AI business ideas funded by YCombinator

Summary

In this video, the speaker discusses how to determine what to build with AI, emphasizing the importance of natural aptitude, deep interest, and the potential for great success. Drawing insights from Y Combinator's portfolio, they analyze emerging trends in AI startups, focusing on industries like healthcare, fintech, and education. The video highlights the dominance of B2B solutions, the significance of technical expertise, and the untapped potential in areas like AI safety, data privacy, and blockchain integration. The speaker encourages leveraging generative AI innovatively and addressing ethical concerns to stand out in the competitive AI landscape.

Takeaways

• 😀 Paul Graham suggests that successful work with AI should align with natural aptitude, deep interest, and the potential for great work.
• 🔍 The speaker recommends observing successful AI companies backed by Y Combinator to identify emerging trends and technologies in AI.
• 📈 Healthcare and biotech are leading industries for AI startups, with 45 companies in the dataset, indicating a strong adoption of AI in these sectors.
• 🏢 A significant majority of AI startups (81.1%) are focused on B2B solutions, reflecting investor confidence in this sector.
• 🛠️ Infrastructure-focused AI startups are less common, with only 14.9% of the companies operating at this layer, suggesting higher barriers to entry.
• 🤖 Automation is the primary use case for AI across industries, with 69.1% of companies aiming to assist human work, while 31% are developing fully AI-driven solutions.
• 📊 Generative AI is the most popular technology among startups, with 78 companies leveraging it, likely due to advancements in large language models.
• 🔒 Addressing data privacy and security is an emerging area with potential, as only a small fraction of startups are currently focusing on these issues.
• 💡 Startups working on ethical AI or AI safety are rare, indicating a significant opportunity for innovation in these critical areas.
• 🌐 The integration of AI with emerging technologies like blockchain and quantum computing is still in its infancy, with only a handful of companies exploring these intersections.
• 🎓 The typical AI startup founder has strong technical expertise, often in computer science or related fields, and many have backgrounds in leading tech companies or academic research.

Q & A

What are the three qualities Paul Graham suggests a project should have according to the video?

Paul Graham suggests that a project should have three qualities: 1) It should be something you have a natural aptitude for, 2) It must be something you have a deep interest in, and 3) It should offer scope to do great work.

What is the significance of looking at Y Combinator's portfolio to identify AI trends?

Y Combinator's selection process has consistently surfaced companies that go on to reshape entire sectors, making their portfolio a valuable indicator of emerging trends and technologies in AI.

What does the video suggest about the current state of AI in terms of ideas and noise?

The video suggests that AI is currently filled with noise due to hype and many bad ideas. It emphasizes the importance of finding good ideas amidst this noise.

How many AI companies were analyzed from the last four Y Combinator seasons according to the video?

The video analyzed 417 AI companies from the last four Y Combinator seasons (Summer 24, Winter 24, Summer 23, Winter 23).

What are the top two industries where AI startups are making significant contributions according to the analysis?

Healthcare and biotech lead with 45 companies, followed by fintech with 38 companies, indicating these are the top two industries where AI startups are making significant contributions.

What percentage of the analyzed AI companies are operating in the B2B sector?

81.1% of the analyzed AI companies are operating in the B2B sector, showcasing a strong confidence in B2B from investors.

What is the primary use case of AI across all industries according to the data from Y Combinator?

Automation is the primary use case of AI across all industries, with 69.1% of companies building assistance to help humans deliver faster, and 31% building entirely AI-driven automations.

Which industries are identified as lagging in AI adoption and presenting opportunities for first movers?

Manufacturing, agriculture, energy, and retail are identified as industries lagging in AI adoption and presenting opportunities for first movers.

What percentage of the analyzed AI startups are leveraging generative AI technologies?

78 companies, which is a significant portion of the analyzed AI startups, are using or building something using generative AI.

What advice does the video give regarding the focus areas for someone looking to build with AI?

The video advises focusing on the B2B sector, underserved industries like manufacturing and retail, leveraging generative AI innovatively, and addressing ethical concerns such as data privacy and AI safety.

What is the typical background of Y Combinator-backed AI startup founders according to the video?

Most Y Combinator-backed AI startup founders have strong technical expertise in fields like computer science, software engineering, machine learning, or data science. Many have educational backgrounds from prestigious institutes or prior work experience at leading tech companies, and a significant portion have co-founding teams.

