Show me YC’s batch & I tell you how the future looks like

Hi friends 👋,

Happy Thursday from sunny Copenhagen.

A short quiz to start the day & today’s article. What is a shared attribute by these unicorns: Airbnb, Stripe, Dropbox, Coinbase, DoorDash, Reddit, OpenSea, Deel, Instacart, Scale, Zapier, Webflow, Jasper.

All these companies have been through Y Combinator. The quiz wasn’t too difficult given today’s title ‘Show me YC’s batch & I tell you how the future looks like’ 😄 We learned: Paul Graham and Jessica Livingston - YC founders - are pretty good at spotting trends early. Here are specific examples:

Coinbase went through YC in 2012. Way before crypto became mainstream. (my grandmother asked me about Bitcoin in 22’).

• Dropbox in 07’. Long before cloud storage became a thing.

• Airbnb in 09’. Long before the sharing economy was a thing.

• Zapier in 12’ & Webflow in 13’. Long before no-code tools were a thing.

• Scale AI in 16’. Long before the AI hype boom.

I believe you get where I am going with this. The companies that go through Y Combinator today give us a glimpse of how the future might look like in 5 or 10 years.

Curious as I am, I spent my evenings reading up on the W23 batch. Let’s see if there are anything interesting insights we have learned.

Here’s what I got for you today:

• AI YC w23 batch ‘landscape’ overview - all startups at a glance

• crunching some numbers - noteworthy trends!?

• the five musketeers - my favorite W23 batch AI startups

• a long list of ¡all! W23 ai startups

The AI YC w23 batch startup landscape

All AI YC w23 batch startups at a glance:

Yes. I did that. I copied & pasted all those logos myself. I guess there is a real AI use case. Create an AI to automate logo copy, pasting & segmentation for VC startup landscape overviews. Might save lots of hours for all those VC analysts. And for me. And no. I’m not a VC analyst. I do this for fun.

Numbers don’t lie

What is the problem that most YC AI startups are currently trying to solve? In this section, I put some numbers into perspective & tried to find some interesting trends & hot topics.

Total W23 batch size: 223 companies. YC has reduced its batch size from last years W22 batch by almost 50% (W22 batch: 397 companies)

28% of the entire batch are AI startups. 61 AI startups in total. The previous batch - s22 - only had 31 AI startups. Both shows, how hyped AI startups currently are. But we already knew that.

The money is in B2B. Of those 61 startups 80% focus on B2B.

The hottest AI categories. ML Ops Platform. Developer tools. Sales & Customer success (each briefly discussed below).

‘ML Ops’ is the hottest category. You will have to get used to the new buzzword: “ML Ops”. These are tools that help you in operating your language model (LLM). Some examples:

• Customize/ optimize your model to your needs.

• Run your language model more efficiently by using less GPU.

• Manage your language model by understanding how many tasks it performs in an hour.

I believe a lot more startups will pop up in this category. As large corporates adopt AI, they will need tools to customize, train, manage and monitor these models. And they won’t build it themselves. There will be a lot of money made here.

Make the developer more productive, aka developer tools. GitHub’s Copilot and Replit’s ghostwriter have shown that AI can turn a 10x engineer into a 100x engineer. Businesses are pushing their developers to use AI to write code. They get 50% more code without hiring a new developer. It’s a no-brainer. And more & more startups want a piece of the pie. Developer tools are a hot category right now.

The time of big customer service centers comes to an end. AI startups want to take over customer success. The current YC batch is full of founders trying to automate customer service with AI.

AI as a superpower for your sales team. The first words popping in our head when we hear AI: saving costs/ reducing headcount through automation. But: AI might also help the revenue side of the balance sheet. Many YC companies are working on building tools for your sales team to close more deals.

My five picks

Here’s a list of five startups that I find particularly exciting. I hope you do, too.

1/ The holy guacamole: EzDubs 
EzDubs are the geniuses behind real-time video translations. I will finally be able to send my german speaking grandmother interesting documentaries I find on Youtue (most of them are english only). Watch this video and be amazed:

2/ The smart play of the day: Codecomplete
You probably know GitHub’s copilot. GitHub’s claim: Developers using their Copilot complete a given task 55% faster than developers who don’t use it. The problem: large Enterprises won’t let their developers use GitHub’s Copilot because of security & privacy risks. Codecomplete is building a self-hosted developer copilot for Enterprises.

3/ The no-brainer: OfOne
Your drive-thru BigMac order will be taken by AI. And you’ll not even notice a difference. OfOne wants to reduce costs for restaurants by replacing humans with AI. This might be unpopular as it plays into the “AI is taking our jobs” argument. Harsh truth: This is the future.

