Keep pulling the thread on Jonathan Siddharth.
Turing has over $300 million in revenue, is profitable, and has raised $225 million in funding at a last valuation of $2.2 billion.
Turing works with seven out of the eight frontier AI labs, including OpenAI, Anthropic, Meta, Google, Microsoft, Nvidia, and Amazon.
The internet data used for pre-training large language models was exhausted approximately three years ago.
Jonathan Siddharth predicts that $30 trillion of knowledge work will be automated by advancing AI.
The scaling laws for AI models, which state that performance smoothly improves with more parameters, compute, and data, are continuing to hold true.
Jonathan Siddharth predicts that Turing will eventually replace traditional consulting and services firms like McKinsey, Bain, BCG, Accenture, TCS, Wipro, and Infosys with agentic AI systems.
Jonathan Siddharth believes that achieving Artificial Superintelligence (ASI) requires solving four key pillars: multimodality, reasoning, tool use, and coding.
Jonathan Siddharth defines Artificial Superintelligence (ASI) as the automation of 90% of the tasks performed by 90% of humans working on a computer.
The AI data industry shifted significantly after the release of reasoning models like O1 and DeepSeek, moving from needing simple data to complex, strategically engineered data.
During the training of GPT-3, OpenAI's objective was to teach the model how to code, perform function calling, and use tools, which are foundational for building AI agents.
Jonathan Siddharth predicts that current AI models will be able to overcome today's most difficult agentic coding challenges within approximately six months.
Jonathan Siddharth does not believe a "rapid takeoff" scenario for AI is likely, instead expecting steady, continuous progress.