Outlines

00:00

🤖 AI Project Selection Criteria

The video script begins with a discussion on the criteria for choosing an AI project, referencing Paul Graham's advice that the work should align with one's natural aptitude, deep interest, and offer the potential for significant impact. The speaker shares their background in tech and AI, emphasizing the importance of identifying good ideas in the noisy AI landscape. The speaker decides to analyze Y Combinator's portfolio to identify emerging trends and technologies in AI startups, focusing on the hottest industries and sectors, and the skills and backgrounds of successful founders.

05:02

📊 AI Industry Analysis from Y Combinator's Portfolio

This paragraph delves into the analysis of AI startups backed by Y Combinator, highlighting the most popular industries and sectors adopting AI, such as healthcare and biotech, followed by fintech and developer tools. The speaker discusses the predominance of B2B over B2C in the AI sector, the focus on application layer over infrastructure due to the latter's complexity, and the common use case of AI for automation and assistance. The analysis also touches on industries lagging in AI adoption, presenting opportunities for innovation, and the variety of AI technologies leveraged by startups, with generative AI leading the pack.

10:04

🛠️ Emerging Tech and AI Startup Trends

The speaker explores emerging technologies such as Edge AI, model efficiency, and multimodal AI, noting the low percentage of companies working in these areas according to Y Combinator's data. The paragraph also addresses critical issues in AI, such as data privacy, AI safety, and explainability, and the startups addressing these concerns. The speaker identifies the potential in sectors like blockchain and Quantum Computing, where a handful of companies are pioneering AI integration. The paragraph concludes with insights into the backgrounds of Y Combinator-backed founders, emphasizing the importance of technical expertise and the value of diverse team compositions.

15:05

🚀 Recommendations for AI Entrepreneurs

In the final paragraph, the speaker synthesizes the analysis to offer recommendations for aspiring AI entrepreneurs. The advice includes focusing on the B2B sector, targeting underserved industries like manufacturing and retail, prioritizing technical expertise, leveraging generative AI innovatively, and addressing ethical concerns proactively. The speaker encourages viewers to conduct their analysis but provides a concise guide based on the insights gathered from Y Combinator's AI startup ecosystem, emphasizing the importance of standing out in a competitive market and preparing for upcoming regulations and challenges.

Keywords

💡Natural Aptitude

Natural aptitude refers to an inherent ability or talent in a particular area. In the context of the video, it emphasizes that the work you choose should align with your innate strengths to ensure success and satisfaction. This concept is crucial because having a natural aptitude for your work makes it easier to excel and overcome challenges.

💡Deep Interest

Deep interest means having a strong passion or curiosity about a subject. The video stresses the importance of being deeply interested in your work to persist through difficulties. This passion drives motivation and commitment, essential for long-term success in any project.

💡Scope to Do Great Work

Scope to do great work refers to the potential for significant achievement and impact in a chosen field. The video highlights the need for projects that offer opportunities for excellence and innovation. This ensures that your efforts can lead to notable and rewarding outcomes.

💡AI Disruption

AI disruption involves the significant transformation of industries through artificial intelligence technologies. The video discusses identifying industries ripe for AI disruption, where innovative AI applications can bring about substantial changes and improvements. Examples include healthcare, fintech, and education.

💡YC (Y Combinator)

Y Combinator (YC) is a leading startup accelerator that provides funding, mentorship, and resources to early-stage startups. The video uses YC's track record of successful startups to identify emerging trends in AI and potential areas for innovation. YC's support helps startups refine their products and scale their businesses.

💡B2B vs. B2C

B2B (Business-to-Business) and B2C (Business-to-Consumer) refer to different market approaches. The video highlights that most successful AI startups focus on B2B, solving problems for other businesses. In contrast, B2C targets individual consumers. Understanding this distinction helps in strategizing the target market for AI innovations.

💡Generative AI

Generative AI involves creating new content, such as text, images, or music, using AI models. The video notes the rapid advancement and popularity of generative AI, with many startups leveraging it for various applications. This technology is pivotal in driving innovation and developing new AI products.

💡Ethical AI

Ethical AI focuses on developing and using AI in a manner that is fair, transparent, and respects privacy and security. The video emphasizes the growing importance of ethical AI due to increasing regulations and societal concerns. Addressing these issues can differentiate startups and build trust with users.

💡Open Source vs. Proprietary

Open source refers to software with source code that anyone can inspect, modify, and enhance, while proprietary software is privately owned and restricted. The video mentions that most AI startups are building proprietary solutions. However, open source projects also play a critical role in AI development and innovation.

