• The startup CTGT has raised over $7 million in a seed funding round led by Gradient Ventures.
• The company addresses the growing demand for domain-specific, customized machine learning models.
• Other investors include General Catalyst, Y Combinator, Liquid 2 Ventures, and Paul Graham.

CTGT, a startup focused on helping enterprise clients train and deploy machine learning models at lower cost, has raised $7.2 million in a funding round led by Gradient Ventures, Google's AI-geared venture arm.

The deal shows investors are still optimistic about startups developing tools to train models, as companies like Google, OpenAI, and Elon Musk's xAI jostle to produce a foundation model more capable than all others.

Cyril Gorlla, cofounder and CEO of CTGT, says many businesses have access to these generalized models but can't deploy them because of the inherent risks. It may hallucinate a response or display bias. For example, a healthcare provider can't risk a medical chatbot giving dangerous advice. A tech company wouldn't want a digital agent referring customers to its competition.

CTGT's solution allows enterprise clients to take one of these off-the-shelf foundation models and imbue it with their brand voice and data. It claims to lower the massive amounts of computing resources needed through a process called feature learning, where a system learns to recognize patterns and structures in raw data without being explicitly programmed.

Gorlla says the platform actively monitors and audits a client's custom models, allowing it to better spot and eliminate unwanted behavior. Enterprises can also refine and retrain models on the fly without the need to take models offline for updates.

Vig Sachidananda, a partner at Gradient, says the startup addresses a need for more reliable and bespoke alternatives to generalized models.

"We're seeing AI models rapidly progress in capabilities," Sachidananda said. "As models take on more tasks and those with increasing complexity there is a growing need from enterprises for systems that can help improve safety, provide more accurate responses, and mitigate biases."

"We're excited about CTGT as they're developing a novel approach to tackle these problems by helping enterprises interpret and control how models internally understand concepts, such as those that can relate to unsafe or unexpected behavior," he added.

Other investors in the company's seed funding round include General Catalyst, Y Combinator, Liquid 2 Ventures, and angel investors like Paul Graham, Zapier cofounder Mike Knoop, François Chollet, creator of the Keras deep-learning library, and one of Facebook's first employees, Taner Halıcıoğlu.

Gorlla, 23, started the company last year with a former classmate, Trevor Tuttle, at the University of California San Diego. The two dropped out and were admitted to Y Combinator, the storied startup accelerator behind Airbnb, DoorDash, and Twitch.

CTGT, which gets its name from Gorlla and Tuttle's scrambled initials, says it's already testing its tech with three unnamed Fortune 10 companies. One of its first customers, the professional accounting firm Ebrada Financial Group, uses the platform to improve the reliability of its frontline customer service chatbots.

• Index raised over $2 million in seed funding led by Bain Capital Ventures and Blackbird Ventures.
• The company offers tools for product managers to track ideas and build roadmaps.
• The Australia-based startup's beta app is now open to all customers on its waitlist.

Australia-based startup Index, which makes a planning tool for product managers and software engineering teams, recently raised over $2 million in seed funding.

Bain Capital Ventures and Blackbird Ventures, a venture capital firm based in Australia and New Zealand, led the round. Bond Capital and Y Combinator also participated.

Index allows product managers and their teams to track ideas, manage feedback, and build roadmaps for their products using tools like virtual whiteboards and tables. It's compatible with Linear, an engineering delivery software. Index also connects to CRM platforms Hubspot and Attio, customer messaging software Intercom, revenue intelligence platform Gong, and others.

Index's customers tend to be product managers at B2B companies with 50 to 500 employees. After Index's launch on LinkedIn in November, product managers signed up in droves, creating a waitlist of at least 1,000 people who work at startups like Ramp, Index cofounder and CEO Simon Kubica told Business Insider. As of Tuesday, though, Index's beta app is now accessible to all customers who had pre-ordered Index and were on the company's waitlist.

Before starting Index, Kubica spent four years as a product manager at Atlassian, the Australian company known for developing project management tools like Jira, a popular software among teams at many tech companies.

