The Full Picture of Thinking Machines Lab, the Company Founded by Former OpenAI CTO Mira Murati

Since 2023, technological innovation in artificial intelligence (AI), centered on generative AI, has progressed rapidly, with new developments from industry leaders and startups attracting attention. Among these, the startup "Thinking Machines Lab," newly launched by former OpenAI CTO Mira Murati, is making waves. This article focuses on and explains the background, objectives, and technical and social impact of this emerging company.

1. What is Thinking Machines Lab?

1-1. Overview of the New Startup

Thinking Machines Lab is a startup with the goal of "making AI more widely understandable and adapting it to people's diverse needs and values." It has brought together an impressive team, with former OpenAI CTO Mira Murati serving as CEO, OpenAI co-founder John Schulman as Chief Scientist, and former OpenAI Chief Research Officer Barret Zoph as CTO.

1-2. Stealth Mode Exit and Blog Post

Thinking Machines Lab had previously been operating in "stealth mode" (a highly confidential state), but recently lifted the veil and revealed its mission and policies through an official blog post. The blog post makes the following claims:

By emphasizing that disparities and transparency issues are growing in the AI field, they have indicated a policy of taking a more open and collaborative approach.

2. Main Features and Initiatives

2-1. Focus on Multimodal Technology

Thinking Machines Lab has announced that it will focus on multimodal AI systems that handle multiple media formats (modalities) such as text, images, audio, and video. This is based on the idea that by allowing users to interact with AI in diverse forms beyond just text—such as images and audio—more "human-like" interaction and advanced problem-solving become possible.

According to their blog, the pillars of their work are "prioritizing collaboration with humans and developing systems that can flexibly adapt to any professional field." This is noteworthy in that it is expected to have a wider range of applications compared to existing large language models.

2-2. Applications in Science and Engineering

Thinking Machines Lab has clearly stated that it will "build cutting-edge models that support scientific discovery and engineering breakthroughs." Specifically, a wide range of applications is envisioned, including research and development in specialized fields such as medicine, chemistry, and physics, as well as improving programming efficiency.
This is also a reflection of their stance to actively pursue experimental and advanced research, leveraging the experience of Murati, who led the development of ChatGPT, DALL-E, and Codex during her time at OpenAI.

3. Commitment to AI Safety

3-1. Emphasis on Governance and Transparency

Thinking Machines Lab has set a policy to "share safety best practices and increase transparency across the research community" by publishing code, datasets, and model specifications, while simultaneously building mechanisms to prevent the misuse of cutting-edge AI models. This can be seen as an open-oriented stance against some top labs where closed development has become the norm.

3-2. Risk Countermeasures from the Perspective of Social Implementation

The blog post also uses the expression, "The truly important breakthroughs come from rethinking the goal setting itself." This suggests the necessity of not just optimizing existing metrics, but also reconsidering the values that society and users need, and facing the risks hidden within them.

Specifically, they state that they will improve safety measures based on collaboration with users and feedback during the operational phase to ensure that the developed models do not autonomously spread misinformation or increase the risk of privacy violations.

4. Mira Murati's Background and Influence

4-1. Achievements at OpenAI

Mira Murati joined OpenAI in 2018, serving as VP of Applied AI and Partnerships before being promoted to CTO in 2022. She led the development of OpenAI's flagship products, including ChatGPT, DALL-E, and the programming assistant AI Codex. All of these technologies garnered significant public attention and accelerated the social implementation of AI.

4-2. Experience at Tesla and Leap Motion

Prior to that, she was involved in product development at Tesla as a Senior Product Manager for the Model X, and was also involved in the initial release of Autopilot, a driver assistance feature that serves as a point of intersection between electric vehicles and AI. Additionally, at Leap Motion, she served as VP of Product and Engineering, where she worked on the development of motion sensor technology that detects hand and finger movements.

Through this practical experience, Murati is well-versed in the development of complex products where hardware and software work closely together, and it is highly likely that this knowledge will be reflected in the development of multimodal AI at Thinking Machines Lab.

5. Trends in Talent and Fundraising

5-1. Joining of Former OpenAI Researchers

Thinking Machines Lab currently has at least 29 AI researchers and engineers on staff, including members who have gained experience at major companies such as OpenAI, Character AI, and Google DeepMind. The reason for such top-tier industry talent gathering here is likely due to high expectations for Murati's leadership and vision.

5-2. Potential for Large-Scale Fundraising

In recent years, investment capital for cutting-edge AI startups has increased rapidly, and there are rumors that Thinking Machines Lab may be aiming to raise over $100 million. Although no specific amounts or investor names were disclosed in the blog post, it is highly likely that major VCs and strategic investors who resonate with their approach to multimodal AI technology and safety will step forward.

6. The Trend of Startups by Former OpenAI Members

6-1. Comparison with Anthropic and Safe Superintelligence

Startups founded by members who have left or graduated from OpenAI include Anthropic and Safe Superintelligence, which have already established a certain position in the industry. While Thinking Machines Lab shares their mission of 'advancing large-scale AI models while minimizing risks to society,' it further strengthens its unique identity in terms of 'introducing multimodal technology' and 'sharing open knowledge.'

6-2. Balance Between Competition and Cooperation

The fact that talent leaving OpenAI is launching new companies one after another in this way has the potential to promote competition in the AI industry while creating a virtuous cycle that generates new innovation. On the other hand, there are concerns about the oligopoly of major IT companies centered in the US and the uneven distribution of research, so 'to what extent collaboration with the open research community can be realized' will be the key to the future.

7. Future Outlook and Challenges

7-1. Balancing Advancement and Democratization

Thinking Machines Lab's philosophy is, at first glance, aimed at simultaneously achieving the conflicting goals of 'cutting-edge frontier AI' and 'democratized AI that anyone can use according to their own values.' How the company tackles this difficult challenge will be its major undertaking.

7-2. Social Acceptance and Regulation

While the social implementation of AI technology is accelerating, complex issues such as privacy, ethics, and legal liability are surfacing.Thinking Machines Lab is expected to take initiatives to increase social acceptance while collaborating with government agencies and international research institutions, given its proactive stance on model safety and openness.

Led by former OpenAI CTO Mira Murati, Thinking Machines Lab focuses on multimodal AI, safety, and the open sharing of research results, aiming to achieve both "peak capabilities" and "usability and explainability" in AI.
Like other startups founded by former OpenAI members, it is highly likely to have a significant impact on the industry and society as a leader in the new trends of the large language model era. However, many challenges lie ahead, including technical breakthroughs, social acceptance, and balancing development with appropriate regulation. The path taken by Thinking Machines Lab will surely serve as an important indicator for predicting the future of the entire AI industry.

Related Articles