💡Technical Expertise

Technical expertise involves having specialized knowledge and skills in areas like computer science, software engineering, and AI. The video underscores the necessity of technical expertise for founding successful AI startups. Founders with strong technical backgrounds are better equipped to navigate the complexities of AI development.

Highlights

Paul Graham suggests that work should have three qualities: natural aptitude, deep interest, and the potential for great work.

The speaker has a background in tech and AI, and is seeking the third quality of offering great success in AI.

Why Combinator's selection process is a valuable indicator of emerging trends and technologies in AI startups.

The analysis focuses on the hottest industries and sectors for AI startups, including untabbed potential and ripeness for AI disruption.

Healthcare and biotech lead with the highest number of AI startups, followed by fintech and developer tools.

B2B sector has a significant majority of startups compared to the B2C sector, indicating investor confidence in B2B.

Automation is the most common use case for AI across industries, with assistance and AI-driven solutions.

Industries like manufacturing, agriculture, and energy are lagging in AI adoption, presenting opportunities for first movers.

Generative AI is the most utilized technology among startups, followed by machine learning and NLP.

Most AI startups are proprietary, with a small percentage focusing on open source.

Real-time AI and multimodal solutions are emerging trends in the AI startup landscape.

Startups are addressing critical issues such as data privacy, AI safety, and explainability in response to regulations.

Enterprise solutions dominate the AI startup scene, with a smaller focus on small businesses.

Emerging technologies like blockchain and Quantum Computing are being explored by a handful of AI startups.

The typical YC-backed AI founder has a strong technical background, often in computer science or related fields.

Recommendations for those interested in AI startups include focusing on B2B, underserved industries, and leveraging generative AI innovatively.

Addressing ethical concerns and regulations in AI is a challenging but potentially rewarding area for startups.