At Atlassian, Kubica led multiple teams and product strategy, but he found it difficult to oversee his products while using "a mix of spreadsheets, documents, whiteboard tools, and essentially trying to stitch together" product workflows, Kubica said. "Atlassian is the preeminent collaboration tools maker, but I found that exceptionally hard to do."

Kubica then contacted his friend, Christian Iacullo, to see if he also found it difficult to manage products across various tools at his job. At the time, Iacullo was a software engineer at Canva, where he worked on the company's developer platform and education product, according to his LinkedIn. Before Canva, Iacullo had also worked at Atlassian. "Christian was kind of my first user research call," Kubica said.

Shortly after the call, the duo began brainstorming ways to simplify product planning. Soon after, they applied and were accepted to Y Combinator's winter 2023 cohort. Iacullo is now Index's cofounder and CTO.

The product management tools industry is crowded. A handful of these include public companies like Monday.com and Atlassian, which both saw recent share price spikes after reporting quarterly earnings.

Privately held competitors include Productboard, which last raised $125 million in Series D funding in January 2022. Productboard allows teams to road map products and integrates with workplace productivity tools such as Jira and Slack. There are also earlier-stage competitors in the space, such as Fibery, another product management platform, which has raised $8.3 million in funding.

Index will use the new funding to accelerate hiring in engineering and other roles across the company, Kubica said. In addition to Kubica and Iacullo, Index employs two software engineers. "We did our viral social media launch in November," Kubica said. "That really blew up, and so now we want to pour fuel on the fire and go faster."

• The storied startup accelerator Y Combinator is seeking applications for its Spring 2025 batch.
• YC shared in a recent "Requests for Startups" that it's looking for various AI-related ideas.
• From AI agents to a new AI app store, here's what 10 YC partners are looking for in the next batch.

The deadline to apply for Y Combinator's Spring 2025 batch is February 11 — and the program this time around is going all-in on AI applications, calling for founders with ideas on AI agents, better ways to handle GPU overload and AI-enabled security.

The famed accelerator each year hosts four batches of early-stage startups for a three-month program filled with development and mentorship. Tens of thousands of startups typically apply for a few hundred spots, and the program takes place in person in San Francisco. YC also invests $500,000 in every selected company and its alumni include Stripe, Airbnb, DoorDash, Coinbase, and Instacart.

For Spring 2025, which runs from April to June, YC partners said they're particularly interested in various new ways to harness and use artificial intelligence.

The investors — who include president and CEO Garry Tan and group partners Diana Hu and Michael Seibel — published their thoughts in the firm's latest "Requests for Startups" post, which is typically published annually. In line with the rapid pace of AI innovation, however, Spring 2025's RFS was published just three months after the previous one.

Here's what the firm is looking for.

AI Agents

Like many tech investors and leaders, YC partners believe that AI agents will play a pivotal role in shaping the next generation of startups, and one major opportunity lies in developing AI personal assistants capable of managing everyday tasks — think scheduling, email organization, and travel planning.

By automating these functions, AI-powered agents could free up time for individuals and businesses, leading to greater efficiency in both professional and personal settings.

"Software is a proven way to bring to everyone what only the rich could afford before," wrote group partner David Lieb. While typically only the wealthy can afford accountants, lawyers, personal trainers, and private tutors, new tech is set to change that, he said.

"Over the next few years, we expect AI to get good enough to do most of these jobs," Lieb said.

AI agents will also play an important role in software development, as well as in other verticals such as customer support, marketing, legal, and healthcare. By tailoring AI solutions to specific sectors, startups can deliver more effective automation and enhance decision-making processes.

AI agents could themselves become customers, said YC partner Dalton Caldwell, who sees an opportunity in a new B2A market.

"With the advent of AI and agents, it seems like it's a good idea to build software and services where agents as customers are actively supported and documented rather than an edge case," he wrote. "For example, APIs to help agents pay for hosting credits, or book travel, or engage into contracts with other parties. In the stock market, it's well understood that humans and programs are trading together, and in the future, this will only increase."

