# The Biological Computing Co. Launches Mischief 1, An AI Video Model Trained On Lessons From Living Neurons

> Source: <https://officechai.com/ai/the-biological-computing-co-launches-mischief-1-an-ai-video-model-trained-on-lessons-from-living-neurons/>
> Published: 2026-10-31 07:49:47+00:00

Newer AI approaches aren’t just innovating with different architectures — they’re innovating on the substrate itself.

San Francisco startup The Biological Computing Co. (TBC) has released its first commercial AI product, an image-to-video model called Mischief 1, fittingly timed to Halloween. CEO and co-founder Alex Ksendzovsky announced the launch on X, calling it a “little mischief” four years in the making.

“For four years, we’ve been brewing real brain cells on electrodes to understand how they compute. Now, we’ve turned those discoveries into a video model anyone can use,” Ksendzovsky wrote. He said Mischief 1 is at least 5x faster than its base model and the cheapest image-to-video model available on both fal and OpenRouter, starting at 0.5 cents a second, with an AWS Marketplace listing coming soon.

## What TBC Actually Does

TBC’s core business isn’t video generation — it’s biological computing. The company grows living neurons, derived from stem cells, on multi-electrode arrays in its lab and studies how those cells naturally process and compute information. Real-world data like text, images, and video gets encoded into the neurons, and the resulting neural activity is decoded back out and mapped onto conventional AI architectures as an optimization layer.

The pitch is that half a billion years of evolution has already solved problems around efficiency and adaptability that silicon is still struggling with, and that borrowing those biological computing principles can make transformer-based models faster and cheaper to run without replacing the underlying hardware. Mischief 1 is the company’s first attempt to put that research into a product anyone can actually use, rather than a research demo.

It’s a similar instinct to the one driving work on direct brain-computer interfaces, where companies are trying to find ways for biological and digital systems to [communicate directly with each other](https://officechai.com/stories/neuralink-telepathy/) — except TBC is going the other way, using biology to make the computing itself more efficient rather than connecting it to a human user.

## The Founders

TBC was founded by Alex Ksendzovsky and Jon Pomeraniec, who serve as CEO and COO/President respectively. Both are trained neurosurgeons rather than typical Silicon Valley software founders — Ksendzovsky and Pomeraniec met as undergraduates at the University of Maryland and the University of Pennsylvania and have maintained a two-decade partnership through medical school, surgical residencies, and academic research roles.

Ksendzovsky has said the idea traces back to his college years, when he was studying philosophy of mind and watched a visiting professor demonstrate a robot powered by a single neuron, which convinced him there was untapped computational power in biological tissue. The pair reportedly made early research progress on the brain’s computing capabilities around 2021 before leaving academic and surgical careers to start the company, initially operating under the name Biological Black Box before rebranding to The Biological Computing Co.

## Funding

TBC emerged from stealth in February 2026 with a $25 million seed round led by Primary Venture Partners, alongside Builders VC, E1 Ventures, Proximity, Refactor Capital, Tusk Venture Partners, and Wonder Ventures. The round coincided with the opening of the company’s flagship wet lab in San Francisco’s Mission Bay neighborhood, where its neuron-electrode chips are grown and maintained.

Primary’s Brian Schechter, who led the investment, has argued that silicon has “carried the field far” but that AI’s next efficiency breakthrough will need to come from alternative architectures like biological computing. Tim Gardner, a co-founder of Neuralink, is among the backers who have publicly endorsed TBC’s direction, saying the most performant compute of the future will likely resemble the brain more than today’s chips do.

TBC has said its neurons are sourced through established, lawful biomedical supply chains, and that it is moving away from rat-derived cells toward neurons grown from reprogrammed human stem cells. Each chip reportedly houses between 100,000 and 500,000 neurons and has a working lifespan of about a year.

## Where This Fits

TBC isn’t alone in trying to commercialize biological computing — Melbourne’s Cortical Labs has pursued a similar living-neuron approach and recently opened what it calls the world’s first biological data centre in partnership with Singapore’s NUS Medicine and data-centre developer DayOne. But TBC appears to be the first to ship a neuron-derived optimization into a publicly available commercial AI product rather than a research platform, with Mischief 1 live on fal and OpenRouter today and an AWS Marketplace listing expected to follow.
