Success Story with Scott D. Clary

Vishal Virani - Rocket.new Founder | The Future Of AI Software Development Belongs to Non-Coders

November 23, 2025

Key Takeaways Copied to clipboard!

  • Founders building disruptive technology, especially in new fields like AI, must prioritize product quality and comprehensive solutions (Day 0 to Day 2) over simply optimizing for speed, as demonstrated by Rocket's 25-minute generation time versus competitors' sub-three-minute outputs. 
  • Success in entrepreneurship, particularly when building from non-traditional tech hubs like Surat, relies on using one's product as the ultimate speaking point to overcome skepticism from investors and stakeholders regarding background or location. 
  • Founders must adopt a 'first principles' thinking approach—breaking down concepts to core fundamentals to invent proprietary methods—rather than merely following existing methods, which is crucial for navigating rapid technological shifts like the one brought by LLMs. 
  • AI will not replace people who adopt it; instead, roles will be redefined (e.g., from 'developer' to 'AI developer'), and mediocre work will be automated. 
  • Hallucinations in AI are often caused by vague requirements or prompts, and users must be highly specific to achieve high accuracy, recognizing that AI, like humans, is not 100% perfect. 
  • Sustainable business in the age of LLMs requires founders to maintain high gross margins (ideally 80% for SaaS) to survive the cost of AI inference, rather than relying solely on continuous funding rounds. 

