AI

YC’s latest batch sure was a lot of ‘maybe AI can do… this?’

Comment

Robotic arm carrying a mechanical part
Image Credits: Alashi / Getty Images (Image has been modified)

Sitting through hundreds of startups on YC Demo Days, you’re not always sure whether you are actually perceiving patterns or if your brain, as coffee battles with monotony, is inventing them in a kind of pareidolia for business plans. This year, though, the theme was pretty obvious: “AI can do that, probably! Maybe.”

Certainly today’s AI models are more capable than yesterday’s, and yesteryear’s. But we’ve seen over and over how these systems demo well but fall down under systematic requirements or as tools with reliable and repeatable results.

It’s hard not to see this batch as the  precursors of a coming wave of AI-powered shovelware. Pick a use case, do a little fine tuning of an available model (no one actually builds their own), cherry pick some good examples for screenshots and bolt on a prefab UI. Congratulations, you’re now the very first AI social media content generation platform for independent bars and restaurants in the Middle East and North Africa. Buy a couple hundred five-star reviews and you’re on your way!

Now, it’s not that restaurants in Cairo and Beirut couldn’t use a helpful tool to gain some traction online and attract new customers. It’s that having AI, as it currently exists, do something for you is kind of like admitting that it doesn’t matter.

Creating an AI-powered conversation agent that answers the phone at your business sounds good when you frame it as a way to never lose a customer. But what does the customer think when the business they call decides AI is the reception they deserve? Personally, I would hang up and try someone else. What about a trade worker who gets an AI calling to make an appointment? Same thing.

Our favorite startups from YC’s Winter 2023 Demo Day — Part 1

Realizing an email to you has been trivially “personalized” by AI is like being told, we can’t be bothered to personalize our emails, but we want you to think we do. Wouldn’t you feel tricked? It’s a systematic imposture upon the customers.

If your first interview with a company is with a conversation agent or a person obviously reading generated cues from the knowledge base or whatever, do you feel like a person joining a team or a part being sized up for installation? You’re not even worth the full attention of a qualified human.

That’s not necessarily the vibe I got from every AI startup in this YC batch, but I sure got it from a few of them. Here’s a partial (!) list of the “AI can do that, probably” companies I jotted down.

  • Type – AI-first document editor.
  • Iliad – Generate game art assets.
  • Layup – Build workflows across apps with one line command, like onboarding a hire.
  • Nucleus – AI-powered onboarding orchestration that understands “the true nature of a business.”
  • Hadrius – SEC-compliance robo-advisor.
  • Speedybrand – Generated marketing content for SMBs.
  • Quazel – Language learning with an AI tutor.
  • Booth.ai – Generative AI “photographer” for e-commerce.
  • Squack – Natural language accountant tools.
  • Berri.ai – Creating ChatGPT apps as a service.
  • Semantic – Financial news insights “enriched” by AI.
  • Credal.ai – ChatGPT-like interface for employees that references company docs but protects business secrets
  • Defog – Add AI data assistant to your app.
  • Linkgrep – Suggests things from knowledge base and adds to chat or notes live in browser.
  • Sail – Automated sales emails.
  • Aiflow – Automate market research based on reviews and feedback.
  • Tennr – Turn knowledge base into a custom LLM.
  • Truewind – AI-powered bookkeeping and finance processes.
  • Flair labs – Collect insights from customer service call data and emails.
  • JustPaid – Automate bill pay, catch over-payments to vendors.
  • Kyber – Automate insurance industry tasks like answering questions and underwriting.
  • Meru – Platform for training your own LLMs.
  • Sameday – AI that calls workers like plumbers and roofers to make appointments.
  • Zenfetch – Analyze customer calls live and surface talking points.
  • Syncly – AI to analyze customer emails.
  • Pair AI – Video courses generated using AI.
  • Latent – Automating electronic health records.
  • Avoca – AI receptionist to answer missed calls at SMBs.

Until about 30 seconds ago, I actually had appended thoughts about the companies to these brief and likely insufficient descriptions. But I realized the list was in danger of becoming a litany of complaints (not to mention way too long). No one likes to read someone just shooting down ideas left and right, especially when many of those ideas are being worked hard on by people for whom they are important. It’s easy to criticize. So easy someone in the summer batch may try to automate it!

But I challenge you to look at that list and not wonder about some of the entries: Is that really what’s needed? Won’t that need lots of oversight? Doesn’t this introduce liability, or decrease transparency? Did anyone ask customers if they want this? Who verifies and audits the results — another AI? Who is displaced by these tools? Who trains people on them?

Practically every company that presented said they’d gone live a few weeks earlier and miraculously were already at some healthy ARR. But a few weeks is hardly enough time for a major automation tool to be even installed and the documentation read, let alone evaluate its performance and whether it’s worth the price tag. I can’t imagine even half of these have been used, really used, by a potential customer.

One example I can’t help but share: A generative marketing imagery company in its slide had the following prompt for the system to work with: Our classic ketchup is made only from sweet, juicy, red ripe tomatoes for the signature thick and rich taste of America’s Favorite Ketchup. The AI’s copy: SWEET & JUICY KETCHUP FOR ALL! If I was a marketer at Heinz and that was in the demo I was given, I would stand up, thank them for their time and open the door.

Our favorite startups from YC’s Winter 2023 Demo Day — Part 2

Some of the companies admitted they’d pivoted halfway through the program and wrote their first line of code for this new application just recently. Of course we must allow for the adventurous and freewheeling nature of early-stage startups, that’s part of the fun and excitement of the space. But do these companies really feel “innovative” to you? They seem rather to be big fans of innovation, sneaking into its room and trying on its clothes. (“Cute… here, you try it on, fintech.”)