4/ The one with big traction: Latent
LatentHealth wants to automate healthcare operations. After 2.5 months of operation, they already signed enterprise contracts with two of the largest health systems in the US worth >300k in ARR. Decent start.

5/ The one that I am rooting for: TorMagnus Pharmaceuticals
Scientists leveraging AI to find new treatments & develop new drugs for Parkinson's treatment. If successful, this might truly change lives.

The long-list of all YC W23 AI companies

Here’s the list of all AI companies in the current YC batch. They are building a lot of exciting stuff! Please reach out, if I missed one.

🧑‍💻 Developer Tools

#1. Buildt (link) - Search, understand and augment your codebase in seconds using AI

#2. CodeComplete (link) - It’s a GitHub Copilot but for Enterprise. Large companies can't use GitHub’s Copilot because of security and privacy risks. CodeComplete is building a self-hosted version fine-tuned to the company's codebase.

#3. Meru (link) - ChatGPT but for data & code. Meru helps developers to speak to their Documents, HTML Files, JSON files in natural language

#4. Lasso (link) - Automation across any platform: It’s like Zapier, but for any application - an API for anything

#5. Traceloop (link) - Traceloop’s AI generates code to monitor your systems or to test changes in your code.

🏗️ Infrastructure

#6. Anarchy (link) - Anarchy provides a platform for building customized chatbots on top of existing LLMs.

#7. BerriAI (link) - BerriAI is a platform that allows you to ingest your data, build, iterate and deploy production-ready LLM apps in minutes.

#8. Helicone (link) - Provides analytics infrastructure to monitor GPT-3 usage, costs, and latency metrics.

#9. Pyq (link) - Pyq is an easy and affordable way to integrate machine learning into your application by providing all the cloud infrastructure and setup.

#10. Waveline (link) - The fastest way to integrate AI into your business.

🤖 ML Ops

#11. Baseplate (link) - Baseplate is a tool that makes it easy to build high-performance LLM apps

#12. Flower (link) - Train AI on distributed data. Federate any workload, any ML framework, and any programming language

#13. GradientJ (link) - Tools for teams to test, deploy, and manage large language models at scale

#14. Poplar ML (link) - PoplarML lets you deploy any machine learning model to a fleet of GPUs as a ready-to-use and scalable API endpoint with one command

#15. Vellum (link) - The dev platform for building production-worthy LLM applications

#16. Ivy (link) - Run any ML code, with any ML framework, on any hardware

#17. Stack (link) - Fine-tune and compose Large Language Models to automate business processes

#18. Rubbrband (link) - Rubbrband helps developers easily train open-source machine learning models by resolving dependencies and provisioning servers

#19. Texel (link) - accelerate your cloud GPU workloads by optimizing scaling, deployment, and latency. You can run a generative AI model or encode a video on our service up to 10x faster than standard solutions while cutting costs by up to 90% and reducing engineering risk.

#20. Chima (link) - Tools to help large companies customize their LLMs.

🧮 Data Analytics

#21. BaseLit (link) - Baselit makes it easy for product teams to get answers from data. For example, a PM can ask a question in plain English without having to know how the data is stored

#22. Blocktool (link) - Build a financial data warehouse that you can plug in to your spreadsheet and BI tools with no-code.

#23. Defog (link) - ChatGPT for data, embedded in your app

#24. Lume (link) - No-code Tool to Generate and Maintain Custom Data Integrations. Automate integrations with natural language

#25. Outerbase (link) - The interface for your database.

#26. Rollstack (link) - Elevate your data storytelling and automate your slides and docs. Rollstack is a tool that helps its users create stunning and high-quality visualizations

🧾 Finance

#27. AiFlow (link) - AiFlow uses LLMs to help private equity firms automate due diligence so they can close deals faster. Their models reduce the pain of digging through hundreds of documents

#28. AlphaWatch AI (link) Natural language search with AI chatbot for financial documents. Makes your market research easier by answering any broad question on industry, stock prices, and company trajectories

#29. JustPaid (link) - AI-powered controller. Save Money By Continuous AI Contract Insights & Avoid Billing Errors.

#30. Truewind (link) - AI-powered bookkeeping and financial modeling for startups

💼 Sales

#31. Bluebirds (link) - Helps sales teams to resell to past customers who switched jobs. Automatically detect job changes, and deliver hundreds of warm leads

#32. Coldreach (link) - AI-powered cold outreach personalization. Coldreach auto-generates hyper-personalized outreach messages that book more demos while saving SDRs 90% of the time.