Segments

Building and Courage in Tech
Copied to clipboard!
(00:00:00)
  • Key Takeaway: Challenging the status quo requires courage, and success in that attempt can significantly impact many lives.
  • Summary: When attempting something new, initial challenges are inevitable, requiring the courage to make the attempt. A successful venture can lead to impacting numerous lives. Before Rocket, Vishal Virani co-founded DhiWise, a Figma-to-code startup that faced difficulty convincing people machines could generate human-acceptable code.
Surat’s Craft and Innovation
Copied to clipboard!
(00:01:23)
  • Key Takeaway: Surat, the world’s largest diamond manufacturing city, fosters a culture rooted in craft, polishing, and handling precious goods.
  • Summary: Surat is the world’s largest diamond manufacturing hub, where nine out of ten diamonds worn globally are processed. This environment instilled a focus on craft and meticulous handling of valuable items. Vishal Virani’s drive toward technology stemmed from a greed to build something significant in the tech world, coming from a farming background as the first engineer in his family.
Strategy of Building in Surat
Copied to clipboard!
(00:05:44)
  • Key Takeaway: Building in a non-traditional tech hub like Surat can be a strategic advantage by allowing a founder to become a ‘mini-Google of Surat’ and attract focused local talent.
  • Summary: The decision to build in Surat was philosophical, aiming to inspire founders that location is secondary to will and curiosity. Strategically, it avoids direct competition with established tech hubs, allowing the company to attract sharp local talent who might not otherwise move. The product itself must speak on behalf of the founders when facing investor friction due to location.
Overcoming Investor Doubt
Copied to clipboard!
(00:15:37)
  • Key Takeaway: When lacking traditional credentials (ex-FAANG, elite college), the product’s capability and demonstrable user traction must serve as the primary strength to build conviction with stakeholders.
  • Summary: Founders often face doubt based on educational background, previous company affiliation, or location; if these are lacking, the product must speak for itself. Investors who tried the product first were impressed by its world-class quality, leading to inbound interest and securing funding. The key is to build something that proves capability, even if it means facing initial rejections.
Pivot from DhiWise to Rocket
Copied to clipboard!
(00:18:16)
  • Key Takeaway: The pivot from DhiWise to Rocket was driven by the conviction that the advent of LLMs represented an ‘iPhone moment’ demanding a larger, future-ready vision.
  • Summary: DhiWise, a Figma-to-code startup, was successful but became obsolete as LLMs emerged, signaling a fundamental shift in software development. Founders must think beyond five years, viewing early AI releases like GPT-3 as merely ‘iPhone version 3.’ Choosing conviction over the comfort of a steadily growing business led to the pivot toward Rocket’s broader vision.
First Principles Thinking Explained
Copied to clipboard!
(00:25:28)
  • Key Takeaway: First principles thinking involves deconstructing decisions into core fundamentals to invent one’s own methodologies, rather than just adopting external methods.
  • Summary: Elon Musk’s philosophy inspired the adoption of first principles thinking, which means getting into the granular detail of every decision. For example, understanding the fundamental behavior of LLMs allows one to invent unique prompting techniques rather than following established ones. Focusing on principles enables invention when methods fail, whereas focusing only on methods leads to stagnation.
Consumption Versus Application
Copied to clipboard!
(00:31:36)
  • Key Takeaway: Accumulating facts through information consumption without testing and applying core principles to one’s specific situation leads to ineffective implementation.
  • Summary: Many founders collect ideas from books without understanding the underlying context or testing the concepts in their unique environment. Implementing a framework like OKRs without analyzing the specific talent and circumstances under which it succeeded elsewhere often leads to failure. Founders must deeply analyze their situation and modify external concepts accordingly.
Rocket’s Vibe Solutioning Platform
Copied to clipboard!
(00:35:39)
  • Key Takeaway: Rocket solves the ‘Day Zero’ problem by generating complete, multi-screen application concepts and code based on a short prompt within 25 minutes.
  • Summary: Rocket is a vibe solutioning platform that analyzes an app idea, geography, and demographics to generate a complete concept, including 8-10 screens of design and necessary components, in about 25 minutes. It differentiates itself by reducing the cognitive load on users, offering solutions via slash commands (e.g., /authentication) instead of requiring complex prompt engineering. The long-term vision is to solve Day 0 (ideation), Day 1 (building), and Day 2 (optimization/scaling) problems.
Speed vs. Depth in AI Tools
Copied to clipboard!
(00:50:25)
  • Key Takeaway: Rocket consciously prioritizes comprehensive output quality over raw speed, delivering an 8-hour equivalent of work in 25 minutes, which users value more than faster, less complete results.
  • Summary: The 25-minute generation time was a calculated risk, validated by data showing users waited because the output was comprehensive, solving problems across multiple screens. While competitors focus on 3-minute generation, Rocket aims to reduce eight hours of manual prompting/iteration work into one comprehensive first result. The philosophy is to focus on output quality and solving the user’s problem completely, even while engineering efforts reduce the time taken.
Adopting AI Requires Restructuring
Copied to clipboard!
(00:57:36)
  • Key Takeaway: Effective AI adoption requires restructuring existing workflows, processes, and teams, rather than attempting to fit AI tools into legacy systems.
  • Summary: When adopting new technology like AI, the mindset must shift from fitting the tool into existing workflows to restructuring processes to leverage the tool’s full potential. Just as smartphones changed habits entirely, AI necessitates a fundamental change in how work is done to achieve 10x or 20x productivity gains. Trying to force AI into old workflows will inevitably lead to the perception that the tool is failing.
AI Cognitive Load & Memory
Copied to clipboard!
(00:59:22)
  • Key Takeaway: AI tools like ChatGPT are improving cognitive load reduction through features like memory, allowing context persistence across threads.
  • Summary: Foundational models are improving daily, but user expectations have risen dramatically since initial launches like GPT-3. The introduction of memory features in tools like ChatGPT is a powerful, small advancement, eliminating the need to re-explain personal context in every new thread.
AI Replacing Developers
Copied to clipboard!
(01:04:16)
  • Key Takeaway: AI replaces mundane work, but developers who fail to adapt to AI integration will be replaced by those who redefine their roles.
  • Summary: If an individual does not adapt AI, they will be replaced; roles are restructuring, such as the ‘AI product designer’ handling conceptualization to code delivery. Mediocre work will be automated, requiring developers to become experts in leveraging AI within their specialized fields to remain indispensable.
Future Software Team Structure
Copied to clipboard!
(01:07:14)
  • Key Takeaway: Founders should focus on Go-To-Market (GTM) while AI platforms handle rapid MVP development, reducing initial team size needs.
  • Summary: Currently, serious systems still require developers to validate generated code security, but founders can prioritize GTM by hiring a few people to manage product delivery via platforms like Rocket. The initial risky phase of building an MVP is drastically shortened, allowing solopreneurs to validate ideas quickly before scaling the engineering team.
Developer Role Evolution
Copied to clipboard!
(01:09:44)
  • Key Takeaway: Future developers must shift from traditional front/back-end roles to AI-centric roles or become 10x productive generalists.
  • Summary: The industry will see new roles like AI developers or data engineers focusing on model fine-tuning and contextual systems. Mediocre developers should learn to become AI developers or leverage platforms like Rocket to become highly productive, potentially handling the work of five roles (product, design, testing, DevOps) single-handedly.
Understanding AI Hallucinations
Copied to clipboard!
(01:15:33)
  • Key Takeaway: AI hallucination rates are often inflated by user expectations and vague instructions, mirroring the inherent error rates in human systems.
  • Summary: Human support systems often operate with 30-40% error rates, yet users expect near-perfect accuracy from AI, magnifying perceived hallucinations. The primary cause is vague prompting; users must define boundaries and context precisely to receive non-generic, high-quality results, even while accepting a small residual error rate from foundational models.
AI Prediction One Year Out
Copied to clipboard!
(01:19:36)
  • Key Takeaway: Within a year, affordable beta versions of personal ‘Jarvis’-like assistants will emerge, eliminating most repetitive and abundant work.
  • Summary: The world will see a glimpse of the sophisticated assistance demonstrated by Jarvis in Iron Man movies, made affordable to the masses. This will ensure users no longer need to focus on mediocre or repetitive tasks, as AI will handle them.
Rocket’s Rapid Success Factors
Copied to clipboard!
(01:21:11)
  • Key Takeaway: Rocket’s rapid funding success was built upon four years of prior hard work, illustrating that overnight success is merely the visible tip of the iceberg.
  • Summary: The founder attributes the quick funding to four years of foundational effort preceding the 15 weeks of Rocket’s launch success. Success is defined by achieving the next milestone and ensuring the entire team aligns toward that vision, often requiring intense daily work.
Defining Sustainable Business
Copied to clipboard!
(01:23:36)
  • Key Takeaway: A sustainable business, especially in the LLM era, must maintain high gross margins (like 80% for SaaS) to offset high inference costs.
  • Summary: Sustainability means surviving the customer acquisition cost (KEC) to lifetime value (LTV) duration without external control. With LLM costs being high, single-digit or 15-20% gross margins are risky long-term, necessitating engineering focus to build strong margins.
Entrepreneurial Mental Models
Copied to clipboard!
(01:24:40)
  • Key Takeaway: Founders must control their business mathematics (burn rate, funnels) and connect all dots before making decisions, avoiding vanity metrics.
  • Summary: Founders must understand their financial equations well enough to control the business, making decisions by connecting current, future, and competitor perspectives. While early-stage decisions require visionary boldness without complete data, stable businesses must combine data analysis with vision, staying focused and avoiding distractions.
Advice to 20-Year-Old Self
Copied to clipboard!
(01:28:00)
  • Key Takeaway: Early-stage entrepreneurs should avoid distraction by focusing 100% effort on one core thing aligned with their life goal to prevent wasted time.
  • Summary: The founder realized that trying to do too many things (service company, product, etc.) in the early days led to wasted years. Distraction causes failure in the beginning; picking one focus and dedicating maximum effort is the path to achieving significant scale.