The world with AI among us is becoming all too real.

A milestone technological advancement, machine learning is not yet ready for primetime; data engineers are still training AI to operate autonomously.

But why wait months (or years) for artificial intelligence to get up to speed, when you could hone a computer in a matter of hours?

Massachusetts-based startup Gamalon Machine Intelligence this week unveiled its own solution: Bayesian Program Synthesis (BPS). The technology promises to simplify AI projects, without Google-level resources.

Founded by MIT-trained computer scientist Ben Vigoda (nephew of late actor Abe Vigoda), the company uses probability models to teach machines to identify something like a dog with just a few samples.

Most deep-learning systems require data scientists to input millions of canine images so the computer can learn every possible variation. Gamalon, however, uses probabilistic programming, allowing engineers to present a picture of a dog with short ears and one with long ears, and let the computer fill in the gaps.

This type of speed-reading technique achieves higher accuracy and uses less computation, according to Vigoda, who boasted to Bloomberg that “You can run our software on a laptop, and it takes 100 times less horsepower to find an answer.”

Gamalon launched in 2013, but only this week announced the alpha launch of its first commercial applications, called Structure and Match.

According to Bloomberg, the niche products are used to scan databases and fix ambiguities; for $10,000 per month per 100,000 rows of text, the software will identify different spellings for customer names and addresses.

The services are available now to select customers as APIs integrated with the Amazon, Microsoft, and Google Cloud platforms.

“We have one customer that, every year, spent nine months and $4 million to structure and match their data,” Vigoda said in a statement. “In contrast, [we were] able to perform the same task in minutes with twice the accuracy.”