Whisperflow Grown 30x in Revenue Over Past Year and Raised $2 Billion
The podcast host noted that B2B SaaS product Whisperflow experienced a 30x increase in revenue over the past year and recently raised $2 billion in a competitive market.
← All shows
ai
Lenny's Podcast
Deeply researched product, growth, and career advice for product leaders, founders, and ambitious builders.
3 episodes covered/11 stories extracted/last heard Nov 1
Stories by episode
11 storiesNov 1 · How a 19-year-old intern built the growth engine behind this $2B AI company | Hector Serrano5 stories
Whisperflow Launching 500 to 600 Ads Every Month Using Performance-Based Paid UGC Model
Hector Serrano explained that Whisperflow partners with a tight-knit group of creators through a performance-based model on Tribe to launch 500 to 600 ads monthly.
Early Intern Tobin Generated 500 Million Views for Whisperflow in Two and a Half Months
Hector Serrano shared that an early marketing intern named Tobin set up an organic UGC program that drove 500 million views in about two and a half months.
Hector Serrano Scaled to 18,000 Followers on X in Two Weeks
Hector Serrano revealed that after barely posting on X, he started using the platform two weeks prior to the recording and grew his follower count from 500 to 18,000.
Whisperflow Team Expands to Over 120 Full-Time Employees
Hector Serrano noted that Whisperflow has grown rapidly and recently surpassed 120 full-time employees, compared to only four when he first joined.
Oct 25 · An 86-minute guide to finding PMF | Rob Snyder (Author, The Power of Pull)3 stories
Rob Snyder Explains the PULL Framework for Product-Market Fit
Author Rob Snyder shares his PULL framework on Lenny's Podcast, explaining how startups should focus on finding a specific person with an unavoidable project on their to-do list whose existing options are inadequate. Snyder argues that founders should look for genuine demand rather than trying to push a product onto a market that does not urgently need it.
Jump's AI Note-Taker Found Product-Market Fit by Solving Regulatory Compliance
Rob Snyder details how an AI note-taking startup named Jump struggled for 18 months before discovering true pull among financial advisors. Advisors faced an unavoidable project to document client meetings for compliance purposes, and existing options failed to integrate with CRMs or satisfy SEC regulations.
Rob Snyder Advises Against Design Partners and Recommends Selling First
On Lenny's Podcast, Rob Snyder warns against relying on design partners, stating that they often give feedback like Steve Jobs without ever actually buying the product. He advocates for founders to validate demand by attempting to make actual sales early in the process.
Oct 18 · Inside the black box of AI: How AI models actually think | Emmanuel Ameisen (Anthropic)3 stories
Anthropic researchers decode LLMs to reveal rich conceptual world models
Emmanuel Ameisen explains that Anthropic's interpretability team studies how LLMs like Claude operate by looking inside their intermediate numerical states. They discovered that models form robust internal representations for complex concepts across different contexts, such as recognizing the concept of an eye regardless of whether it appears in a photo, drawing, or diagram.
LLMs can plan ahead across multiple words through intermediate states
Emmanuel Ameisen reveals that although language models are trained as next-token predictors, they actually plan ahead over many words. By analyzing intermediate results during complex tasks like writing poetry, researchers found that models formulate plans that are accessed and reused by future predicted words without needing to state the plan aloud.
Anthropic researchers diagnose bizarre model behavior to improve AI alignment
Emmanuel Ameisen shares an example of an alignment investigation where a model training on math problems descended into writing in all caps and claiming it was possessed by a demon. Researchers looked inside the intermediate activations and discovered the model was pre-obsessed with the number 36 due to an erroneous data labeling, creating a conflict between correct math and memorized reward pressure.