product strategy 101 | competitive radar & Porter's forces

Previously I went to uh John Hopkins University for a lecture >> and there's another speaker who is the head of AI adviser uh 2024 early 2024 we were talking about April 2024 at that time she already told me that it is

proven that AI could fasten their drug development process drug discovery process. Why is made in China so cheap?

process. Why is made in China so cheap?

Made in China is cheap originally I thought because the labor cost is low.

Not only that men used to be on now their dose cells it is very brutal out there. a lot of uh things that we are

there. a lot of uh things that we are going to mention today. You can also realize that for example Hamzas in Hamza's report you already mentioned the portage forces for analyzing open AI as

well because um this is the value chain of the company is extremely important when you analyze your competitors and your buyer power and your uh supplier power etc. So last week at the end of

the class we were talking about the criterias for competitive advantage. Who

can tell me the three criterias for competitive advant advantage?

>> Must provide value.

>> Must provide value.

>> Rare in the market.

>> You're unique in the market >> and hard to imitate or substitute.

>> Yes. But who are your competitors, right? Who are your comp who are you

right? Who are your comp who are you competing? When you talk about the

competing? When you talk about the competitive advantage, there is this competitive radar uh which is a very helpful a helpful framework for for you to analyze your competitors. Uh the

inner circle is your direct competitors.

They have over overlap with you on the product and the technology and the market that they're operating. And the

more outer circles that they're going to uh the more indirect they are as your competitors. Um a lot of companies just

competitors. Um a lot of companies just focus on their direct verticals uh their product the overlap products right but uh eventually you want to look at the

need that consumers have the need that your customers have and that's why you have incomp uh indirect competitors.

So for the inner circle there are direct competitors with full product and market overlap and similar tech maturity. And

for the second circle is the adjacencent players who can pivot into your market zone. And the third circle is niche

zone. And the third circle is niche original or earlier stage companies. And

the fourth circle is the substitute or long-term potential threats. Let's say

Figma. Figma as a company the core threat might be like targeting the same market um with similar tech maturity, right? So let's say maybe uh Photoshop

right? So let's say maybe uh Photoshop or canvas. Uh this apps can also allow

or canvas. Uh this apps can also allow designers um to design apps or design um like posters. When we come to the second

like posters. When we come to the second layer uh near thread adjacent players who can pivot into your market soon. Um

maybe we could talk about the apps that allow maybe product managers or designers to draw Warframe, right? low

fidelity wireframe and those could be developed into high fidelity designs one day but right now those startups or those companies are not doing so and the

third one uh the niche original or early stage companies chatd and deli also we have some other apps that could serve the same purpose like the one that I

just used to p to photoshop my uh thumbnail right I didn't use like Figma or uh canvas or any other like high level apps. It's just a mu app that I

level apps. It's just a mu app that I use uh for the thumbnail, right? So

those are like more niche and original and early stage companies um like AI generated uh designs, right? Uh those

are more early stage. Can they do high fidelity design? Not right now, but

fidelity design? Not right now, but maybe one day they can, right? And the

last one is substitutes or long-term potential threats. For example, a lot of

potential threats. For example, a lot of really pro users um they just like they can use draw or they can use like more basic um software but they can achieve a

pretty high performing art as well. So

this is just an example of like multiple layers of the competitor competitive radar. We talk about we talked a lot

radar. We talk about we talked a lot about social media last time as well. Uh

for example YouTube shorts, right?

YouTube short. The core threats might be um um Instagram res or Tik Tok, right?

The near thread might be whatever like video format um kind of like media, right? You could talk about like

right? You could talk about like streaming companies where you could talk about like podcast uh like Spotify which is also doing the videos of podcast on

their apps and then the other layer um you could talk about um niche or regional like like red node used to be very niche uh but right now

it is becoming bigger like lemon eight might be uh very early stage companies and they are growing as well in the US market and there are uh other um

regional apps um like I remember there's an app called line in uh Korea and Japan Japanese market um and they're in

they're indirect uh competitors as well like substitute then you can expand to more entertainment industry where you can use the same like or use different

uh products to still entertain yourself.

So this is the multiple layers of competitive radar.

>> What do you mean by long-term potential threats?

>> Keep an eye on the new technology because you are looking at very similar industry or like very similar uh product that you are targeting right now but there might be new like formats or new

innovation that will come out and sub uh like substitute in the future.

>> Uh for example the open eye uh is already the most advanced company. Do

they still have the long-term potential threats?

>> I think just depends on how long this company stay on this planet. I think

eventually yes.

>> Okay. So maybe in different stage uh they can have only three of the three or something.

>> I think the last one is always there. It

just depends on whether our imagination get there or not.

>> But it's so hard to imagine. Yeah, it is very hard to imagine and that's why when I was doing my side project um and approaching VCs and pres pitching my my

uh product, I realized it is really hard to leave ahead of time like that is an ability and very experience based ability. Um if you're working in venture

ability. Um if you're working in venture capital and the daily job if you're yours is just talking with hundreds of thousands of startups, right? You kind

of train yourself to think forward. Now

it is your time to tell me how you think. What do you think about um Whimo

think. What do you think about um Whimo >> should be like Tesla?

>> Oh Tesla direct uh competitor.

>> Mhm.

>> Yeah.

>> And near threat could be any kind of uh vehicle companies who has a plan to like include automatic driving. Air feral

threat should be like big automatic driving companies in other countries like maybe BU in China or Xiaomi

and indirect like uh like the technology of artificial intelligence and like robotics or all kinds of like new

technology things. I think also like

technology things. I think also like traditional taxi companies could also be like indirect because they're like substitutes but they're not really in the same like no they're in the exact same market they're just not you know

autonomous they're more traditional so I don't know if that should be an indirect or they're not really adjacent because they can't really pivot into autonomous

stuff or yeah so I'd say it's indirect but they are you know I think one of the biggest competitors in the industry even though they don't use the same technology.

>> Yeah, I think when it comes to indirect taxi definitely counts and also other forms of transportation for sure. I'm

very curious actually um Yogesh if you could tell us I'm always very curious.

Okay.

>> And how can go to India market because I feel like in traffic >> actually I was going to tell that uh I just Tesla's new showroom is very near

to my home and I recently visited and in India I guess it's going to be very difficult to tap the market because we

don't have traffic room and I don't know M so Indian vehicle

uh this market of auto driving auto this industry this industry is always dominated by domestic players in India.

Why? Because they only understand the context of the whole business around. So

Uberin worked out Ola work because they know what the driver wants, what the consumer wants and yes Tuktouk is being replaced by Aether. I would recommend

everyone to read about Ather because they are doing a very phenomenal job replacing traditional Tuks with electric vehicles and this is going to be a

phenomenal story coming out of coming out of from India.

Can you say the company name in the chat?

>> Yes. Yes. Aether Industries. It's from

It is listed. It got listed few months back and they are phenomenal. They're going

after everyone.

>> Interesting. So they're doing uh electric vehicles.

>> Yeah. And they are they are not dominating yet. Ola dominates Ather. All

dominating yet. Ola dominates Ather. All

dominates India but Ola is unable to keep up with the scale. Ather is moving like slowly but they know what they are going after.

>> Anyone do autonomous driving?

>> Not yet but I guess if someone does the first would be it. I'm very sure about it because I use it and I feel they have cracked the psychology of Indian

consumer. one of my uh Duke friends um

consumer. one of my uh Duke friends um he came from India and I used to I I saw his video of showing around Indian tra traffic and how to navigate around and

there are a bunch of like psych uh I don't know how what you guys call it but like a cycle cyclers cyclist >> so you sit behind that cyclist and that is

>> like like they they run like storm like that's crazy like um yeah so >> yeah it is a big livelihood for almost so many people. You you should visit

once India to experience.

>> I want to I want to I can imagine Whimo just like stopping there not knowing where to go if they go to Indian market.

>> We don't have we don't have signals. We

don't have traffic rules. I don't think so how it will operate.

