Interview, Fireside Chat
Turing CEO Jonathan Siddharth: Who Wins in Data Labelling & Why 99% of Knowledge Work Will Disappear
- The industry is transitioning from data labeling firms and traditional SaaS to research accelerators focused on generating complex, real-world workflows for training agentic systems capable of executing economically valuable tasks like legal work or software development.
- Automation of all knowledge work involving computer interaction is projected to occur within approximately one decade, shifting the primary software interface from human-centric GUIs on phones to agent-centric tool calls and multimodal inputs via wearable devices.
- While front-office adoption in financial services, life sciences, and pharma is expected to accelerate, back-office automation may face slower growth due to change management constraints, though budget shifts from human labor to AI have already begun in sectors like customer support and SEO.
- The market for vertically focused data acquisition is currently in "innings one" with significant growth potential, and future competitive moats will rely on data-driven feedback loops rather than technology access, potentially with only a few winners in the 10-year data provisioning landscape.
- A "slow, steady takeoff" of AGI is anticipated rather than a rapid one, a pace believed to allow time for workforce upskilling and educational reform, with superintelligence expected to make human labor 100x more productive and reduce the cost of AI access to roughly $20 per month.
- Entrepreneurship will be democratized as non-technical founders can leverage specialized AI agents (e.g., marketing, engineering, product management) to build companies without significant capital, while the definition of a software engineer will expand to include non-technical individuals.
- Incumbents failing to adopt new tools due to data or permissioning hurdles face a decline over the next 10 to 20 years, whereas companies with "research DNA" capable of rapid adaptation will succeed in the shifting AI ecosystem.
- Financial services clients may lag approximately two years behind the state of the art in AI capabilities within two years, while approximately 50% of the time, top models already produce work indistinguishable from human experts on specific tasks.
- Robotics and embodied AI represent a significant, untapped opportunity requiring vast data resources that do not currently exist, alongside breakthroughs in AI-driven drug discovery and research automation.
- Human activity will shift toward solving problems at higher levels of abstraction, such as curing diseases, reversing aging, and space travel, rather than leisure, driven by a belief that frontier labs are responsible and safety-conscious in model deployment.