeing a lot of theseco pilots. So as a parent, Why would they want to get their kids to learn
[00:10:47] Tamir Shklaz: What's your view on that? Greatquestion. So I think it's not at the place. So first and foremost, AI is not
nearly at the place where it's going to replace software engineers, where it's
where it is having an [00:11:00] impact is onlike the menial work that software engineers need to do.
[00:11:03] Tamir Shklaz: And it's acting as a hugeproductivity tool, but the large scale systems thinking AI is not yet capable
of doing. So to design full applications and full systems we're not there yet.
A good example of this is if you look at, there's this website called LeetCode,
which is a repository of these really challenging programming problems.
[00:11:25] Tamir Shklaz: GPT 4, the most recent AI modelat at the time of this recording aces, the ones that humans have already
solved. So the, hard ones, but if, there's an existing human solution to it,
the really hard ones where it's, we don't where there aren't many solutions available,
AI completely flops.
[00:11:44] Tamir Shklaz: It's, unable to solve. To solvethose problems, and so there's certainly still a long way to go before we're
looking at a massive shift within the world of software engineering. Having
said that, I think there's 2 other aspects to look in this on [00:12:00] why coding is still going to be animportant skill.
[00:12:02] Tamir Shklaz: 1 is that AI is just going tocreate a different programming language. So instead of you needing to write
Python, you're going to need to write AI to code. A program, but you're still
going to need to know how to write that. So if I put a software engineer in
front of chatGPT and say, use chatGPT to build me a website, or if I put
someone who's never written software and say, use chatGPT to read me a website,
the software engineer is going to be 10x more productive and 10x better at
knowing exactly what to ask and how to...
[00:12:35] Tamir Shklaz: How to make it work, becausewhat you need to do, what software engineering is doing at its core is solving
problems and writing very detailed specifications. When this happens, do that.
When these things happen, do this. The way in which that logic is implemented
at the moment is in Python. But as AI or any other programming language, as AI
gets better and better, we're still going to need to specify that logic, but
it's just going to be in more [00:13:00] of ahuman readable text as opposed to programming.
[00:13:03] Tamir Shklaz: So programming is just going toevolve. It's not going to be replaced. The last one, and this is an interesting
idea we haven't spoken about we talked about. The most important thing is
learning how to learn. I see coding as becoming an incredibly powerful tool of
learning how to learn other subjects.
[00:13:19] Tamir Shklaz: So in a similar way that readingin the early 20th century became a prerequisite to All of existing schooling,
like all of our schooling depends on you needing to know how to read because
that's how you then learn geography and history and all these other things. I
foresee a future where coding becomes the equivalent of reading for you to
learn these things.
[00:13:41] Tamir Shklaz: The reason why that is, is thatcoding provides an incredible interface to how we actually understand things.
So there's a theory of, learning called constructionism, which was created by
this guy called Simeon Papa. I'm probably butchering his name back in like the
1980s or 70s. He was an MIT [00:14:00]researcher.
[00:14:00] Tamir Shklaz: He, was originally involved inwhat eventually became Lego. And his, theory was that we don't consume and we
don't learn by consuming knowledge or consuming information we learn by
building, we learn by constructing and coding provides an incredible interface
for us to create mental models of the world around us.
[00:14:19] Tamir Shklaz: Let's take like a physicsexample like if I want to understand. The Newton's laws of motions, like how,
if I throw a ball across the sky, how that ball is going to move, I could go
ahead and learn the math the way that it's traditionally done and memorize
these formulas, but a much deeper understanding will be given to me is if I
code that, if I take the math and create a simulation that shows what the board
[00:14:45] Tamir Shklaz: And now I start playing aroundwith the variables. I start saying, okay, let's make gravity 30 instead of 20.
Let's make the ball three times as massive. And what you can then start to do
is you play around with your knowledge, and where it doesn't work, it shows [00:15:00] a flaw in your logic. It shows a flaw inyour understanding.
[00:15:03] Tamir Shklaz: Like when you get an error incode, it means okay, there's something I didn't understand here. And coding
will then give us a way of getting instant real time feedback on if the models
that we have about the world are accurate or not. And that's going to be a huge
tool for learning other subjects..