The Twenty Minute VC

Turing CEO Jonathan Siddharth: Who Wins in Data Labelling & Why 99% of Knowledge Work Will Disappear

Jonathan Siddharth is Founder and CEO of Turing, one of the fastest-growing AI companies advancing frontier models. Jonathan has led the company to an astonishing $300M ARR with just $225M raised and a profitable company. A Stanford-trained AI scientist, Jonathan previously helped pioneer natural language search at Powerset, which was acquired by Microsoft. ----------------------------------------------- Timestamps: 00:00 Intro 00:51 Redefining “Talent Marketplaces” Today 03:46 Data, Compute, Algorithms: What is Most Abundant? 16:59 The Biggest Challenges Enterprises Have with AI Adoption 20:57 Why Will 99% of Knowledge Work Will be Gone in 10 Years 28:53 How Will Data-Driven Feedback Loops Replace Technology as the Moat 34:20 Is Revenue BS in Data Labelling? Are Players Calling GMV Revenue? 43:43 Are We in an AI Bubble? 52:22 Why is SaaS Dead in a World of AI? 01:00:32 Will the Phone be the Primary User Interface to an AI World? 01:07:46 Quick-Fire Round ----------------------------------------------- Subscribe on Spotify: https://open.spotify.com/show/3j2KMcZTtgTNBKwtZBMHvl?si=85bc9196860e4466 Subscribe on Apple Podcasts: https://podcasts.apple.com/us/podcast/the-twenty-minute-vc-20vc-venture-capital-startup/id958230465 Follow Harry Stebbings on X: https://twitter.com/HarryStebbings Follow Jonathan Siddharth on X: https://twitter.com/jonsidd Follow 20VC on Instagram: https://www.instagram.com/20vchq Follow 20VC on TikTok: https://www.tiktok.com/@20vc_tok Visit our Website: https://www.20vc.com Subscribe to our Newsletter: https://www.thetwentyminutevc.com/contact ----------------------------------------------- #20vc #harrystebbings #jonathansiddharth #turing #datalebelling #ai #data #saas

Jonathan SiddharthguestHarry Stebbingshost

Nov 30, 20251h 17mWatch on YouTube ↗

WHAT IT’S REALLY ABOUT

Turing CEO predicts AI agents will automate nearly all knowledge work

Jonathan Siddharth, CEO of Turing, argues that traditional data-labeling firms are obsolete and being replaced by "research accelerators" that build complex reinforcement-learning (RL) environments to train agentic AI systems. He describes how Turing powers the “data pillar” for 7 of 8 frontier labs, creating synthetic but realistic workflows across industries so models can learn to perform economically valuable, multi-step knowledge work. Siddharth predicts that virtually all digital knowledge work—about $30 trillion worth—will be automated over time, with slow but steady AI capability takeoff and a massive shift of budget from human labor to AI systems. He also foresees the decline of classic SaaS, the rise of custom fine-tuned small models inside enterprises and governments, and a future where individuals are 100x more productive through fleets of agentic AI “exoskeletons.”

IDEAS WORTH REMEMBERING

5 ideas

Data-labeling is giving way to research accelerators that build RL environments.

Turing no longer focuses on simple annotations but on constructing rich reinforcement-learning ‘mini-worlds’ that mimic real workflows across industries, enabling models to learn tool use, multi-step reasoning, and agentic behavior.

AI training data needs have shifted from simple tasks to complex, domain‑expert workflows.

As models become smarter, the marginal value comes from high-skill, vertically specific data (e.g., complex coding tasks, legal workflows, underwriting), which requires expert humans and realistic environments, not commodity labelers.

Custom small models fine-tuned on proprietary data will be a durable enterprise pattern.

Siddharth argues many enterprise use cases (e.g., insurance underwriting) are best served by smaller on-prem models fine-tuned on decades of internal decisions, offering better accuracy, speed, and data control than trillion-parameter general models.

Nearly all digital knowledge work will be automated, but via slow takeoff.

He believes any job done on a computer using tools, keyboard, and mouse will be automated over time, yet adoption will be gradual, especially in back-office functions, giving society and enterprises time to adapt and redesign workflows.

Moats will come from data-driven feedback loops and deployment, not raw tech.

Similar to Google’s search advantage, winners will continuously collect real usage data, see where models fail in production, and use that feedback to generate targeted new training data, creating compounding performance gaps.

WORDS WORTH SAVING

5 quotes

I think the era of data labeling companies is over. It’s now the era of research accelerators.

— Jonathan Siddharth

All knowledge work is going to be automated. It’s only a matter of time.

— Jonathan Siddharth

SaaS, as we know it, I think is over.

— Jonathan Siddharth

There is a very significant model capability overhang… the models are capable of X, but what we are getting out of the models is X minus delta.

— Jonathan Siddharth

Whoever wins the superintelligence race will probably win search, consumer devices, operating systems, and cloud. You’re playing for everything.

— Jonathan Siddharth

Shift from data-labeling companies to research accelerators and RL environmentsTraining agentic AI systems to perform real-world, multi-step knowledge workCustom fine‑tuned small models for enterprises and future government/sovereign modelsAutomation of $30T of digital knowledge work and labor-to-AI budget transferFuture of SaaS, software engineering, and moats in an AGI-centric worldAI adoption dynamics in incumbents vs startups; front-office vs back-office useLong-term AI trajectory: slow takeoff, superintelligence, new interfaces and devices

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.