Adept, which aims to build AI systems that can understand and automate any software process, emerges from stealth with $65M led by Greylock and Addition
Adept is an ML research and product lab building general intelligence … Ben G / Ben Goldhaber's Newsletter : April 2022 BenGoldhaber(.com) Newsletter Tweets: Scott Belsky / @scottbelsky : developing AI models to use our existing everyday software tools (and underlying APIs) on our behalf is a fascinating approach for near-term productivity gains. https://www.adept.ai/... congrats @jluan and team. https://twitter.com/... Reid Hoffman / @reidhoffman : Adept takes a different path from other AGI companies. Rather than building general intelligence to take over valuable tasks, they're building AI tools that empower humans to get stuff done. David Luan / @jluan : A bunch of top ML folks from Google, DeepMind, OpenAI, etc have come together to build Adept! It's a pleasure to be working with this kind and extremely talent-dense crew, incl. the folks who invented Transformer. We're doing something a bit different... (thread) https://twitter.com/... Saam Motamedi / @saammotamedi : 6/ The next phase of AI will come from building products that people use today - but that also contain the seeds of generality for the future. Adept is doing exactly this, and we're excited to bring Greylock's expertise in enterprise products to bear.
Context & Ripple Effects
Adept came out of stealth in May 2022 as an ML research lab betting that general intelligence should be built to operate existing software tools and APIs on users' behalf — a framing Reid Hoffman publicly distinguished from other AGI companies' approaches. The bet attracted fast capital: within a year the team had followed this $350M Series B at a $1B+ valuation.
What makes the launch worth revisiting is where it ended: Amazon hired Adept's CEO and co-founders into its AGI team while taking some of the company's technology, and David Luan has since departed Amazon's AGI lab less than two years after arriving through that deal. The stealth round is the starting point of an arc that became a case study in how independent agent labs fare against hyperscaler acquirers.
First-order effects
- Greylock and Addition secured an early position in what became one of the most closely watched agent startups, with Adept funded to pursue models that turn text commands into computer actions across everyday software.
Second-order effects
- Adept's rapid rise to a $350M raise validated the workflow-automation category for rivals and investors alike, but its eventual absorption — Amazon hiring the CEO and co-founders and taking some of its tech — showed hyperscalers could outbid the private market for the teams building it.
Third-order effects
- If the pattern holds, 'automate any software process' startups become acquihire targets rather than durable independents, and the talent keeps moving: Luan's departure from Amazon's AGI lab underscores how little stability these reverse-acquihire arrangements guarantee even for founders.
The trend: AI agent labs launched on large VC rounds are increasingly ending up inside hyperscaler AGI teams through acquihires, with founder departures following soon after.