Slightly Sharpe
Published on

AI, Tincre and Competition

  • avatar
    Jason R. Stevens, CFA
The logo atop a skyscraper at night in neon purple, generated by OpenAI's ChatGPT 4.

AI isn't a catch all.

Aside from all the techno-babble from theorists that want your internet time to sell you ads (or their content), AI is still pretty much just massive-scale machine learning. That means it consists of three basic branches of statistics and computer science, combined into learning schemes:

  1. supervised,
  2. unsupervised, and
  3. feedback-based (what most refer to as reinforcement learning techniques).

What most software users experience today is some combination of all three. That's exactly how we leverage AI/ML within Tincre software.

Been there, done that

We are at the forefront of implementation with new AI tools. Tincre’s technical Co-founder, your author, has been using OpenAI’s products since 2017 (that’s 1000 years in AI-time). Our firm fully integrates LLMs, supervised learning, unsupervised and reinforcement learning across the business for our customers, our internal employees/processes, and our ad content.

LLM features

Our latest client-facing feature is a very detailed in-dashboard agent experience that can answer logic-based questions, in detail, about campaigns, their metadata and performance data for end-users. And the agent can answer those questions in any major language. This agentive chat feature leverages new AI models called LLMs. But at Tincre, LLMs do much more. In particular, LLMs write, produce, recommend content and help us organize the entire process of running ad campaigns.

Here's one example of taking the static image from this post and turning it into a short video.

What's an LLM, by the way?

What most of you refer to today as "AI" is Large Language Models (LLMs), one very narrow and specific type of AI model.

In highly over-simplified terms, an LLM is a deep neural network trained on a bunch of data (mostly text) - sentences, paragraphs, words and phrases - to calculate the distances between those pieces of data. Dig deeper with AWS's explanation in case you're interested.

These LLM's are incredibly useful to us and a boon to the Tincre business. That said, much hearsay throughout the tech industry focuses on increased competition due to LLMs.

Competitive landscape with LLMs

So what do we think about competition due to LLMs and how it impacts Tincre?

Ad platforms are walled-gardens

First, LLMs need lots of data to "learn." And the platforms on which we run ads have nothing static, scrapable, or inherently "learnable" about them. Those platforms - like Meta ads and Google ads - are what we call walled-gardens and aren't privy to LLM training, which any LLM requires to function. Even if they were learnable, those platforms change on a daily basis, which renders a stochastic parrot (LLM) useless.

Spend allocation

Secondly, the way we choose where ad spends should go, called allocation, combines all three branches of "AI" (again, simply statistical or machine learning), which have been used for several decades in finance.

Fortunately, the current AI revolution is merely a boon for our ability to scale and expand this allocation technology faster by making it easier for us to make quality decisions for our client users. That’s because we can test more content and get more information, improving cross-platform ad interactions per amount spent.

Motivated people build

Lastly, our business vision is that the next wave ads platform is like E-Trade was for retail stocks trading in the 1990s. Give motivated people access, tools and insights and they'll enter the market and dominate it. After all, software is still eating the world.

LLMs help Tincre achieve this vision faster and better, that’s it.

On competition and the future

Our firm’s competition is incumbent: our main user - a small-mid-sized company's ad manager and/or team - is incentivized not to use our integration platforms, similar to how taxi drivers have no incentive to use Uber, as it basically puts them out of business.

What the most productive have realized, including former Taxi drivers, however, is that leveraging new technologies helps them produce more. In our case, that means the ad manager or team adds more value to their employer or business.

And that's what we see as Tincre's core role: provide ads people the ability to add that value to their employer or business in the easiest and most cost-effective manner possible.

Where we're going

Tincre’s first integration partners hit their one-year mark as of Nov. 30th. We simply abstracted what you know as, and so that other businesses could spin up online digital ads agencies.

The first integration generated over 150k in revenue. There are 20k companies in the US alone similar to them. If you think the digital ads business is big today, just wait for tomorrow with Tincre.

Press button, get digital ads agency

Those integrations are our business focus. will soon become the web’s home for cross-platform ads. You can run ads, you can spin up an agency just like, or you can learn about online ads from every angle.

So what's a digital ad, anyways?

All that said, ads are the business of the internet. They're what we call the web's last mile.

Every word you read, every video, sound, and piece of content you consume, is an ad. Every platform you use exists because it can sell you "information/content" (ads).

There will be (and are) many, many valuable firms within this ecosystem and we're one of them. And the beautiful AI revolution makes content people want cheaper to produce, better in quality and abundant.

That's incredible for our client users and our firm, because that content is information, which is what gives us more data sources. In other words, just like a portfolio of financial assets - stocks, bonds and real estate - with more information each dollar marketed through our platform takes less risk and earns more return for what we sell - the world's best and easiest turnkey cross-platform ads.

The Rap 🆙

If you have any comments or questions we'd love to hear from you. Email Jason directly.

And of course don't forget to sign up to the newsletter, if you haven't already.

Subscribe to the newsletter
👋 We use cookies to enhance your experience. In using this site you agree to the storing of cookies on your device.