The generative AI landscape has seen warp-speed growth in recent months.

Take ChatGPT, for example. The world-famous chatbot boasts 25 million daily visitors and currently holds the record for the fastest-growing consumer Internet app—reaching one million users in just five days.

Globally, popular platforms such as Duolingo, Khan Academy, and Stripe are also embracing the generative AI trend, introducing AI-powered features into their platforms.

But perhaps generative AI’s meteoric rise is most evident in the startup space. On Y Combinator’s Startup Directory, the prestigious startup accelerator known for backing successful ventures like Airbnb, Coinbase and Dropbox, the number of startups tagged as “Generative AI” in its Winter 2023 batch made up 22 per cent of the cohort, surpassing the combined number of startups with the same label over the past five years.

Across the globe, generative AI is now being used to power a diverse range of industries, from content marketing to deep tech engineering. At the time of writing, the number of generative AI startups under Y Combinator stands at 108 and counting.

The homegrown startups making waves in their respective verticals

Standing out in one of the fastest-moving tech spaces is no easy feat.

Against all odds, these three trailblazing homegrown startups have received the prestigious backing of Y Combinator and are continuing to lead the pack in their respective verticals.

Also Read: Use AI to clone yourself with Oleksandr Vysotskyi

We speak to the Founders to learn more about what they think of humanity’s future with generative AI.

Hypotenuse AI (YC20)

Hypotenuse AI was founded in 2020 by Singaporeans Joshua Wong, a former machine learning scientist at Amazon Alexa and Cambridge University alumni, and Low Lin-Hui, an early hire in Stripe’s APAC team and ex-founder of her own e-commerce brand.

What does Hypotenuse AI do?

Hypotenuse AI is an AI content marketing platform that manages and creates content for businesses, including blog articles, product descriptions, AI images and advertising copies. Hypotenuse’s AI is built using proprietary machine learning techniques designed especially for marketers and copywriters to write high-quality, high-conversion content that matches a brand’s tone and style.

How did Hypotenuse AI come about? What inspired you to start it?

Through a personal pain point.

During my time as a machine learning researcher at Amazon, I was helping a colleague to start an e-commerce website selling vegan soap bars. We realised that the most frustrating part of this process for us was getting the copy written for product descriptions and marketing.

Given my background, I started building out AI models to automate this process and realised that while it was a challenging problem, it was possible to solve. I went to speak to a dozen businesses and realised that it wasn’t only a problem for us but for many online companies out there.

So I left Amazon to solve this problem for all businesses out there.

Generative AI tools are becoming increasingly popular in recent months. What are your predictions over the next year?

Generative AI will become widespread in marketing teams globally. The hallucination problem (of AI making up facts) still won’t be solved, but it will be greatly reduced to the point that it beats human error rates in many areas. The cost of large language models will continue to fall. Multimodal models will rise—ones that work across video, audio, text and more. Advancements will keep pouring out of research, and the hype around AI will still continue to grow. It won’t slow down yet.

Smaller open-source AI models will gain popularity, but the biggest advancements will still come from large language models that are too gigantic for the average user to train. Many generative AI start-ups will die because they created fancy demos but never found product market fit. A handful more billion-dollar generative AI start-ups will be spawned.

Governments will begin to regulate the development of the most advanced AI models. Educational institutions will be in an existential debate, but good ones will embrace generative AI carefully and adapt parts of it to fit into their curriculum. PhDs in AI become less relevant, but engineering with these models become increasingly in demand. The face of things we took for granted as constant in our lives would change—including how web search is done, how we find the right words to explain concepts, and

Yuma.ai (YC W23)

Yuma.ai was founded in 2023 by Guillaume Luccisano, a French software engineer and serial entrepreneur living in Singapore and a seed investor in over 50 startups in the US and France. His first startup, Socialcam, was acquired by Autodesk in 2012.

What does Yuma do? 

Yuma is ChatGPT for customer support. We focus on helping Shopify merchants with their support burden. We are a one-click install app on top of their helpdesk software, such as Gorgias or Zendesk. Once installed, we start suggesting the best draft for answering support inquiries.

How did Yuma come about? What inspired you to start it?

Yuma got started by accident. This is my third YC startup after Socialcam (W12) and Triplebyte (S15). I started playing with LLMs in late 2022 and released Yuma as a prototype for fun in mid-December 2022 and was overwhelmed with demo requests. That’s when I knew I was onto something and had to turn this into a real company once again.

Generative AI tools are becoming increasingly popular in recent months. What are your predictions over the next year?

The pace of change is currently astonishing. Although we may not yet fully realise it, AI will have a profound impact on our day-to-day lives, potentially as significant as the introduction of the smartphone and the internet.

Many roles will need to adapt or change, and we will have to learn how to harness the power of AI. Education, for example, will never be the same. On a global scale, AI will become a major geopolitical focus and will be a key competitive advantage for nations that own the best technology.

Also Read: Regional expansion, careful approach to fundraising remain key for SEA fintech startups to grow

Defog.ai (YC W23)

Defog.ai was founded in 2021 by Medha Basu (left), a former journalist, and Rishabh Srivastava (right; interviewee), a full-stack developer and data scientist living in Singapore.

What does Defog.ai do?

Defog is an AI data assistant that developers can add to their apps—no machine learning expertise is required. For example, fintech companies use Defog to help their users analyse complex expense trends just by asking questions in plain English.

How did Defog.ai come about? What inspired you to start it?

Prior to starting Defog, Rishabh had run a data company for six years. His biggest limitation in scaling the business was the time spent in servicing ad-hoc questions asked by customers. With LLMs improving, he realised that an AI system that could generate and execute Python or SQL code for answering user questions was becoming feasible. And that led to the start of Defog.

Generative AI tools are becoming increasingly popular in recent months. What are your predictions over the next year?

Generative AI will start moving from consumer and SME usage to the enterprise. Privacy and governance considerations will be critical for this, and we will see smaller models that can be deployed on-premises to do specialised tasks get more adoption.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. Share your opinion by submitting an article, video, podcast, or infographic

Join our e27 Telegram group, FB community, or like the e27 Facebook page

Image credit: Canva Pro

The post Singapore’s YC-backed AI startups making waves globally with generative technology appeared first on e27.