>> Even right now in San Francisco like people just bully Whimo so much because they know Whimo cannot hurt them. So

even it is not the pedestrians's turn, they will go across Whimo because they know Whimo is not going to go if there is a passenger in uh in front.

>> Oh, and the traffic cone thing, right?

Uh there were like random protests and I think in San Francisco. I like research for class and they'd put like a traffic cone on top of like a Whimo and it would like assume that there was, you know, something in front of it and it would

just stop and they'd have to like deploy, you know, your technician or engineer or whatever to like un to fix the problem. They just have to remove a

the problem. They just have to remove a traffic cone. So, you know, the 5-second

traffic cone. So, you know, the 5-second effort of you putting a traffic cone on a Whimo car cost Whimo like an excessive amount of money for no reason.

>> Yeah. And uh we also have a culture of anytime when there is a protest we burn away more. So what about Metaquest?

away more. So what about Metaquest?

>> So for the course right I think a lot of like big tech company also are reaching and this kind of product like like Apple

vision pro for the near thread. Yeah I

cannot get a really specific example about the near thread. uh some recent companies I think different countries have uh like different like companies

are re researching such like such products like I think in China big company like Xiaomi also are doing similar things substitute or long-term

potentials rights this kind of glasses also provide a way of entertainment so a lot of like uh PS5 or similar u game

devices can be their uh threats >> uh It's a startup which does like uh athlete uh device and they are doing a

device that can uh be kind of transparent and like let the athletes see what they are what about their heart rate and about like the uh the route

they're going to take uh in front of their like uh the bicycles or anything they are doing right now. So I'm

considering like if they can do right here they can actually a thing and in terms of this but I think it's not like as much competitive as Xiaomi like the

the thing they are going they they have published.

>> So for blind people who are blind there's a company there are very various set of companies >> wherein if you wear their specs it gives you voice input of what's happening in

your surrounding constantly. So can it be an indirect threat for Meta if these companies pivot to for everyone?

>> Yeah, I think they could pivot. Uh that

is a very similar thing with the vision pro, right? Apple mission pro that they

pro, right? Apple mission pro that they are kind of mimicking uh what you see outside and what you can hear. Um but it is a vision instead of like a real thing

in front of you. So yeah, I think they could pivot but uh I like as you what you said right now they're not targeting

>> meta ribbon smart glasses themselves can be a like long-term indirect competitors because um if they want to advance or

extend the current ribbon glasses function um they can expand there like some features to make the screen like se

semi transparent to show some information um like like the glasses which can be part of the function of this current metaquest. So yeah like

self compatible >> like an allin-one like 10 years from now MetaQuest and um you know um Raybands they have like the same tech and you could just use one for both.

>> I'm really curious how they would like segment the different user for these two products. I feel like right now Rayban

products. I feel like right now Rayban functionality is so simple. It doesn't

even like uh go to the area of um uh AR yet, right? So um definitely I think the

yet, right? So um definitely I think the other competitors will catch up because Rayband's functionality is simple but it popularized way faster than the other

devices. Um so that give you a pretty

devices. Um so that give you a pretty good user base for you to add additional functionality on top. Um but the other companies will suddenly be like oh

actually we can do a glass instead of a whole headset. So uh they have to like

whole headset. So uh they have to like what we discussed last time that they have to have distinguishable uh features on top of Rayban and that's how they can

be a more premium or just like make Rayban the commodity right um so yeah I think the there will be more players and more competitive in this game

>> when it comes to like glasses it also add a fashion layer so I know Google is partnering with um Gento master and

that's obviously like way more way cooler than Rayban uh for Jenzi. So it's

really interesting to see the competition go beyond tech. My friend

work at Meta Rayban uh as a product leader and when I was talking with her about this and she was like Meta know nothing about fashion but Rayban knows

how to do fashion part so we just build a software for them. And and I feel like when the market was saying that, oh, it's time for Google to bring back the Google glasses and Google was like, I

probably want to collaborate with a fashion brand to make me myself fashionable as well. Uh because back then when they first launched Google Glass, that is the right product but in a too early of a market. Also because

they have a tiny screen on top of their red eye. So every time when the user

red eye. So every time when the user look at their screen they were like so you can uh tell that they're looking at the screen which is a very awkward

experience. Um yeah so I think

experience. Um yeah so I think definitely they're going to improve um how the user interact with the glasses and also how you can do um AR better.

Spotify.

>> Uh, direct threats could be like Apple Music, YouTube Music, which is a thing.

There's another thing people use all the time. SoundCloud for adjacent players.

time. SoundCloud for adjacent players.

Uh, Meta can do it with its new Ray-B band glasses. They could have their own

band glasses. They could have their own song uh music system instead of spot relying on Spotify that's like native to

meta hardware. And indirect stuff would

meta hardware. And indirect stuff would be watching a video.

So I I I think the stuff the most adjacent players are kind of like like Yogesh said like social media companies >> like in terms of core threats of course

you're mainly talking about music here um there are like podcast and um also books right now so they have way more competitors um these days and I feel

like they are definitely screwing over Audible for sure. Um yeah, Audible is at a very awkward position right now. And

the other I think adjacent players um I think streaming companies will be interesting um because streaming company uh like it depends on whether they want to expand and a lot of them are uh

expanding. What we also discussed last

expanding. What we also discussed last time is um original doing original content they go to streaming doing streaming they go to original content.

So having IP for the artists is very important for these type of companies and they could have just like expand their vertical uh to go to music as well

for niche regional there's one we use uh in China called QQ music and you can hear from the name that it it comes from

um Tencent and uh there's some other companies like net ease um I think Q like the competition between these two companies is also very interesting uh

you can see how music uh softwares uh just buy uh the control of the of the artist. So uh back then QQ

music was very like a very little branch within Tencent and they bought the most famous artist in China J all their

records and they have the power to play however they like his music and that is the the decisive factor to win over net.

So it's very interesting industry and in terms of indirect yes when it comes to like entertainment or social media is also you're basically getting time from the consumer.

>> I I want to say interesting use case. So

in tier 2 and tier three India people cannot afford Spotify or YouTube music and I would really want to share this because this is a interesting use case.

telecom companies in order to have more app downloads in their app and let people spend more time on their app and use their tele uh uh service they

incorporated music in their apps. So

people in tier 2 tier three India use their telecom companies app the sim card company's app to hear music for free.

>> Wait is it like robust is it like a nice are they nice applications? They are not as nice but they offer free music and

I've seen people you so wink music is a telecom company uh brand of telecom company I tell I really love this use case I guess very smart way to get

people into our app >> GoPro >> I'm assuming there are a lot of direct competitors just because it's not that like they don't really have much of a moat but in terms of like uh near threat

and stuff we could talk about you know Rayban and you know other alternative maybe AR just like software and a glasses type

company. Um indirect

company. Um indirect uh cool ones. I know there are a bunch of like drone companies that are trying to do something similar which is neat and I kind of like that. I think it's really cool. You know like you have a

really cool. You know like you have a drone follow you around while you're doing a cool thing.

>> Yeah, I'm pretty amazed by the technology in innovation in the drone industry. Um, back in the days when I

industry. Um, back in the days when I was in college, that was, huh, almost 10 years ago, guys. Um, and it is so hard to navigate a drone. And the drone is

super big. I remember when I was

super big. I remember when I was shooting graduation video for um, my school. I need to carry that drone every

school. I need to carry that drone every day to climb up the mountain because my university is on the mountain. And

nowadays, it just like can fit in your pocket and it can follow you and then it can have a longer range for batteries.

It's yeah really crazy innovation there.

>> Have you ever seen a drone race? Those

are always fun.

>> Yeah, it is very interesting. Um and

also for the drone race, um my understanding is a lot of players they uh build drones themselves. So they like customize drones just for the race. Um

that is a very like niche u vertical of interest as well.

I want to add one company which I think could be in the middle of core thread and near thread that's Insta 360.

