Prentis, new AI lab co-founded by Reid Hoffman, Mark Pincus in talks to raise $100M

A new AI venture called Prentis is reportedly in discussions to raise roughly $100 million, with Reid Hoffman and Mark Pincus listed as co-founders. The lab's stated thesis is that automating routine computer tasks will eventually surpass coding as the dominant application of artificial intelligence.

The capital chase is the headline, but the more telling detail is the bet itself. Coding assistance has been the proving ground for generative AI over the past three years, absorbing billions in venture funding and producing tools that are already embedded in developer workflows. Prentis is positioning itself for the next layer down: the mundane, repetitive interactions that fill most of a workday, from navigating software interfaces to moving data between systems.

Hoffman and Pincus bring credibility and network access, though neither is a technical founder in the conventional sense. Hoffman's track record with LinkedIn and his early bet on OpenAI give him standing as an AI-adjacent operator. Pincus, best known for Zynga, is a less obvious fit, but his involvement signals that consumer-facing software instincts are part of the mix. The pairing suggests a lab that wants to bridge enterprise automation with the kind of product polish that consumer companies obsess over.

The $100 million target, if achieved, would place Prentis in the upper tier of well-capitalized AI startups at a moment when investor patience for speculative AI labs is thinning. The pitch will need to demonstrate more than demos. Routine task automation is a crowded field, populated by established players like UiPath and a growing roster of AI-native challengers. Prentis will need to show that its approach is either substantially cheaper, substantially more reliable, or substantially more capable than what already exists.

For the broader market, the story is a signal about where smart capital thinks the next wave of AI value will accrue. If two of Silicon Valley's most connected operators are willing to attach their names to a $100 million round for task automation, they are betting that the bottleneck has shifted from raw model capability to the unglamorous work of connecting those models to the software people actually use every day.