AI infrastructure

As AI grows increasingly complex, YC says startups have a big opportunity to improve test-time compute, reduce latency, and enhance model performance for real-world deployment.

"There's room to rebuild the stack here: better software at inference-layer tooling, cheaper ways to handle GPU workloads, and optimizations that let AI apps scale without bleeding money," Hu, one of the group partners, writes.

Founders with expertise in systems programming who can build foundational infrastructure to support the next generation of computing are also encouraged to apply to YC. As Caldwell put it, there's a big demand for software innovations, from optimized compilers and new programming languages to breakthroughs in operating systems and networking.

"Being able to understand and innovate on the entire software stack can be an unfair advantage to a startup founder," he said."

YC also sees opportunities in data centers and AI commercial open-source software or AICOSS.

"We need more data centers that can be built faster and at a lower cost to support the infrastructure needed for AI progress," Hu said. "Hyperscale datacenter projects take years to complete, and with the current interest and funding, we need new companies and more innovative solutions to accelerate this buildout — whether in power infrastructure, cooling, material procurement, or project management."

Caldwell added that while big tech companies like Google and Facebook have historically open-sourced some tools, they don't typically provide commercial support for businesses — which means opportunities for upstarts. Silicon Valley is currently in a frenzy over DeepSeek, a Chinese company with open source powerful AI models that compete with OpenAI's at a fraction of the cost.

"There are going to be many winners in the space of open source AI, but the launch of DeepSeek should provide ample new ground that a founder might want to cover to help businesses make use of these systems," he said.

Security

YC is also recommending that Spring 2025 applicants address document security, audit automation, and AI software distribution. Seibel, one of the group partners, is looking for an AI-powered "DocuSign 2.0."

"Everyday, complicated documents requiring e-signature create friction in people's business and personal lives," he said, adding that current solutions make it difficult to properly format documents and correct errors. He's looking to back solutions that easily create new document templates, auto-fill information, or offer a voice agent to guide their human counterpart through difficult sections.

Startups will also be of interest to YC if they can build AI-powered compliance tools that streamline risk assessments, automate reporting, and monitor regulatory adherence in real time, helping companies reduce costs and stay ahead of evolving regulations.

Seibel is also ready to see a new AI App Store that protects user data while offering developer infrastructure and tools to help users find AI apps.

"Imagine a travel AI that's great at finding flights that knows you usually travel with your nine-year-old who loves staring out the window, or an AI helper that recommends the original text an idea appeared in when you're reading essays or books," he said. "Done right, this will create more opportunity for startups and founders: their apps will be smarter with shared memory, and this will turn into a new marketplace that solves both distribution and monetization."

• Google's cofounder Larry Page made an appearance at a Y Combinator meeting in December.
• He spoke to a group from the fall 2024 class of startups.
• He talked about the origins of Google and listened to founders talk about their startups.

Google's cofounder Larry Page made a rare public appearance last month during a Y Combinator event for startups in San Francisco, Business Insider has learned.

Page showed up at an "office hour" for a group from the fall 2024 class of startups during the final days of the event, where he listened to founders talk about their business ideas, a person who attended said. Lulu Meservey, a Y Combinator spokesperson, confirmed Page's attendance at the event.

Y Combinator is a startup incubator that holds seasonal classes — known as "batches" — where budding startup founders can network and get advice from Silicon Valley luminaries.

Page was brought along to the event by the Y Combinator group partner Paul Buchheit, the attendee said, adding that Page also shared some anecdotes about the early days of Google. The person asked to remain anonymous because they had not been permitted to speak with the media about the event.

A public sighting of Page is extremely rare these days.

He and his Google cofounder, Sergey Brin, stepped away from their executive roles at Alphabet in 2019, giving the pair time to pursue other business interests such as flying cars and disaster relief. The pair still sit on the Alphabet board and hold shares of a special voting stock that gives them ultimate control of the company.

