Reid Hoffman and Mark Pincus Are Building an AI Lab Focused on Task Automation — and Raising $100M to Do It

Prentis, a new AI lab co-founded by LinkedIn co-founder Reid Hoffman and Zynga founder Mark Pincus, is in talks to raise $100 million. The startup is betting that automating everyday computer tasks will become a bigger AI use case than code generation. Here's what we know about their thesis.

Reid Hoffman and Mark Pincus — the co-founder of LinkedIn and the founder of Zynga, respectively — are joining forces on a new AI venture called Prentis. According to TechCrunch, the lab is currently in talks to raise $100 million in early funding, signaling significant investor appetite for what the founders are calling a "neolab" model.

The Core Bet: Task Automation Over Coding

The central thesis behind Prentis is a pointed one: that automating routine computer tasks will soon surpass code generation as AI's most commercially significant use case. This is a deliberate pivot away from the current conversation, which has been dominated by coding assistants like GitHub Copilot, Cursor, and a wave of similar tools.

The argument isn't that coding AI is unimportant — it's that the addressable market for automating repetitive, non-technical desktop and browser-based tasks is larger and less contested. Think the kind of work that occupies large swaths of knowledge workers daily: data entry, form filling, report generation, navigating enterprise software.

Why These Founders, Why Now

Hoffman and Pincus bring complementary credibility. Hoffman has been an active AI investor and thinker — he's a board member at Microsoft and an early backer of OpenAI — while Pincus built Zynga into one of the defining social gaming companies of the early 2010s. Both have deep networks in Silicon Valley and experience scaling consumer-facing products.

The "neolab" framing is notable. It suggests Prentis sees itself not just as a product company, but as a research-forward organization with ambitions to define new categories — similar in posture to labs like Adept, Cohere, or Inflection, rather than a pure application-layer startup.

What This Means for Founders and Operators

For startup founders and enterprise buyers, the Prentis announcement is worth watching for a few reasons:

The task automation space is heating up. Competitors including Anthropic (with Claude's computer use capability), OpenAI (Operator), and startups like Relay and Bardeen are all pushing into agentic, task-level automation. A well-funded neolab with Hoffman and Pincus attached raises the stakes.
Funding gravity matters. A $100M raise at the formation stage suggests VCs believe the category is real — which often accelerates enterprise procurement conversations and talent acquisition across the whole sector.
The non-technical user is the new target. If Prentis's thesis proves out, the next wave of AI ROI won't come from developer productivity — it'll come from automating the workflows of operations, finance, HR, and sales teams who never touch code.

The Competitive Landscape

OpenAI's Operator product, launched earlier this year, already demonstrated how much demand exists for AI that can navigate a web browser on a user's behalf. Anthropic's computer use API gave developers programmatic access to similar capabilities. What Prentis appears to be betting on is that neither of these companies — primarily focused on foundation model development — will fully commit to the product and go-to-market depth needed to own this category long term.

That's a reasonable thesis, but it's a crowded race. The $100 million, if closed, buys Prentis runway to find out whether their take on agentic automation can differentiate at a product level — not just a narrative one.