#33. Fabius (link) - Fabius takes the questions you ask in pipeline reviews for a handful of deals and answers them for your entire pipeline. For example, Fabius flags open deals where the rep hasn’t discussed the budget (which is critical today as companies re-evaluate spending)

#34. Rubber Ducky Labs (link) - Helps e-commerce companies avoid tone-deaf product recommendations - like featuring ski jackets in June

#35. Salient (link) - An AI-powered sales platform that personalizes your outbound at scale, automatically responds to customers, and proactively re-engages with prospects at the right time.

#36. Tennr (link) - Wants to turn every sales rep into a top performer. Tennr uncovers what your stakeholders care about and shows you how you’ve won similar deals in the past using conversational AI trained for your business.

🤗 Customer Success

#37. Buff (link) - Infrastructure to build AI support agents. Optimize deflection rates, response times, and CSAT, all without increasing headcount

#38. Kyber (link) - ChatGPT for Enterprise Support. For example, when Branch Insurance's appointed agents have a question while selling a policy, instead of having to wait hours for help from support staff, they can simply ask the Kyber Bot and immediately receive a factual answer with a link to where the answer was taken from, allowing them to close deals faster

#39. OfOne (link) - Replaces the human at the drive-thru. Instead of a human, an AI receives and processes the order.

#40. OpenSight (link) - AI-powered customer support automation for fast-growing companies. OpenSight claims: Get 24/7 live customer support; With near 100% accuracy; At 93% less cost

#41. Parabolic (link) - Parabolic is an AI assistant that autodrafts ready-to-send responses for your customer support team.

#42. Vector (link) - Vector makes it easy for account managers to understand what customers are actually saying, so they can focus on the right retention experience.

#43. Syncly (link) - We help revenue operations and customer success teams to prioritize customer issues by automatically analyzing their email communication.

🎁 Marketing

#44. CreatorML (link) - CreatorML helps YouTubers come up with better video ideas and packaging so that their viewers are more likely to click and watch longer

#45. SpeedyBrand (link) - Digital marketing solution for SMBs powered by Generative AI. Speedy generates website traffic for SMBs by improving their online presence: SEO-Optimized digital content

#46. Type (link) - The AI-first document editor that helps you write fast.

🔬 Life Science

#47. DiffuseBio (link) - Generative AI for drug design. Diffuse Bio is building a push-button AI software platform for molecular design, leveraging breakthroughs in generative AI

#48. Latent (link) - Automate healthcare operations. Their product streamlines insurance authorization by analyzing EHR records and surfacing the most relevant data

#49. Lavo LifeSciences (link) - AI-accelerated quantum chemistry for rapid drug development

#50. Scanbase (link) - The API for Diagnostic Test Analysis (COVID-19, FLU, RSV, STD

#51. TorMagnus Pharmaceuticals (link) - Leveraging genetics, AI and computation to find new therapies for disease.

🤳 Image

#52. Booth (link) - Create pro quality product photography with AI. Quickly generate high quality lifestyle photos by simply specifying your desired shot and uploading some sample product images.

#53. Iliad (link) - Creative tool for generating game assets.

🎤 Voice

#54. Play.ht (link). Generate or clone any voice and use it to turn any text into. Makers of podcast.ai

#55. EzDubs (link) - Real-time AI translation in videos. Break down language barriers to videos and live streams. Engage viewers from multiple demographies.

📹 Video

#56. Decoherence (link) - Create video from prompts. Decoherence makes it easy for anyone to create videos with AI in minutes rather than hours. We enable creators to quickly generate music videos or B-roll that matches their content.

#57. Linum (link) - DALLE for Video

🧑‍🏫 Learning

#58. Pair (link) - TikTok but for learning. A mobile course platform with AI coaching tools

#59. Quazel (link) - Language learning through natural conversations with an AI tutor

🌳 Climate Tech

#60. KeelingLabs (link) - Next-Generation Energy Management for Grid-Scale Batteries. Helps grid-scale battery operators adapt to constant changes in the grid with energy management software that automatically adjusts its behavior to maximize revenue.

🧠 Knowledge management

#61. Credal.ai (link) - AI written Cliff Notes on what your team got done. Credal saves managers hours every week by automatically drafting slides and memos for all hands, retros and leadership alignment meetings. We identify the key changes each week in your planning documents, spreadsheets and SaaS, and then writes an update that anyone at the company can understand.

I hope you found this insightful.

Chris

*p.s.: I’m hosting a Podcast where I interview historical figures about the modern world (using AI). Check it out: spoti.fi/3Sl3NrD