Interaction Models: Mira Murati's New AI

The problem Thinking Machines wants to solve

The AI models available today work in a sequential way. The user types or speaks. The model waits. Then it processes and replies. However functional, this pattern creates a deep break from natural human communication.

Thinking Machines describes this limitation clearly: "Today's models experience reality in a single thread". In other words, the model perceives nothing until the input is complete. It doesn't see the user's hesitation. It doesn't pick up the tone of voice. It doesn't interpret the visual context.

The announcement of interaction models therefore comes from a clear ambition: closing the gap between artificial intelligence and genuine human collaboration. As reported by The Verge, Thinking Machines is aiming for models that "think, respond and act in real time."

What interaction models are: architecture and how they work

Thinking Machines' interaction models are designed to process three input channels at once: audio, video and text. They do it continuously, without waiting for the user to finish the interaction.

This real-time multimodal approach is an architectural leap from traditional large language models. Today's models, even the most advanced, treat every conversational turn as a discrete event. An interaction model, by contrast, keeps an active, persistent awareness of context.

In practical terms, the system can notice when a user hesitates before finishing a sentence. It can read a facial expression during a video call. It can adapt its reply to the emotional tone it detects in real time. In short, it's a model that doesn't just react, it participates.

To dig deeper into multimodal AI and its technical foundations, MIT Technology Review offers sharp analysis of how these systems are evolving.

Who Mira Murati is and why the market is watching her

Mira Murati served as CTO of OpenAI until 2024. In that role, she oversaw the development of GPT-4 and ChatGPT. Her departure from OpenAI drew wide attention across the industry.

The founding of Thinking Machines confirmed that Murati intends to build something structurally different. Not an alternative to ChatGPT, but a new paradigm of human-machine interaction. As a result, the market follows every announcement from the company with interest.

As with other startups founded by former executives of major AI labs, Thinking Machines enjoys instant technical credibility. Still, the distance between an announcement and a commercially mature product remains significant. This is especially true for technologies this ambitious.

According to Gartner analysis, multimodal AI technologies are maturing rapidly. As a result, business adoption times are getting shorter than in previous cycles.

Immediate impact for Italian SMEs: what really changes

For an Italian SME, the practical question is: does this development change anything today? The answer is nuanced.

In the short term, interaction models are not yet available as a commercial product. Thinking Machines has announced a direction, not a launch. Still, the indirect impact is already measurable. Announcements like this speed up the evolution of the entire AI ecosystem, including the platforms already in use.

In particular, companies working in relationship-intensive areas, such as customer service, consultative sales and internal training, should keep an eye on this trajectory. And if you're weighing investments in AI solutions for your business, you need to factor this paradigm shift into your planning.

At SHM Studio, we work every day with SMEs asking how to integrate artificial intelligence into their operations. Understanding where the technology frontier is heading is therefore a core part of our advisory approach.

Work still in progress: limits and unknowns

It's important to stay realistic. Interaction models come with serious technical challenges. Processing audio, video and text simultaneously in real time takes considerable computing power. Latency also has to be low enough to keep the interaction smooth.

There are privacy issues too. A system that continuously perceives a user's visual and audio surroundings raises significant regulatory questions. In Europe, the GDPR and the AI Act impose clear constraints. Enterprise adoption of these technologies will therefore have to deal with the regulatory framework.

Finally, the question of the interface remains open. How do you design a user experience for a system that doesn't wait? How do you handle interruptions or corrections in real time? These interaction design problems are still largely unsolved. For those working in web design and digital interfaces, this is already a timely topic to think about.

What to do now: find your direction without chasing the hype

Faced with a technology announcement of this scale, the most effective response is neither passive waiting nor uncritical, immediate adoption. Instead, it helps to build an informed strategic stance.

First, map the business processes where real-time interaction could create value. For example: technical support sessions, customer onboarding, staff training. Then, assess which tools already available come close to this paradigm and test them in controlled settings.

For SMEs that want to build a coherent digital strategy, digital marketing and SEO services remain fundamental pillars. But AI is increasingly cutting across these areas. Ignoring it means giving up a growing competitive advantage.

If you run campaigns on professional channels like LinkedIn, you can already use AI to optimize targeting and messaging. Our LinkedIn campaigns and Google Ads services already include automated optimization logic. Likewise, SEO copywriting benefits from AI tools for semantic research and content structuring.

Outlook: where this trajectory leads

In 2027-2028, it's reasonable to expect real-time multimodal models to become a standard component of enterprise AI platforms. Thinking Machines won't be the only player in this space. OpenAI, Google DeepMind and Anthropic are all working on advanced multimodal capabilities.

According to McKinsey projections, the adoption of generative AI in business is set to accelerate significantly over the next two years. As a result, SMEs that start building in-house skills and AI-ready processes today will gain a structural advantage.

That's why SHM Studio supports its clients not only with their day-to-day digital activities, but also in understanding the changes underway. If you'd like to dig deeper into these topics, visit our blog or contact us directly for a consultation.

In short, Thinking Machines' interaction models are a clear signal of where AI is heading. It's not yet time for operational adoption. But it is the right time to understand, plan and position yourself.