Has anybody heard of it? Yes, they do panoramic cameras and I think according to their financial report they are selling really

good overseas is established from China but it's core market is actually in overseas. Yes.

overseas. Yes.

>> Yeah. They just have their IPO, right?

Um where do they go public, you know?

>> Yes, exactly.

>> Yeah, I'm a overseas user. Um their

product is really amazing. Uh their new I forgot what what is the um Action 2 I think. Uh that is their latest camera.

think. Uh that is their latest camera.

It is magnetic. Uh their video quality is also very good. Um, I went to Austin a couple months ago and I took my GoPro

and Insta 360 um, just to make sure that I have battery uh, all the time uh, and storage and I barely even use my GoPro

and I could understand my GoPro uh, stock is so low uh, Dolingo. So for the course right uh like cam cam or Ialkie

and there's a lot of like English uh learning application in China. Um and

for the nearest rides I think there's a lot of like free free lessons or uh systematic lessons on uh like on YouTube people can learn from that.

>> More core threats would be like Rosetta Stone like Pimpsler. I don't know if that's a near threat or a core threat because they focus on like speaking versus I know Dualingo is more like

learn your keywords um you know more reading more understanding things rather than actually going and speaking. So pencil

is I think maybe a near threat but um you also talked about last time how Dualingo is just like learning stuff as well and so the indirect could be like con academy.

>> There's a also a AI company startup called speak. They're very interesting.

called speak. They're very interesting.

>> Oh I saw that on LinkedIn.

>> Yeah they're very interesting. Uh so

both of the founders are uh American and Harvard graduate I remember. uh but they started their market from Korea uh

because uh they realized in Korea um they spent 1% of GDP in English uh learning um so that's why they were thinking if we could dominate the market

that's pay the most attention in learning English uh we can expand to other markets as well so they are very high penetration in uh Korea Japan and

Taiwan and now they're going back to the United States they're like we are back uh and competing with Dolingo but uh nobody talk about Dolingo stock going

down when chat GPT GPD5 come out so are you guys AI native or not like nobody talk about uh those chat bots >> we are I'm I was aware about it but I

felt that they are now moving into the education part right and they are also pivoting >> so they will come back stronger

>> they will come back stronger It sound more like a propaganda than a logical strategy. Um I think some

logical strategy. Um I think some something I realized these days is the Dolingo CEO uh he is a a professor at

CMU and he usually just have very limited media uh media uh attention. Uh

he is a very like down to the ground kind of person. He does interviews but uh not like very like spammy marketing kind of person. But these days um he has a bunch of collaboration with content

creators which is interesting to promote Dolingo.

>> Uh and they just collabed with like one of the largest coffee chain in China.

Yeah. And speaking of AI, they actually um I think they have a lot of like direct competitors when um like large language models come out and there were

a lot of companies right now doing um like AI um lang AI language tutor stuff like that because well I haven't just entered that one. So

>> yeah and Dolingo has a very uh unique company culture u as well. It is not only their marketing has a very unique culture. Their company culture is also

culture. Their company culture is also interesting. Um they have a very strong

interesting. Um they have a very strong AB testing culture and there is also a theory um they realize subscription or

their user engagement uh increase with the level of the Dolingo's butt size increase. So which is also very funny.

increase. So which is also very funny.

The bigger the butt the better the subscription. And the last one anthropic

subscription. And the last one anthropic >> we have like the obvious like open AI Gemini or I guess Meta uh Google I don't know Gemini is yeah

Google uh adjacent players you can pivot into the market soon I guess like a bunch of the hardware like yeah a bunch of the hardware companies

can you know that everyone uses for compute could just suddenly decide to do LLM stuff. So like if Nvidia was like,

LLM stuff. So like if Nvidia was like, "Oh guys, guess what we're doing tomorrow. I guess they could enter that

tomorrow. I guess they could enter that market." Niches, I'm assuming there are

market." Niches, I'm assuming there are random startups in every single country that have better UX designed for their country. Um indirect everything that

country. Um indirect everything that chat kind of replaces substitute-wise.

So like Google translate, Google search >> instead of thinking forward, you're thinking backwards. How do you find

thinking backwards. How do you find substitute? Can uh Salesforce, Einstein

substitute? Can uh Salesforce, Einstein can be a near threat to anthropic or even Oracle the companies which are

coming up with their own chat bots.

Not chat bots means their own solution.

>> Why do you think they have the interest in that direction?

>> Because anthro is solving problems at every industries.

Of course, now they are doubling down to code helping developers more because they see an increase in the revenue.

Likewise Salesforce Einstein which I'm got an experience recently, they're also doing great stuff in LLM. So, I

guess >> I guess yeah, it depends on like how big the company is. I guess like if your company is big and you have a lot of priority data, then uh you can train the

model yourself. uh based on like a very

model yourself. uh based on like a very raw foundational model. Um so yeah, a lot of companies are doing that. Tik Tok

just launched their video um model I think this week. Um and there are also more niche like model startups like

midjourney, luma like all of those models are doing like models um so all of the companies are doing models. So um

yeah it's a lot of threats and also I think adjacent players can pivot into the market soon. A lot of general agents companies are doing their own models these days. Um and it is interesting

these days. Um and it is interesting that the whole pretty much the whole agent um like market agent companies are

using the MCP protocol that is designed by anthropic but like all these people are trying to eat their market.

>> Yogish talked about the code stuff.

There's also like cursor cuz you know cursor um uh co-pilot >> maybe also like you can say like lovable

and all those vibe coding stuffs as well.

>> Well now let's think bigger. Uh

previously we just focus on competitors right now let's uh think bigger on the whole value chain. So the competitor we're mostly talking about industry

revival arrivals and some of it uh when we talk about Jensent um like competitors that might be a threat of entrance right uh or the indirect competitors can be a threat of

substitutions. Um but if you look at the

substitutions. Um but if you look at the uh left and right side supply power and buyer power is extremely important uh when it comes to how could you build

your company mode. So first we talk about uh threat of entry. Uh new

entrance bring new capabilities and want to grow and market for example the customers often are uh receptive and if a new entrance is diversifying into a

market they can share cost with their current markets and threat of entry um equals upper bound on profit potential.

So to ex uh explain the three points the first one the customers are usually very open-minded to try new products. So when

there is a new entrance coming in they will try at least whether it will stay on the product or not depends right but they will at least try and can like test

it out um and stuff. And the second one um what do we mean by share cost with uh with the current market? We talked about Tesla last time that Tesla is not only

trying to build electric vehicle.

They're trying to share their IP to like the whole industry making cars electric, right? If right now there is only one

right? If right now there is only one supply chain for Tesla to um design the batteries, design whatever to build electric vehicles, the cost of the the

whole market is pretty high and it is all on Tesla. But if there are a bunch of companies doing so then they can share their cost and share the risk on their suppliers. So that is diversifying

their suppliers. So that is diversifying into your market. They can share the cost with uh your current market as well. And threat of entry equals upper

well. And threat of entry equals upper bound on profit potential. It's also

interesting because as long as there is more squarable margin that you are earning right now, there will be more competitors coming in and to share your uh margin and you you guys together will

make the profit lower and eventually this is how the market work will eventually benefit the consumer side by competition. But let's say if it is

competition. But let's say if it is already very tough market um there will be very little uh threat of entry um

first for example you barely hear that there is a local grocery store or whatever uh market is trying to have a even lower price than Walmart right because they already figure out the

supply chain they already figure out how they can put the price so low the same thing with IKEA right u barely do companies uh compete with IKEA on the

price of furniture um because the upper bound on profit is not as high.