I know I’m underestimating the amount of work it takes even to build the most perfunctory AI-powered B2B SaaS service, but a lot of these feel like our old hackathons where someone would make an API available and everyone would try to shoehorn it in to the most realistic-sounding application, hoping to get that $1,000 gift card from SAP or whatever. There’s joy in the process of creation, but the results don’t really stand on their own.

Probably I’ll be proven wrong when one of these companies goes unicorn and everyone laughs at the TechCrunch writer who doubted them. But I can’t shake the worry I felt in hearing founder after founder say with such conviction that their AI could do something better, when I suspect that conviction has been cultivated upon false pretenses.

More TechCrunch

Looking Glass makes trippy-looking mixed-reality screens that make things look 3D without the need of special glasses. Today, it launches a pair of new displays, including a 16-inch mode that…

Looking Glass launches new 3D displays

Replacing Sutskever is Jakub Pachocki, OpenAI’s director of research.

Ilya Sutskever, OpenAI co-founder and longtime chief scientist, departs

Intuitive Machines made history when it became the first private company to land a spacecraft on the moon, so it makes sense to adapt that tech for Mars.

Intuitive Machines wants to help NASA return samples from Mars

As Google revamps itself for the AI era, offering AI overviews within its search results, the company is introducing a new way to filter for just text-based links. With the…

Google adds ‘Web’ search filter for showing old-school text links as AI rolls out

Blue Origin’s New Shepard rocket will take a crew to suborbital space for the first time in nearly two years later this month, the company announced on Tuesday.  The NS-25…

Blue Origin to resume crewed New Shepard launches on May 19

This will enable developers to use the on-device model to power their own AI features.

Google is building its Gemini Nano AI model into Chrome on the desktop

It ran 110 minutes, but Google managed to reference AI a whopping 121 times during Google I/O 2024 (by its own count). CEO Sundar Pichai referenced the figure to wrap…

Google mentioned ‘AI’ 120+ times during its I/O keynote

Firebase Genkit is an open source framework that enables developers to quickly build AI into new and existing applications.

Google launches Firebase Genkit, a new open source framework for building AI-powered apps

In the coming months, Google says it will open up the Gemini Nano model to more developers.

Patreon and Grammarly are already experimenting with Gemini Nano, says Google

As part of the update, Reddit also launched a dedicated AMA tab within the web post composer.

Reddit introduces new tools for ‘Ask Me Anything,’ its Q&A feature

Here are quick hits of the biggest news from the keynote as they are announced.

Google I/O 2024: Here’s everything Google just announced

LearnLM is already powering features across Google products, including in YouTube, Google’s Gemini apps, Google Search and Google Classroom.

LearnLM is Google’s new family of AI models for education

The official launch comes almost a year after YouTube began experimenting with AI-generated quizzes on its mobile app. 

Google is bringing AI-generated quizzes to academic videos on YouTube

Around 550 employees across autonomous vehicle company Motional have been laid off, according to information taken from WARN notice filings and sources at the company.  Earlier this week, TechCrunch reported…

Motional cut about 550 employees, around 40%, in recent restructuring, sources say

The keynote kicks off at 10 a.m. PT on Tuesday and will offer glimpses into the latest versions of Android, Wear OS and Android TV.

Google I/O 2024: Watch all of the AI, Android reveals

Google Play has a new discovery feature for apps, new ways to acquire users, updates to Play Points, and other enhancements to developer-facing tools.

Google Play preps a new full-screen app discovery feature and adds more developer tools

Soon, Android users will be able to drag and drop AI-generated images directly into their Gmail, Google Messages and other apps.

Gemini on Android becomes more capable and works with Gmail, Messages, YouTube and more

Veo can capture different visual and cinematic styles, including shots of landscapes and timelapses, and make edits and adjustments to already-generated footage.

Google Veo, a serious swing at AI-generated video, debuts at Google I/O 2024

In addition to the body of the emails themselves, the feature will also be able to analyze attachments, like PDFs.

Gemini comes to Gmail to summarize, draft emails, and more

The summaries are created based on Gemini’s analysis of insights from Google Maps’ community of more than 300 million contributors.

Google is bringing Gemini capabilities to Google Maps Platform

Google says that over 100,000 developers already tried the service.

Project IDX, Google’s next-gen IDE, is now in open beta

The system effectively listens for “conversation patterns commonly associated with scams” in-real time. 

Google will use Gemini to detect scams during calls

The standard Gemma models were only available in 2 billion and 7 billion parameter versions, making this quite a step up.

Google announces Gemma 2, a 27B-parameter version of its open model, launching in June

This is a great example of a company using generative AI to open its software to more users.

Google TalkBack will use Gemini to describe images for blind people

Google’s Circle to Search feature will now be able to solve more complex problems across psychics and math word problems. 

Circle to Search is now a better homework helper

People can now search using a video they upload combined with a text query to get an AI overview of the answers they need.

Google experiments with using video to search, thanks to Gemini AI

A search results page based on generative AI as its ranking mechanism will have wide-reaching consequences for online publishers.

Google will soon start using GenAI to organize some search results pages

Google has built a custom Gemini model for search to combine real-time information, Google’s ranking, long context and multimodal features.

Google is adding more AI to its search results

At its Google I/O developer conference, Google on Tuesday announced the next generation of its Tensor Processing Units (TPU) AI chips.

Google’s next-gen TPUs promise a 4.7x performance boost

Google is upgrading Gemini, its AI-powered chatbot, with features aimed at making the experience more ambient and contextually useful.

Google’s Gemini updates: How Project Astra is powering some of I/O’s big reveals