Transcripts

what should I build with AI if this is

the question that you are struggling

with watch this video till the end Paul

Graham who is one of the founders of why

combinator the startup school he says

that the work that you do should have

three qualities first it should be

something that you have a natural

aptitude for second it must be something

that you have deep interest in so that

you can you know overcome challenges

that will come your way while you are

working on that project and third thing

is that particular project that thing

should offer scope to do great work now

I have been working in Tech I like

solving problems in the world of

engineering and AI I have deep and Trust

in AI but now it comes down to the third

question the third quality okay what

should I work on that will offer you

know great success and AI as you all

know is filled with noise at this point

partly because of you know there's so

much hype and everybody wants to write

the wave so the whole field is filled

with too many bad ideas at this point it

comes down to how to find good ideas and

in order to answer this question I

thought why not follow someone who has a

track record of identifying and

nurturing successful ideas in the tech

industry why combinator why combinator's

selection process has consistently

surfaced companies that go on to rehap

shape Industries entire sectors and that

makes their portfolio a valuable

indicator of emerging Trends and

Technologies so all I had to do was to

look at the kind of AI companies and

Founders VC is backing specifically I

wanted to learn what are the hottest

Industries and sectors for AI startups

which Industries have untabbed potential

and you know industries that are ripe

for AI disruption U what all companies

and what all startups are solving

uh problems in emerging Technologies

like blockchain or Quantum Computing and

there must be many companies working on

AI safety because we have so many

regulations coming in so companies

working on data privacy AI safety

accessibility explainability

observability those are the kind of uh

insights that we want and lastly you

also want to understand the typical

background and skills these Founders

have so common traits of YC Founders in

order to understand how practically

feasible it is for you to pursue similar

kind of projects now for those who do

not know why combinator is a leading

startup accelerator that provides seed

funding mentorship and resources to help

early stage startups succeed basically

they invest $500,000 Us in each startup

that is accepted into their three-month

program uh in exchange for a small

Equity stake and uh this program aims to

help help these startups dramatically

improve their product help them with

user growth and also uh increases their

options to raise additional funding now

coming to the data collection

process so I collected the data from

YC's startup directory you have more

than 5,000 companies over here that they

have bagged so far I was only interested

in AI companies that to from last four

CS summer 24 Winter 24 summer 23 winter

23 and the tags that I've selected

artificial intelligence AI generative AI

so all of these AI companies are listed

over here and if you look at you know a

sample page from YC U the name is

provided the description over here the

founder details so I've captured all

these along with all the tags that they

have and cleaned the data captured it in

this a table sheet okay so I have

company name description category these

are 4 17 companies that I collected

while looking at a subset of these

companies I have found many exceptional

use cases and in fact uh part of the

data collection process has been uh done

with the help of company called gum Loop

which is backed by YC which was

previously called agentive okay and I

find myself using gum Loop more than I

had imagine now coming to the analysis

I've tried to capture my entire analysis

in this newsletter a article of mine uh

it's called high signal AI uh the link

will be provided in the description

below the first part of the analysis was

to look at the hottest Industries and

sectors that have adopted AI quickly

where people have found really good use

cases and looking at this plot you'll

see that Healthcare and biotech is the

leader here with 45 companies solving

problems in this industry which accounts

for 10.8% of my data of all the

companies that I have collected followed

by fintech with 38 companies 37

companies are building some sort of

developer tools 34 companies are trying

to automate some sort of sales or

marketing workflow and then 18 companies

in education so if you look at these

sample companies I have this study which

is trying to innovate in this education

industry an AI math tutor for every

student so you have these examples in

each and every industry and Healthcare

and biotech is the leader so far

now next you would want to understand

should you build in B2B or b2c the

numbers here are going to amaze you you

have

338 companies out of

4177 building in the B2B

sector 81.1% of the companies are

solving B2B problems only 18.9%

companies are operating in this b2c

sector and you can find a few examples

like gig ml which is helping inter

Enterprises build and deploy large

language models and then b2c you have

Rex pocket pod shortbread these are the

kind of companies and kind of problems

that people are solving in the b2c

sector so these numbers showcase you

know strong confidence in B2B sector uh

from investors uh and uh b2c has a lot

of untabbed potential uh as you can see

that only 20% of the companies are

operating in this particular sector

infrastructure versus application so

this is obvious U as we in traditional

software engineering majority of the

people are going to build in the

application layer that means they are

going to build some sort of application

using the you know underlying

architecture underlying infrastructure

and for obvious reasons infrastructure

layer is hard to build the skill set

required is also rare and investment

requirements are also high so this data

as well coming from YC may not be

representative of the number of

companies operating in the

infrastructure layer uh which obviously

is going to be you know low uh in number

so 14.9% companies in infrastructure

layer and 355 companies operating in the

application layer uh so 85.1% and

14.9% over here automation has been the

biggest use case of AI across all

Industries and there are two types of

automations one is completely AI driven

and another one is uh co-pilots or

assistants so AI assisted human work

that means you are trying to help humans

deliver faster you are automating some

part of the workflow now here 69.1% of

the companies are trying to build some

sort of assistance to help uh humans

deliver faster and 31% of the companies

are building entire AI driven

automations there are companies like

ofon that are trying to automate order

taking at fast food drive-throughs and

there are companies like constructible

that are building co-pilots for

construction teams helping streamlined

projects ideation reduced losses during

you know due to bad data we've seen

which Industries have adopted AI quickly

but what about the industries that are

still lagging which need more and more

Innovation which need more people to

incorporate AI to solve their problems

and these are manufacturing agriculture

energy retail 16 companies only combined