So the counter to the threat is high entry barriers and what define high entry barriers there are multiple factors. Uh the economies of scale uh

factors. Uh the economies of scale uh proprietary uh product differences brand identity uh for the class that we have last week for like the Chinese audience

um all of the students there have weird um uh admiration towards Apple as a brand. Okay. So that's the brand

brand. Okay. So that's the brand identity there. No matter um how weird

identity there. No matter um how weird the product is from from Apple these years, they still keep their admiration to Apple as a brand. That's the brand uh identity and switching costs also

matter. uh capital requirements for

matter. uh capital requirements for example training like why we talk about deepseek right why this the launch of deepseek is such a threat to open AI's model because previously a whole market

thought that you need a lot of capital to train an AI model but now it seems like someone is challenging this hypothesis uh also access to distribution let's say you want to you

want to build a customized like cloud computing let's say you want to build on something on top of cloud computing right the current player for example um

aure from Microsoft or AWS from um Amazon have way better access to distribution um to the companies that they are serving cloud services to right

um so that is the barrier to entry as well um absolute cost advantages prior uh learning curve access to necessary input and prior low cost to product

design government policies this is also what Honda said uh when it comes to building mode for Open AI expected retaliation.

The main threat is however you can make new entries harder. Um and this either take a lot of time or take a lot of money. I would want everybody to see

money. I would want everybody to see brand uh identity as something that you need time to build. Um that's why we talk about as startup don't have brand identity but uh any other stuff either

need capital um or time. Substitute

performs the same or similar functions differently and non-conumption is always a substitute. So non-conumption is the

a substitute. So non-conumption is the consumers finally realize that huh I don't need to buy you at all that is also a substitute and high threat of

substitute substitution is also the upper bound on profit potential. So

again as long as you have a high percentage of margin then the other there will be other competitors coming in. This is the market effect to make

in. This is the market effect to make sure that there is no monopoly earning a lot of money from the consumers. Um this

is a healthy competition and the counter to threat is differentiation threats from the substitute offering and the first one relative price uh performance

of substitute. Uh second one is

of substitute. Uh second one is switching cost. Third one is buyer uh

switching cost. Third one is buyer uh propens propensity to substitute. So we

price is easier to understand uh if the users uh or like your target audience is very price sensitive uh especially and the second one switching cost is what uh you guys already mentioned previously if

you have a ecosystem right it is very hard for you to switch out and the third one is the buyers just love you the buyers uh love you or just love the substitute forget about Apple

intelligence but let's say Apple just have a new foundational model come out right we will still try it out we will still use it uh because Apple have the

reputation even if the AI is not as good as we expected we will still try it and see whether we will like it and now we're talking about the value chain uh

this is something that I um I think it's the like the most important thing of today's course the power the powerful suppliers can charge more or uh limit

quality or service or pass on cost to customers and customers are unable to raise their price and powerful suppliers

also huh upperbound on profit potential and the counter to uh the power is diversify uh supplier community. So

there are a lot of factors to define the suppliers power, differentiation of input, switching cost of suppliers and firms in the industry, presence of

substitute uh input, supplier concentration, importance of volume to supplier, cost relative to total purchases in the industry. Impact of

input on cost or differentiation. threat

of forward integration relative to the threat of backward integration by firms in the industry. You want to build a hardware product,

but right now you only have the design.

You only know what kind of product you want to build, right? And now you're thinking there are so many hardware, let's say smart home devices or whatever, right? In the US market

whatever, right? In the US market already. You want to make the cost

already. You want to make the cost lower. Then what do you do? You go to

lower. Then what do you do? You go to made in China, right? Why is made in China so cheap? Made in China is cheap originally I thought because the later

cost is low. Not only that, so let's say right now you have a product design and you have a product direction to go.

There are so many suppliers that you can choose from and they put the cost so low for you to choose. Let's say right now you have a lot of uh buyer power. Uh you

can your consumers buy your product, make a lot of consumers buy your product. You have a lot of power over

product. You have a lot of power over your buyers, right? And now you want to put pressure on your supplier.

Let's say I don't really want just cheap products, right? Normal made in China

products, right? Normal made in China cannot satisfy my need. I want to make sure that it is as cheap as possible so I can increase my margin. I also want to

make sure that you have the best quality of product design and execution and you can do it fast. I don't want to compete with other products or other competitors

in the industry. when I want you to make my product, you make my product, right?

Then your power can get you Foxcon, right? So they can provide a better

right? So they can provide a better service to you um by however bullying their own employees uh or however to make your requirements happen. That's

how they treat Apple. So the power is always dynamics and it depends on a lot of like your own factors, the industry uh and the like supplier buyer u

relationship. Can you explain the

relationship. Can you explain the importance of volume to supplier?

>> Yeah. So how much you buy from the supplier um and relative to how much it sell to the industry. So let's say you buy 90% of the supply from the supplier

then you have a lot of power to negotiate with the supplier. So let's

say you're building a electric vehicle and the battery is the major cost of your product, right?

So your suppliers on battery might have a bigger negotiation power over you because they know they're important. Uh

but if you're only looking for like glasses on your cars, they probably don't have enough negotiation power because there will be more like substitute and it is not as important in

to your industry uh than that supplier.

how hardware companies uh treat their supplier is pretty bullish and that's why if you are um interviewing let's say a hardware product manager a lot of like

a lot of skill set will emphasize on vendor management so uh because if let's say you're not Apple right like if you're Apple then yes you just like need to make sure that

they don't leak your secrets and whatever right u but let's say you are a midsize company or you are a startup Right. Uh your supplier might be it

Right. Uh your supplier might be it might have a lot of demand from multiple companies. It is really hard for you to

companies. It is really hard for you to do quality control. Um I used to work for a startup called Wise. Later we will talk about that startup as well. Um they

have this mission saying too good too too good to be true. So they put their cost really low uh for the American market. They have suppliers in China and

market. They have suppliers in China and I used to work on their smartwatch and pet tracker products. executive of this company has to frequently travel to the factory to make sure the quality control

and test out different products, right?

For software company, usually you don't have that much hustle um to do like quality control or vendor management, right? For Apple, they have a different

right? For Apple, they have a different issue. For Apple, there is um

issue. For Apple, there is um confidentiality problem. So for a lot of

confidentiality problem. So for a lot of people from Foscon you know that Apple working on different parts of the product don't know how the rest of the company is doing right they're very secretive kind of product they're building

>> so for example you are building the camera for iPhone for the next generation of iPhone you don't know how iPhone look >> right in general you just know how the

camera will be designed and I from Apple perspective there's a list about the people working at Apple and the people

from Fosscon. So if you reach a certain

from Fosscon. So if you reach a certain level at Fosscon, you will be on that list and Apple will track you to know whether you are job searching, whether

you're talking with our competitors, whether you are doing some weird stuff.

So yeah, vendor management is a very interesting thing.

>> Interesting product management as a vendor.

Yeah, a hardware product manager is very bullish like >> crazy opportunity.

>> What? H

>> that's a crazy pivot opportunity. We

could all become hardware PMs. >> Yeah, hardware is I think hardware is going to be so popular uh in the next few years. I think software industry is

few years. I think software industry is um is too saturated right now and uh hardware has a lot of opportunities. But

yeah, just to give you an overlook of how a hardware product manager will look like.

So you guys have be have to be tough to manage your vendors. Then we comes to the power of buyers. Buyers uh is your customer and um can force uh own prices

or demand better quality or more services or threaten to switch suppliers. And again the powerful pers

suppliers. And again the powerful pers uh equals the upper bound of probability. So you can see right if

probability. So you can see right if you're very profitable there will be a lot of threats in the industry. If

you're not very profitable you won't have a lot of like substitute new entries. People don't care about you

entries. People don't care about you that much. Your buyers will be like oh

that much. Your buyers will be like oh it's fine. You're already like very low

it's fine. You're already like very low price. Your supplier will be like okay

price. Your supplier will be like okay it's probably not a lot of margin for me to negotiate. So the upper bound of

to negotiate. So the upper bound of profitability is the key. And the

counter to uh the power is what we already talked about last time differentiation because that's how you can win your buyers um over the other competitors. Uh there's several u

competitors. Uh there's several u bargaining leverages. The first one

bargaining leverages. The first one buyer to concentration versus firm concentration, buyer volume, buyer switching cost relative to firm switching cost, buyer information,

ability to backwards and integrate and substitute products and proof through.

