Picsart Launches AI 'Personas' for Creators

## Character Consistency: The Core Technical Challenge At the heart of Picsart's Persona and Storyline tools is the solution to a significant challenge in generative AI: character consistency. Previously, generating the same character across different scenes was difficult because AI models had no persistent memory, starting each creation from a state of randomness. The breakthrough lies in using a reference image to create a mathematical "fingerprint" or identity embedding of a character, which the AI then uses as a guide for new image and video generation. This technique, often utilizing methods like LoRA (Low-Rank Adaptation) for fine-tuning, allows a creator to "lock" a character's appearance, including facial structure and clothing, for use in multiple scenes. Picsart's Storyline feature streamlines this by allowing a user to upload a reference image, which then preserves the character's key features throughout a video narrative. While Picsart hasn't detailed its specific underlying models, it does state it uses a multi-model approach, integrating various powerful AI models to handle these complex creative tasks. ## The Rise of the "Faceless" Creator and AI-Powered Side Hustles These new AI tools are directly targeting the rapidly growing trend of "faceless" content creators on platforms like YouTube and TikTok. This approach allows creators to maintain privacy and scale content production by focusing on storytelling and information rather than their personal brand. For software engineers and indie hackers, this opens up new avenues for side projects without the need to be on camera. Successful faceless YouTube channels are already generating significant revenue, with some earning thousands of dollars a month through ad revenue, affiliate marketing, and digital products. For example, the horror storytelling channel Mr. Nightmare has amassed over 6.8 million subscribers and earns an estimated $21,000 to $63,000 per month. These creators often use a combination of AI tools for scriptwriting, voice-over generation, and creating visuals from stock footage or animations. ## Building a Business with AI Avatars The accessibility of AI avatar creation is enabling solo entrepreneurs to build and launch businesses with unprecedented speed. One software developer, Daniel, built an AI-powered chatbot and CRM platform as a side project, which grew to a $100,000/month business, by using AI tools to replace the need for a team. Another indie hacker, Ramsri, built two AI side projects that generate a combined $100,000 a year while maintaining a full-time job, working on them for only one to two hours a week. He emphasizes building distribution by sharing his journey and the problems he's solving in public. These examples highlight a key strategy for engineers: identifying a niche problem and using AI to build a scalable solution. The "faceless" approach, powered by tools like Picsart's Persona, allows builders to focus on the product and system rather than personal branding. This aligns well with a developer's skillset, turning a technical side project into a potentially profitable venture. ## The NYC AI Startup Scene: Funding and Opportunities For those looking to join or build a startup in New York City, the AI scene is active, with a notable focus on enterprise and vertical SaaS applications. In 2025, AI companies accounted for 22% of the $42.3 billion in venture capital deals in the city. VCs in NYC tend to favor startups with clear revenue models and real customer traction. Several VC firms are actively investing in AI-driven consumer and social apps. Greycroft, a firm with offices in New York and Los Angeles, has a stated interest in "intelligent enterprise and consumer applications" and believes AI will dominate the software category. BoxGroup is another early-stage fund that focuses on technology companies, including those in the consumer and prosumer space. For engineers considering the startup path, these firms are key players in the local ecosystem. While Picsart currently lists a remote hub in New York, specific engineering roles for that location are not detailed, suggesting a distributed team model.