While Brin has returned to work on AI projects at Google, Page has remained more distant. He continues to financially support Pivotal, a startup working on electric vertical-take-off and -landing vehicles.

Got more insight to share? You can reach the reporter Hugh Langley via the encrypted messaging app Signal (+1-628-228-1836) or email ([email protected]).

• A few months ago, I started using an AI notetaking app, Granola, in meetings.
• I take notes and then after the call, the AI builds a more fulsome outline of the conversation.
• Taking notes on what's most important helps us get more from meetings, Granola's CEO said.

A few minutes after I'd hopped on a call with a tech founder, he mentioned that he'd started using an "amazing" AI notetaking app.

It was helping him capture the various decisions and to-do's that came up in the many meetings that punctuated his calendar.

I was intrigued. I'd tried artificial intelligence tools for summarizing interview notes and transcripts. The results were often great at capturing themes, yet the AI tended to sweep past the details, pithy comments, or intriguing ideas I would tend to highlight.

It was like getting a book report from someone who'd only skimmed the reading.

Not long after my call with the tech founder, I downloaded the app, which is called Granola, on my Mac. It's a desktop tool, for now. An iOS version is on its way and Windows after that.

I've been using Granola since midsummer, and it's changed my meetings. To be clear, I also use a different app to get a full recording of the call to ensure my reporting and quotes are accurate. But what delighted the founder who tipped me off to Granola is also what I like best: I get to shape the outline for the notes that the AI generates.

My kind of notes

When I began using it, I allowed Granola to synch with my calendar. A few minutes before, I get a prompt to join a meeting. When the call begins, I then get permission from whomever I'm talking with to record the conversation. (Granola also has a prompt that pops up at the bottom that reminds users to get the OK to transcribe calls.)

The notetaking window in Granola is pretty much a blank page, which I like because it's a clean UX. I can drop in a title or use the one populated by what's on my calendar.

Once things begin, I only type what's most important, and the AI follows my lead. I can type just a few words and know that, after the call, with a click, Granola will build an outline around the points I flagged.

That's a huge help and different from the summaries I often get from other AI tools. Plus, I also always look back at the untidy notes I took in case something in the AI version feels off.

If I take no notes at all — which is rare — Granola will still deliver a pretty sharp summary complete with subheads and bullets.

The biggest benefit for me is that I worry less about scribbling down each thing that I might later deem important. In essence, I can be more present.

That's a frequent comment from users, Chris Pedregal, Granola's CEO, told me over a call in which we each took notes with the app.

In fact, given the whac-a-mole way many of us work — quickly triaging the messages that bombard us throughout the day — AI notetaking apps could have our back.

Pedregal said he was surprised when the company began hearing from users that they'll often zone out during a meeting to respond to an urgent Slack or WhatsApp message, then go back to Granola and pop up the transcript to read what they missed.

That's notable, in part, because in a recent survey, 57% of Granola users reported being in leadership roles. Pedregal said that supports the narrative that many top execs might be more excited about AI than some rank-and-file workers.

Pedregal, 38, cofounded Granola in March 2023. He's from the US, though he and the company's staff are based in London. Granola is focused on the American market and has US investors, he said. The company recently completed a $20 million Series A round. Google acquired Pedregal's prior startup, Socratic, in 2018.

Finding the sweet spot

The benefit of having an AI notetaker, I've found, is more than knowing I don't have to worry as much about details in the moment (though I'll always double-check afterward). Pedregal said the reason the app doesn't record audio is to make it less invasive.

The things I type are often the points that stand out because they're unique — or questionable — and that I want to think or ask more about.

Pedregal says jotting down a few notes during a meeting — but not being slavish about capturing everything — is the sweet spot. Unless we're trying to multitask, that middle path often enough, he said, to keep us tethered to the conversation and engaged with what speakers are saying.

I admit I've felt good while in meetings on busy days knowing that the safety net is there.