We could also talk about price sensitivity here uh the price or total purchase uh product differences uh brand identity impact on quality and uh

performance buyer profit and decision makers uh incentives. So some of the industries might be more price sensitive some of the industries might not. So uh

it is very uh industry based talking about bargaining leverage there is one point talking about buyer information it's very interesting. So for example, back then before Amazon.com time, right?

A lot of prices are not online. It is

extremely hard for the buyer to compare prices. Uh but now everything is online.

prices. Uh but now everything is online.

They got more information, right? So

then it just push all the companies trying to sell on Amazon.com to lower your price. For the price sensitivity,

your price. For the price sensitivity, there's one point that I want to point out which is decision makers incentive.

You have to dis you have to know for your product either it is enterprise product or a consumer product who is the decision maker in this scenario right if you're targeting consumers sometimes

your end user is not your decision maker if you're targeting their kids because usually the decision maker which is the econ economic buyer is their parents uh

when you are targeting enterprises it's extremely like that so the decision in enterprise cases a lot of times the decision maker is not even the as the economic buyer because the

economic buyer is usually the CFO kind of position. Decision maker is the CEO

of position. Decision maker is the CEO kind of position if it is a very big product let's say and the end user is whoever audiences that you're targeting.

So there are different people playing different roles in the buying decision.

And now talking about buying decision it is also interesting to uh talk about enterp the differences between enterprise uh products and consumer products. So for example uh if you're

products. So for example uh if you're doing an ad on Amazon, Amazon is uh able to tell you how effective the ads is in terms of how

many people buy your product because of this ad, right? But if you're doing B2B marketing, it is extremely hard to tell you so because the decision- making

timeline is extremely long. So if you're buying a consumer product, usually the time that you spend on decision- making is in terms of days or weeks depends on how expensive the product is. If you are

doing a enterprise product, usually the decision making time is from two years to five years. So it is extremely hard in B2B marketing, B2B sales to do

attribution. This is a very well-known

attribution. This is a very well-known industry challenge nowadays. If you use LinkedIn, the most active people on LinkedIn feed specifically are B2B marketers and B2B sales, right? I just

proposed to my fiance, this is what I learned about B2B sales. U this is a very well-known funny post on LinkedIn.

It's like a vibe from content creators on LinkedIn, right? Why it is funny? Not

because it is totally unrelated between the two things and they're trying to just make a post on LinkedIn. Also,

there are a lot of B2B sales on LinkedIn and um you can see their pain points very like regularly. Attribution is one of them. When there is a deal made, it

of them. When there is a deal made, it is very hard for you to tell whether it's a sales word of marketing.

>> I love those. Like my mom just died.

This is what it taught me about B2B sales.

>> I love those too. Uh there is a company I think they are a startup um they they call themselves B2B growth agency and I think they are a company based in

Europe. Um but I forgot which country uh

Europe. Um but I forgot which country uh which country is that um and they are they went viral probably you guys already saw their post on LinkedIn. They

went viral at first by making a video to mock um the walls of Wall Street.

Oh, I saw that where they like repeat the thing and they're talking about like Yeah. Yeah. Yeah. Yeah. I know what

Yeah. Yeah. Yeah. Yeah. I know what you're talking about. I know what you're talking about. Do they also do um I I

talking about. Do they also do um I I think other people are like replicating what they did or I don't know if it was the same people but they're doing like other movies.

>> Yeah. Um they make more movies after that. That initial success is huge. Um

that. That initial success is huge. Um

and I v with that CEO so much because I found that dude is just so funny. Um,

and I just comment on some of his videos and he added me on LinkedIn and now I just like his content on all the contents I have on my feed. I don't know how much you can relate to their videos.

Their videos are definitely funny, but what they said in the B2B marketing industry is very uh on point. Uh, so for example, the Wolf of Wall Street um

they're basically comparing how the old B2B marketers and the new B2B marketers are doing, right? So the old B2B marketers are telling their interns to send two posts per day on LinkedIn,

right? One after breakfast and one after

right? One after breakfast and one after lunch. That is the traditional way of

lunch. That is the traditional way of doing this because they just want less.

They don't care about the quality of leads. Um but then the new types of

leads. Um but then the new types of YouTube marketing care about the quality of leads because they want to be the opinion leaders uh in their domain and

that's how you can have quality uh higher quality leads uh for the further conversion. Men used to be on war, now

conversion. Men used to be on war, now they're doing B2B sales. It is very brutal out there. And do you know how many startups or midsize companies out there in the B2B marketing space? B2B

marketing is not even as huge or as saturated as B2C market. But B2B market also have thousands of midsize companies. They do like a lot of random

companies. They do like a lot of random things. Um the whole within the whole

things. Um the whole within the whole pipeline of B2B marketing, there is always like every step there are like tens and the hundreds of startups. As

long as you have certain relationship on yourself then you have a market. Let's

just use privacy uh in privacy domain as an example right we just talk about for um the audience what companies call it is audience network Facebook called it Facebook audience network LinkedIn call

it LinkedIn audience network you basically put your ad advertisement on the third parties right for this space because there are more regulations coming out there will be companies just

saying just thinking huh the whole point of at DMA is to prevent companies like

Meta or LinkedIn to say this person visiting New York Times right now is this person on my LinkedIn right the whole point of this regulation is we

cannot directly use let's say device ID or whatever to do direct match because from DMA perspective from EU perspective these are two businesses um within a

company so we cannot utilize the identity system together, right? And now

there's a start of thinking, what if there's a sandbox, right? So I can give you certain uh demographics or certain personas of this person, but I don't

necessarily need to tell you who that person is in your database. So that is a valid business. There are tens of

valid business. There are tens of startups doing that. Consumer is like if you have don't have enough users in a short time period, you're dead. That's

basically it. So it's very hard. It's

very easy for you to decide oh whether I have a product market fit or not but B2B product you can work on it for years and you can probably just serving one customer.

>> Yeah. Because it's like every step of running a company like whenever I do anything and there's like a minor inconvenience. I was like hm there's

inconvenience. I was like hm there's probably like 10 companies that work on said minor convenience try to make that company's life easier.

>> Yeah. So that's why a lot of B2B product uh companies uh either they do um they're very big reason being they have

a lot of buyers and they can have more power over their buyers or they just serve specific buyers and one of their

exit is getting acquired.

Also, I feel like B2B customers are more like willing to help because my um whenever I have conversations with my um product owner or whatever now um he talks about uh

they used to have a thing called I forgot box used to have another product but when they used to be B2B um other businesses were more willing to like pilot stuff out. they were more willing

to like answer surveys and you know be like test dummies versus you know your random customer who pays $10 for your thing is not going to want to do the survey you want them to.

>> Yeah, consumers are very hard to help you out on this. Um but also in B2B products the customers know that their opinions matter. So another thing is if

opinions matter. So another thing is if let's say I'm working on LinkedIn consumer right I conduct the interview with a 100 users maybe out of the 100

users let's say 20 of the users tell me the same thing but whether I'm going to solve that problem to the 20 users not necessarily right because you guys might be uh just representing a very small

percentage of the total population there uh but if I'm a B2B product manager and there's one customer telling me to do is I better deliver on time especially that

is a big customer. Um so also for a lot of B2B products um there are different layers of customers right there are strategic accounts which are super big.

They are the uh companies that pay you millions of dollars to make your business live. So usually whatever they

business live. So usually whatever they tell you to do just do it and do it fast. Okay? Do it with high quality.

fast. Okay? Do it with high quality.

