Winning the AI arms race seems like it’s dependent on just two factors — scale and money.
Flush with cash, the biggest AI developers are focused on gathering massive amounts of data, developing compute power, and building ever-larger models.
But that isn’t the only way to get ahead.
“It’s definitely true that if you throw more compute at the model, if you make the model bigger, it’ll get better,” Aidan Gomez, the CEO of Cohere, said in a recent interview with 20VC. “It’s kind of like it’s the most trustworthy way to improve models. It’s also the dumbest.”
He said there are several other ways smaller companies can compete.
“There’s a lot of pressure on making smaller, more efficient models, smarter via data and algorithms, methods, rather than just scaling up due to market forces,” he said.
Gomez said many gains in the open-source space have also come through improvements in handling data. That includes better algorithms to scrape higher-quality data from the internet and innovations in synthetic data.
Cohere, which builds AI technology for enterprises, was valued at $5 billion after its latest funding round in July. Earlier this year, it announced a family of language models called Command R, which its COO Martin Kon previously told Business Insider was part of an “emerging category of scalable models.” As the company continues to provide technology for enterprises, Kon said that affordability and reliability are more important than offering companies the latest in AI technology.