together in these industries so you need

more and more people and these

industries present opportunities for

first movers in AI adoption now let's

talk about the Technologies specifically

AI technologies that these startups are

leveraging now when I talk about AI

Technologies I talk about machine

learning generative AI natural language

processing computer vision uh media

Generation video processing so on and so

forth and take these numbers with a

pinch of solt because a lot of these

Technologies are overlapping so there

may be companies that are using multiple

AI Technologies to solve their problems

and at the top for obvious reasons we

have generative AI because llms are are

advancing pretty quickly 78 companies

using or building something using gen AI

then we have machine learning 56

companies NLP 47 and computer version 18

as I said there will be many companies

which are using both NLP or generative

AI using both machine learning computer

vision so on and so forth moving on to

open source versus proprietary uh now

this data may not be correct because VC

obviously would not want to you know

fund a lot of Open Source companies so

95.7 % 399 companies are proprietary and

4.3% of the companies are building in

open source please note that there are a

lot of companies out there which are

coming out of Open Source projects if we

talk about other Technologies and other

different types of technical trends that

we see in these AI startups Edge AI so

models running on your phone uh Apple is

doing a lot of work in this particular

category but uh when we look at YC data

only two companies .5% of the companies

mentioned that they're solving something

in Ed aai only five companies mentioned

that they're working or trying to solve

something with model efficiency that is

reducing the computational resources

that it needs to train one large

language model so 1.2% of the compan is

focusing on AI model uh efficiency 46

companies are building something with

realtime AI uh that is voice agents uh

primarily multimodal so approximately 22

companies 5.3% appear to be worth

working on multimodel AI Solutions now

as AI is evolving there are a lot of

regulations there are a lot of concerns

around data privacy AI safety

explainability so on and so forth so

there are layers of critical issues that

are required to be solved now there are

startups that are addressing data

privacy and security concerns 18

companies 4.3% explicitly mentioned that

they're solving something in data

privacy and security sector so cyber

security and data privacy one such

company is Coria there are startups only

five uh so far which are working on

ethical AI or AI safety uh so there's a

lot of potential over there startups

making AI accessible for non-technical

users so there are companies like Creo

which is trying to build internal tools

with AI without coding so no code tools

there are three companies that are

working on explainable a adding more

transparency then we have uh 11

companies solving challenges in climate

Tech uh three companies uh trying to

address issues with bias and fairness

using AI we have ai for small businesses

versus Enterprise Solutions so if you

look at this 70.7% of the companies are

Enterprise solution only 88.9% are built

for small businesses looking at these

numbers we definitely need more and more

people to work on these critical issues

around safety regulation data privacy uh

security and uh there is a lot of

potential to grow uh Within These uh

sectors coming to some hard uh emerging

Technologies which is blockchain and

Quantum Computing now obviously these

Technologies and the fields are so

complex that you would not have a lot of

companies solving these problems but we

have a bunch of them Quantum Computing

there are two companies uh which are

trying to incorporate AI with Quantum

Computing solve something over there

then we have three companies working on

blockchain so conductor Quantum

harnessing Quantum computing to

potentially solve complex problems

beyond the reach of classical Ai and

then we have companies like salio or

kadalo merging blockchain with AI for

enhanced data integrity and

decentralized intelligence so there is

still a lot of potential uh Within These

two sectors specifically I'm I'm more

interested in blockchain how you

integrate blockchain with AI to build

something uh amazing now let's come to

the background of the people who are

willing to work on these problems work

on these startups so typical background

of a YC banked founder here you can see

most more than 75% of the founders have

strong technical expertise in computer

science or software engineering machine

learning or data science especially if

you are an AI founder you have to be

very strong in programming so

educational background most around 20%

of the found ERS are coming from

prestigious institutes like Howard MIT

Stanford Berkeley 25% of them have

previously worked at strong leading tech

companies like Google Facebook meta okay

15% of the founders have prior startup

experience so they have worked on you

know some of their own startups before

and uh 8% of the founders come from

academic research backgrounds phds

postdoc University professors and

45% have co-founding teams so that means

a technical founder plus a business or

operations founder so they together make

uh a good team

24% have backgrounds that position them

to disrupt traditional industry so they

have already uh done deep work within

those Industries and they would bring in

perspective that will help the technical

founder or the business founder to take

it further within that particular

industry now not to mention that if you

have done exceptional work in the past

uh without having these titles or

prestigious institutes in your

background you can still make a mark uh

all you need is you know drive and

showcase some evidence that you can

actually you know uh overcome those

challenges that are going to come when

you will work on something hard

something that y would love to uh you

know invest in now now putting this

entire analysis together has what I

would recommend to anyone but obviously

you know feel free to go through the

entire analysis and build the answer for

yourself I would suggest that you focus

on the B2B sector at this point okay uh

look at underserved uh Industries

manufacturing retail if you know

somebody that's good uh prioritize

technical expertise because you would

need uh technical expertise to thrive in

this business if you do not have that

expertise partner up with somebody who

does then uh leverage generative AI in

an Innovative manner to stand out

because there are many companies which

are just rappers and they'll soon be uh

you know out of business because as soon

as open a releases their next Model A

lot of businesses are going to shut uh

because of that then address ethical

concerns this is a category which is

hard to solve but will have a lot of

potential because there going to be many

regulations coming in new bill are going

to pass and privacy cyers these are

issues that a lot of companies are going

to have to deal with so if you can build

a solution around it yeah you are going

to stand out in that particular category

then so yep that's been it and I hope

that you found this video useful

insightful and if you did please give it

a thumbs up that's the best way you can

help me subscribe and I'll catch you

guys in the next one until then keep

learning and keep building with a