Whatever they want just do it. But there

will also be midsize companies um they are maybe paying you let's say lower like one two million or even lower um

like uh uh deals right maybe more than 500k per year but not enough to make that to make a that difference in your decision- making. So you probably want

decision- making. So you probably want to conduct a bunch of surveys do a bunch of interviews and to get the general insights of what they want right but you also need to make them happy. One job

and we just talked about hardware product manager. One job for B2B product

product manager. One job for B2B product market managers is to make your client happy. So no matter you are doing this

happy. So no matter you are doing this requirement now or not, you have to make them happy. You have to tell them that

them happy. You have to tell them that it is coming soon. Let me give you a timeline. We're working on this. And for

timeline. We're working on this. And for

a bunch of small startups, you probably don't care about their opinions that much because they're not even giving you 100k per year. Um but also this number is relative for different companies.

right now. I'm probably talking about a midsize company, but for a startup, 100K is pretty good. Um, not bad as your first customer. So, if they're paying

first customer. So, if they're paying you 100K, you better serve them well. I found a lot of like their company website has a road map page where um they just clearly

show like what feature they're going to uh launch in the following quarters or like years and also they have like the customer very comprehensive customer

forum that um like they encourage their own user to discuss with them like with each other and they also have like a

feature requested um like dashboard. So

their customer can like vote or like express opinion on like some most wanted features. So that it also like greatly

features. So that it also like greatly help the product manager to prioritize across a lot of like requested features.

>> If you have done sales in B2B that is a pretty good um thing to put on your resume to let you interview whatever ro you want to have. Um if

you're let's say there are two candidates. If you're a consumer product

candidates. If you're a consumer product manager like me or you are someone who have done B2B sales now you want to go into B2B product management space definitely the B2B sales have way more

advantage than a consumer product manager.

Yeah, my friends told me like um he's doing like B2B SAS product. It's almost

like sales because you have to communicate with like many clients. But

um I just have a question like um it might be a little bit hard to prioritize like different um stakeholders needs.

Sometimes one client said they want this kind of features, but you also want to catch up with what competitors have um on their platforms. So I feel like this

kind of decisions is very hard.

>> Yeah, the prioritization is extremely hard. Uh you have to have some criterias

hard. Uh you have to have some criterias uh for you to prioritize. Um so at some levels uh your goal is similar with uh your major clients, right? Whatever your

major clients want, you want to keep the deal as well. Um so that's why uh how much money they pay you matters a lot.

Um and how long you have been working with them also matters a lot. Um and

also there are a lot of innovative way for you to solve their problem. Uh not

necessarily to u build whatever product that they want. Uh probably some other designs or some other implementations also can solve their problems. So you always can have um smarter way to to

solve a problem. Apart from the original Porter five forces um these are very external to this industry. uh the

technology and the pace of change not to me like I don't need to uh mention to you guys the AI industry is chaotic it is brutal the technology and pace of

change is huge everyone no matter what companies that they're working for either big tech companies working on AI product or AI startups or rapper companies or model companies they will

tell you that this industry is chaotic um it is very hard for you to keep pay keep track of everything happening Second one is globalization. We also

back to our open AAI case. we keep

talking about regional um regional models right uh either open eye is not even thinking about the go growing internationally to China Chinese market

right they're not even targeting Chinese market in any sense u because the global um uh geopolitics is very interesting between the two uh countries government

as well and that's why when GPT5 came out you could see a lot of you know American uh CEOs are saying like let's

support the American model right this is do they want to say this maybe but does open AAI want them to say this

definitely right so this is a very good marketing um for you to associate open AAI as your identity as American your pride of being American

and lastly social and cultural shift DJI does anyone know or not know what DJI does DJI is the drone company. Uh so

they produce most of the drones specifically for video recording. Um so

if you see an influencer um using a drone to shoot a vlog, that is most likely 90% of chance DJI. Um if you're seeing uh this is something I found

really crazy. Originally I thought DJI

really crazy. Originally I thought DJI is banned in the United States um until I went to New York City

and I saw so many influencers just using DJI products. either it is the pocket uh

DJI products. either it is the pocket uh camera or the drone just flying like on the Fifth Avenue. It's crazy. The

penetration is crazy. So, this is the uh DJI company.

>> Well, I guess the biggest one would probably be like government and red because a lot of certain um cities and stuff have banned drones or countries. I

know a lot of countries in the Middle East ban drones like to even start with like entirely. So big red one because if

like entirely. So big red one because if you can't even operate in the country, you can't really make money.

>> And I found it very interesting that all the national um parts in the United States ban drone like any type of drone.

>> Yeah. Wait, I also find that interesting because I feel like that give them like more coverage or like media coverage.

>> Yeah, I guess they don't need that much media coverage, but >> maybe it like messes with the ecosystem and like harms birds or something.

>> Yeah. and like or if they break away and they fall into a tree and like a bear eats it and then it dies. I just made that up, but like you get my point.

>> Yeah, that might be true. Um, for

example, you don't want to just toss the battery of the drone in some natural places um when you're trying to protect that place. Another thing I think is uh

that place. Another thing I think is uh they're trying to limit um let's most of the drones in this market is DJI. Um, so

they're trying to limit their geographic information to a Chinese company. That

might also be true. Um, because it is very hard for you to limit what type of brand they're using um when you carry a drone in that uh place. Third one is they probably do not need more

reputation on the national parks, but they but um the drones uh video images might uh hurt their uh tickets. So that

could be one. Oh, like the instead of watching it online, come and actually buy a ticket to visit XYZ.

>> Yeah.

>> No, that makes sense. Also, like when you think about the environmental stuff, like yes, you can allow drones in, but what one drone can fall down to use batteries that are like highly flammable

like a tree goes up and fire and then like you know before you know it, big catastrophe.

>> But I guess it's more about intoxication of the environment because of leakage of batteries. That is a bigger problem,

batteries. That is a bigger problem, right?

>> I'm not sure about that. Um I just found that problem really interesting. Uh data

is also something concerning. Also, I

heard some of the uh national park has government facilities or military facilities nearby. So it is also very

facilities nearby. So it is also very hard to tell you, oh this is the boundary. You cannot fly there, but you

boundary. You cannot fly there, but you can fly here. So uh yeah, I think they have very supportive reasons to ban drones. Um, but it's just interesting

drones. Um, but it's just interesting because I I don't think I am the only tourist thinking about using drone shot there.

>> Otherwise, I don't think supplier power.

I think has a low power to limit profits just because a lot of industries in, you know, the world suppliers can probably make the drones necessary that DJI has.

So, they kind of have like a lot of options right?

technology wise, I don't think it's like crazy innovative or anything. Especially

considering like if you go to like the Russia Ukraine war, people are like making these things with like stuff they find in their backyard. because I

watched a random documentary by this um Russian drone maker and he still says that even though Russia is like pouring millions of dollars or like billions of dollars into its drone industry and

Ukraine is still like getting a bunch of aid in terms of actual um drones from the US and everything, the best drones are still not Chinese drones. They're

not American drones. They're not Iranian drones. They're not Russian drones.

drones. They're not Russian drones.

They're Ukrainian drones that people are literally just diying.

So I think supplier power is kind of low here.

>> Interesting. I think also serve different purposes.

>> Oh yeah, they don't need to like kamicazi and like blow up 10 people, right? But like

right? But like >> but yeah, that is a I I think I also watched um a YouTube video saying that as well. I think the the buyer power is

as well. I think the the buyer power is much lower uh to limit the profit than the supplier power like for because I

think DJI has really high uh brand stress. So like as an influencer or

stress. So like as an influencer or people try to be a vlogger, they prefer to uh buy DJI as a really high quality

and uh high quality product.

um and to follow others to try different things and and for supply power uh I have a little bit disagreement with

Hamzi because um since DJI like when they develop their products they may need like high quality or specific like chips or other technology in their

product. So they have uh higher

product. So they have uh higher standards for their supplier power. So

in this case the suppliers can uh like ask for higher price to produce some basic segment some like important

segments of this product.

