Turning the unthinkable Thinkable, thanks to this Miami AI startup

By Riley Kaminer

In a sea of startups building glorified ChatGPT wrappers and marketing teams slapping AI on any product splash page they get their hands on, it can feel like this artificial intelligence moment is more hype than substance.

But Miami startup Thinkable will make you think again.

The company is developing a patent-pending technology that makes it easy for developers to harness the full potential of Large Language Models (LLMs) to solve enterprise-grade problems and execute complex, domain-specific tasks ranging from developing apps to building neural networks. 

“LLMs are very good at predicting the correct response to an input prompt, they are very bad at planning and executing complex tasks,” co-founder and CEO Eli Blatt told Refresh Miami.

“Solving complex problems using LLMs requires a Generative AI Application to chain together prompts and executable operations able to plan and execute that app’s target task, all of which have to be manually engineered,” he explained. “This makes solving enterprise-grade tasks using LLMs difficult, expensive, and time-consuming.”

Using this core no-code technology, the Thinkable team has already forged partnerships to develop some advanced AI apps that are used by 5,600 people collectively. One of them is PlotDot, an AI-powered platform that helps users create scripts.

“If you ask a language model how to write a screenplay, you might receive a simple list of steps,” Blatt explained. “But that wouldn’t be sufficient to create a screenplay. You’d need a detailed series of prompts, starting with defining your main character and progressing to the antagonist – a character arc, a hero’s journey. The challenge lies in effectively linking these prompts, a task that grows more complex with each additional layer.”

Other projects in the works include PatentDraft, a legaltech product co-founded with a lawyer and exited founder of a Peter Thiel backed startup, and Instagrant, an AI-powered grant writing tool.

Thinkable’s business model will be threefold: SaaS revenue for building and hosting LLM apps, joint ventures to develop Thinkable-built apps alongside strategic partners, and revenue stemming from licensing patented technology.

Blatt launched the platform alongside co-founders Matt Sanders (COO, former Deloitte leader), Richard Pocklington (CPO, Stanford PhD like Blatt, who has filed hundreds of patents and sold a patent portfolio to Google for eight figures), and Reda Khyatti (CTO, formerly CryptoKitties/Dapper Labs) in 2022.

David S. Rose – the founder of NYC Angels and Gust, and Associate co-founder of Singularity University – led Thinkable’s pre-seed round. Thinkable is now raising a seed round.

The company’s ultimate goal? Become the Shopify for AI development, where Thinkable is to AI development what Shopify is to web development. As Shopify earns revenue from building ecommerce sites and their ongoing operation, so will Thinkable earn revenue from the products that are built on their platform.

“We are in an innovation boom that has not really ever been seen before,” said Blatt. “We are genuinely an AI company, and our work centers around a discovery that is now pending patent approval. This discovery is fundamental to our operations and the value we provide, and I believe it will become a foundational AI and ML technology.”

Photo at top of post: Three of Thinkable’s cofounders exhibited at SXSW this year: left to right,COO Matt Sanders, CEO Eli Blatt, and CPO Richard Pocklington.

READ MORE IN REFRESH MIAMI:

Riley Kaminer