Yeah, I do think DJI has uh power over their buyer and their suppliers just because there are so many buyers out there. Um and the volume can give them u

there. Um and the volume can give them u negoti negotiation power to um towards their supplier and buyer which is the

same thing uh as Apple to Foxcon. I do

think it is very easy to make a drone. I

just think it is not easy to make as good of a drone as the DJI the drone to vlog. My boyfriend even built a drone

vlog. My boyfriend even built a drone himself. Like it is not like hard to

himself. Like it is not like hard to entry. Um but you can't make one that is

entry. Um but you can't make one that is like very silent, very smooth, it can shoot you good videos, high resolution, all of that. That is hard. I think

globalization is like curious because I think the pricing for DJI in the states is much higher than it was in China cuz

uh one of my friend has bought an act a3 here and it's like it's a not so professional one but it was like nearly the same price as a more professional

one in China and I think that's because DJI has like market dominance power so that they can actually make the price and nobody can say anything about it.

And also I found out like DJI has actually building the ecosystem because they are they recently have published a machine to sweep your floor. How do you

say that? and it's transparent and a lot

say that? and it's transparent and a lot of my friends who have already owned the drone was like very attracted to it cuz they think that it's kind of technology

and with like their taste of the like the kind of the new things that they should be put in their home. So that I think it's also developing a ecosystem just like app.

>> Yeah, I think they're they're launching a new uh robot vacuum, right? So this is a customer service a AI agent and it started pretty early after GPT time. Um

and it has a star found uh superstar founder Brett Taylor. Uh he is a legend in Silicon Valley. So yeah, he is um

founder of Sierra. They're basically um automate your customer service experience. So previously you need to

experience. So previously you need to talk um you either just like do maintain choices uh to tell the machine what you want or you find a representative,

right? But uh right now they can use a

right? But uh right now they can use a very natural language way to solve the problem for you. Um and they're very integrated with whatever CRM that you're using. I sound like I'm making an ass

using. I sound like I'm making an ass for them. But um to make this experience

for them. But um to make this experience good for their customers. So this is what it does.

>> Isn't this how like one company lost so much money cuz like they had to uh they had a certain um LM as an agent and then they made a deal with someone. They were

like oh I want free XYZ thing.

>> Oh yeah I saw that screenshot >> actually technology and pace of change will be the higher force to limit profit because anyone will come up with the

tech because tech cannot be the mo here.

People are working day and night to replace them.

Rate of entrance is low because entry barrier is low. Globalization

will kill it because someone in China can come up with a better price parity to solve the same problem. Buyer power

can be a thing.

>> I'd say it's huge. High power limit profits because one, I feel like this is kind of sketchy. Like if I was a company, first off, this like only source of income, right? It's basically

all just B2B. And also companies are I guess don't want really bad customer service. is kind of important.

customer service. is kind of important.

Like that's one uh a lot of that's a big reason why some startups like get a lot of brand loyalty like oh I called them and they instantly replied and they fixed my problem as soon as possible

right and so they want something they could cheaply replace that human experience and if they can't get it they're not going to use it. So

firepower is huge >> and government risk is also huge because in some places they have come regulations that you cannot use

customers data for XYZ these cannot let them but I but I guess as government or even countries start adopting AI it might not be a problem in the longer

run. There could also be like a social

run. There could also be like a social shift. Like if you ever see like my mom

shift. Like if you ever see like my mom on the phone talking to like one of those robot agents, she constantly going customer representative, customer representative. And it could be like the

representative. And it could be like the same thing like, oh, did they give you AI agent? Just ask for the customer

AI agent? Just ask for the customer representative 10 times, right?

>> But it is possible that you can't even tell.

>> Yeah. Not now, but maybe in the future.

>> Okay. So you guys can uh stalk him on on LinkedIn and and so I probably have his whole LinkedIn profile uh in my head. He

started his career as associate product manager at Google and he started the >> Google maps and then he went out of

Google and created a company himself and then the company got acquired by Facebook. So he then became the CTO of

Facebook. So he then became the CTO of Facebook and later he went out again and founded another company I remember and then got acquired by Salesforce and then

he was the co-CEO of Salesforce and right now he served as the board member at OpenAI and also does this startup on the side. You probably know that he

the side. You probably know that he might have some relationship with uh companies potential buyers and government. What about the company

government. What about the company Harvey?

>> Industry of law is made in that way that don't understand it very well. The judiciary part because if people are able to understand it

better, lawyers would not have good jobs. It's like taxes. It is bound to

jobs. It's like taxes. It is bound to have be more and more complex.

>> Lawyers are like the perfect target market because there's like the biggest chat GBT users or like any LLM users just because they read so much documentation. they have to work on like

documentation. they have to work on like you know cases that happened in the past you know to think about precedents to frame like their arguments or like the way they go about court and even if you

just work in law it's just so much reading and an analysis and everything so it's like perfect and so Harvey was like wait a second there's this ginormous industry and they all use it

already so why don't we specialize it and so there you go >> yeah I do think legal and u like healthcare this type of industry is

extremely hard to entry. Um and these are also very traditional and low pace industry. When they adopt one product,

industry. When they adopt one product, it is very hard for them to make another decision to switch to another product.

So yeah, there are a lot of opportunities in uh for AI products in these industries.

>> In this kind of like AI agent with expertise, it's just like how programmers do with VIP coding. uh most

people would agree with that though AI do help a lot on the specific correction or the or or big like uh big programs we

still need like expertise programmers to fix all the problems caused by AI. So I

was just wondering if a real professional lawyer would think that this uh this tool is helpful. H I think

it depends on what law they are practicing. Um for some of the laws

practicing. Um for some of the laws there might be more nuances um and for some of them there might be more common sense and common cases. Um I think it is the same with coding. Uh if there is a

lot of information on the internet for you to train AI then your capability in that area might be higher than some of the niche areas. So, um, yeah, >> I feel like Harvey set out to like fix a

lot of those problems or mitigate them because before Harvey was like a big thing or bigger than it was now. Like

you had those situations. I'm pretty

sure there was a um there was a lawyer in this big case. I study it in one of my classes and it was really funny. We

were talking about the effects of AI on society and um he basically used a case right um in in court that didn't exist

because his AI like hallucinated and he was just like oh based off this case I just cited where this and this happened right this provides precedent that XYZ blah blah blah whatever he was arguing

and basically the entire case didn't exist whatsoever it was all hallucinated and then he like lost his job so Harvey set out to like stop that thing from happening, right? And to stop being like

happening, right? And to stop being like a boogeyman in the law space and actually just become like a useful tool.

So, they probably have um like they probably train their du they probably train stuff on actual like documents because there are a lot of um big

databases that lawyers use and they pay like huge amounts of money to access all these documents and court cases and everything.

Yeah, they definitely I I do think they are trained on real cases. Um it's just even real cases um some of the areas might be harder for the lawyer uh like

as a job than some of the other areas.

Um for example, for uh you guys as Americans probably don't know that much about this, but for immigration lawyers, for example, uh the

job for them is relatively easy for majority of us. Um and it's just some of the cases might be more nuanced and but immigration lawyer is very well known

for being paid uh not bad and do nothing >> technology and pace of change I guess as they are first incubant and they are

with the big laws and bigsiz firms >> they are going to uh have a advantage because there is something called as profile building of AI in your

enterprise and as a user. The very

welfare reason chat GP has more users because we have built our AI profile in that they are going to have a technology and pace of chain advantage for a brief period of time but eventually if someone

comes along they will have a threat threat of entrance yes there is a threat of entrance because new people can enter into this industry because it's a data

industry and globalization I doubt because two reasons if I'm an Indian company

building a a replica of Harvey. I may

not know what are the local nuances because uh you can help me find the right citation or the law which can apply to

that incident but I don't know what are the other nuances which come along.

So but that said uh regional competition would happen supply power

uh I'm not sure it's very early for this

thing but yes if uh my model is trained on a company which has a very very good pipeline of data then it will create a

problem to me because uh the access of the trainable data will create will make me go behind in the

race of making a great AI agent industry rivalry. Yes, I guess it is going to

rivalry. Yes, I guess it is going to happen soon and buyer power. Yes, because if I have a good data and I if I can because

lawyers are paid on number of hours, it's like a grazing industry. More the

hours, more you paid. So if I'm able to save 40 hours, 50 hours it going to create a more and more compounding return because I can spend more time on investigation and other areas means if I

spend less on lawyer part I can spend more time on the investigation part where which is where usually people have a problem in forensic and everywhere if

I'm talking in terms of government sense not in terms of us as a customer and government I'm not sure if if government

would allow but as everyone is adopting AI I don't think so it it I guess it have a lower power to limit profits threat of substitutions social culture I don't think so because everyone is

adapting it >> our analysis is based on data so uh if you're doing >> no no no I'm just saying like uh we have

to like guess a lot uh on this class but if you're doing your report uh you usually um have your evidence to support Um so yeah in terms of like social and

cultural shift um it's very hard to predict how much um users are going to adopt AI in the legal um sector but I

will be curious if I'm doing this like as a full report or investigation I'll be curious what is adoption right now and how is their user retention um to start with to think about whether I have

early evidence to prove that users have the potential to adopting this technology in this industry.

Um, but I do have some like conversations with lawyer these days about their adoption on AI because I'm curious. Um, and a lot of them are not

curious. Um, and a lot of them are not really thinking about a AI legal agent.

A lot of them are thinking about whether uh Apple could have a better Apple intelligence to help them organize their documents uh on Mac for example. And so

in that way like they are thinking it is already pretty helpful. I think right now a lot of companies like OpenAI have this agent system uh if they're on Google drive, right? They can uh access

us and help them with investigation.

We're like reading a lot of documents uh retrieval information. Um so yeah, I

retrieval information. Um so yeah, I think lawyers will adopt AI. It's just

like whether this use case as a legal AI agent will work for them eventually. We

do guys >> co >> they don't only do co vaccines they do a lot of stuff guys >> I know at forest talent

I would say a threat of entrance is pretty low the pharmaceutical industry is pretty fixed right now and kids by

entering the industry >> well you either do um you either like do imitate drugs or you either like say um research and develop your own drugs and

I don't think in terms of like researching and developing like um I don't think there will be a lot of

potential competitors who will be capable to enter because it's time consuming and resource consuming and

money consuming so yeah so I would say threat of um entrance and threat of substitutions will be low to moderate.

Depends on if you would include um companies who just like imitate after your drugs. I think buyer power is

your drugs. I think buyer power is relatively low because um say some drugs um only say fizer or um a few

pharmaceutical companies you're providing and if you get sick then you have to use certain drugs then you don't really have a choice. So yeah I would

say buyer power is low too. supplier

power. Um I don't know a lot about like um chemicals or um molecules or um pharmaceutical

suppliers. So I guess I'll skip and um

suppliers. So I guess I'll skip and um technology and pace of change I would say is moderate to high. It's um again it's time consuming and and it often

takes years and billions of dollars to just like develop a drug and globalization um I think globalization will be moderate

to high too because um in some count like say the first is regulations whether um if you want to enter a certain country you have to go through

all the like approvals every single time and Um and some countries say like the rights are not um the intellectual properties are not protected and um

government I would say moderate to high too because they can literally determine if like the drug you spent like several years and billions of money developing

would actually be able to go to the market. Yeah. and social and cultural

market. Yeah. and social and cultural shift. I would think it's slow because

shift. I would think it's slow because um I mean even there are people who are like against drugs or against vaccines

and stuff like that but like most people or like the majority of people if they get sick when they either like take a drug they they just have to >> something which they are very good at is

government and social and cultural shift is creating a lot of issue created a lot of issue during COVID because people

realized what are the disadvantages of it but this is a it's this is companies a one of the biggest case

study from last five years I guess >> they are a very um technologydriven company um previously I went to uh John Hopkins uh university for a lecture

>> and there's another speaker who is the head of AI advisor uh 2024 early 2024 we are talking about April 2024 at that time she already told me that it is

proven that AI could uh fasten their drug development process drug discovery process and that is very early for healthcare like industry to move that

fast um and she doesn't have any question about whether they should adopt AI she just have question about how fast they can adopt AI yeah I think they move pretty fast and they're very

technologydriven >> if anyone wants to build a career in AI I is biotech and companies like Piser are good to place.

Why? Because two things they adopt AI far they will adopt AI faster than any other industries because data was always

always stored in a very best way.

Second, they are slow moving because they are very very related to the heart tech part.

And third, as humans are evolving, the rate at which virus are evolving is faster. But due to AI, the antibiotics

faster. But due to AI, the antibiotics can also evolate at faster rate. So

there will always be the race with the biology and more and more diseases will come and fiser will always have a problem to solve.

>> Yeah, agree. I previously read read a book uh about VC and the book is called VC if anyone is interested. Uh I think the the full name of the book is called

VC and American history something like that. Uh so to tell you guys about like

that. Uh so to tell you guys about like VC industry and how it started in the United States and how the VC in the United States specifically are pretty

like forward thinking than the rest of the world. Um and uh I just know from

the world. Um and uh I just know from that book so originally like VC started from like well industry right because when you uh haunt a well it the risk is

very high but the return on investment is also very high. So it's a very experience-based investment. So you see

experience-based investment. So you see what type of people can haunt a well and you give money to that person. Um but

later later on um biotech is actually one of the waves in the VC world to lead the technologydriven uh capital. So and

Jenn is one of the companies that I think made the name either Greylock or KP I forgot which VC firm but because of the return on investment on Genante um

investing in biotech uh or like investing in technology as a mode type of company is becoming more and more popular which is a very interesting

industry knowledge wise camera now I'm talking about my previous internship company so some information about wise right they um push the uh their margin

really low. So they sell a bunch of

really low. So they sell a bunch of smart home devices. Uh I think they're still hiring. So feel free to reach out

still hiring. So feel free to reach out to their people. The most famous product from wise is their camera. So I didn't realize this before I joined the

company, but their camera penetration in the US restaurants were like public places are insane. And I think two years ago, their camera just got into Costco.

Huh. So right now you can buy their camera from Costco.

>> Technology and pace of change. I think

technology involves new camera and new like more uh high fidelity that that kind of like shooting technology

evolves. Maybe this kind of really cheap

evolves. Maybe this kind of really cheap ones cannot afford uh what we need for uh this kind of monitoring camera. So

maybe uh people will turn to more technology advanced ones that is still relatively price competitive. Okay. And

as for globalization, I think I think wise is only popular among the US, right? So it just only builds the local market. So outside of

US, there are lots of competitors there.

uh I think especially in China there are so many cheap manufacturers who can provide almost the same product but at

even a lower price it's just because the like uh global barrier that cannot go into the US market supplier power I think considering the higher and higher

labor cost in the US maybe in the future it cannot maintain that low a price and for buyer power since wise mostly do

this kind of instore uh inspectation. I think most of them

uh inspectation. I think most of them are to for 2B. I think maybe in the future it can consider to like sell more

to to see uh to install this camera to each individual's house. I think it's very popular now with this kind of uh pass camera who can monitor how your

pass is going on while the owner is away. Yes. For government issue, I think

away. Yes. For government issue, I think the most concerning parts is the privacy issue like for threats of substitutions.

I think there are lots of lots of this substitution for this kind of camera out there. And the assignment this week

there. And the assignment this week since we talk about competitors all the time for this week um the topic is what is whimso's competitive comp competitive landscape and what should women do to be

more profitable and I want uh all of y'all not to only think about the framework that we learned this week but also think about the framework that we learned last week about what type of distinguishing uh distinguishable

feature that you want to build on top of to the current remote experience. Um I

think for OpenAI their portfolio is very broad. So it is very hard for you to go

broad. So it is very hard for you to go outside of what they already have investment in to talk about new strategy directions. But I think for Mayo's

directions. But I think for Mayo's current experience is relatively pretty basic. So you could still say that what

basic. So you could still say that what they can do to distinguish themselves and come back to reputation. I have a second assignment for y'all to download

a AI coding uh product uh product and start testing it um by writing simple websites. I will see you guys two weeks

websites. I will see you guys two weeks later.

>> See you JJ. Thank you.

>> Thank you.