Skip to content

Finding certainty in tech's unknown future

Senior Managing Director of Teachers' Venture Growth Avid Larizadeh Duggan discusses the beginning of a new tech super cycle.


At a glance

  • Over the next decades, AI will drastically change our lives and work, much of which is still unimaginable.
  • There are things we know for sure that will collectively underscore the importance of strategic adaptation, innovation, and efficient resource utilization in leveraging AI for long-term success and competitive advantage.
  • Clean, reliable, and accessible data is crucial for AI success and serves as a significant competitive advantage. Without strong data foundations, companies risk falling behind in the long term.
  • As the technology matures and new tools emerge, we’ll move from adapting AI to our current workflows to adapting our work around this powerful software because of its promise to boost productivity and effectiveness, increase profits and enhance our lives

Recent developments in AI represent a technology leap akin to the rise of the internet, possibly even more powerful

Over the next decades, AI will drastically change our lives and work, much of which is still unimaginable, just as it was in 1994 when the first websites mirrored real-world businesses. AI adoption and value creation will take time as people and organizations, especially large ones, resist change and need tools, processes and time to adapt. However, the fear of missing out has resulted in billions of investments from incumbents, making them somewhat less vulnerable to disruption at least in the short term, thus making historical pattern recognition and outcome prediction more challenging. So, what can we count on today with certainty?  Here’s what we know for sure.

1. Commoditization across layers

Charles Gorintin, who is the co-founder and Chief Technology Officer (CTO) of our portfolio company Alan, a leading European digital healthcare platform, once told me “Every layer tries to commoditize the one above or below.” Here are a few examples where that’s true in the case of AI:

  • Cloud service providers (Google Cloud Computing, Azure, Amazon Web Services) are trying to commoditize Large Language Models (LLMs) by partnering with them.

  • They are also building their own chips to commoditize the infrastructure layer currently dominated by Nvidia, a world leader in accelerated computing.

  • Facebook open-sources LLMs to commoditize them and use the cheapest solutions in their products. 

  • LLMs aim to commoditize Cloud Service Providers (CSPs) by forming partnerships with companies within this sector.

  • They are also trying to commoditize the application layer by integrating vertically into it.

2. Product is king

Research alone doesn't make a company — product and distribution do. The combination of product vision, execution capability, technology innovation, business acumen and go-to-market leadership lead to sustainable businesses. Companies must build products that meet a clear need for a significant number of customers, show critical usability improvements to existing solutions and elegantly slot into existing workflows to minimize adoption pain. 


3. “Constrained creativity, a driver of innovation in science and AI in particular”  

I am quoting Arthur Mensch, CEO of Mistral, a European-based developer of advanced AI models and tools, told me the company observed first-hand the amount of waste large corporations have with their compute. Throwing money at the problem is not a winning solution as a business model predicated on raising billions is not a sustainable strategy. Mistral has been able to develop one of the most powerful LLMs in market at a mere fraction of the price.  Efficient, cost-effective models will drive innovation, leading to another technological leap as AI becomes more efficient and smaller, revolutionizing our devices.

4. The data moat

As a first step, every company must ensure their data is clean, reliable, accessible and can be easily manipulated and analyzed. Data is the fuel of AI and a moat against competitors. Without it companies will lose against the competition in the long term. It’s a guarantee. 


5. Technology-user experience gap

We’re trying to harness powerful yet immature technology, making usability complex. In the midterm, we will talk about AI systems and solutions rather than LLMs or agents. Ultimately, mature AI building blocks will form modern software solutions focused on solving problems, not just applying new technologies. Those who grasp this quickly will be the likely winners.


6. Focus on capabilities, not use cases

Companies focusing on outcomes, identifying critical problems to solve and building capabilities, which include AI as one building block of many, will dominate. Those finding one off use cases to apply AI will fall behind. An outcomes focus will lead to building capabilities which can be repurposed across functions in the organization from marketing & communications to finance and legal, leading to faster innovation. 


7. Changing user interfaces

The user interface of our devices and products will likely change. We will figure out AI at the edge by making models more efficient, cheaper and smaller. The computers in our pockets will behave very differently from the way they do today. How? Your guess is as good as mine. Perhaps we will invent a completely new device that can understand you, see for you, listen for you, talk for you and earn your trust. What is certain, however, is that those who get it right will follow in the footsteps of Facebook, Instagram and Twitter by reinventing interaction and information access.


8. Rapidly changing AI technology but slow enterprise adoption

AI is still immature but evolving quickly. The pace of change is not about to slow down any time soon because it is driven by an ongoing LLM war fueled by hungry investors and formidable talent which have captivated consumers and enterprises with both excitement and fear. Conversely, these rapid changes will cause slower enterprise adoption. Enterprises need robust tools that solve a problem for their organization, not immature technologies which lead to long proofs of concept and delay scaled deployments.



9. The incumbent winning strategy

Incumbents who combine leveraging existing distribution channels to incrementally improve the user experience, while at the same time taking risks to create new destinations for brand new concepts and products, will be ahead of others. This is a concept that Spotify’s Co-President and Chief Product and Technology Officer Gustav Söderström illustrated beautifully at the Sana AI Summit in Stockholm while describing their approach in delighting users by enhancing existing products with AI and taking risks betting on new ones with the AI DJ.


10. Co-creation between human and machine

It is the first time that the enterprise is faced with a piece of software that doesn’t reliably provide the correct answer. It’s probabilistic not deterministic. Therefore, initially, until we solve this challenge or change our processes to adapt to it, AI solutions will see the greatest adoption in co-creation models where humans are in the loop, augmenting creativity, efficiency and productivity. 


11. Nimble teams and adaptive cultures

Because of the constant change, companies thriving during this period need flexible teams and cultures who embrace change, and are prepared to test, learn and iterate quickly. This applies to both solution creators and adopters and is a crucial investment criterion for those looking to deploy capital into the sector. 


12. Collapsing the talent stack

This is another concept which Alan’s Charles Gorintin elegantly put forth in one of our discussions. At Alan, which is using AI in several ways to enhance its services and improve its digital healthcare experience, he’s seeing the role of the developer changing and a rise in the importance of the technical product manager, who understands product/customer needs and can build using LLMs. For other knowledge workers the shift will be felt first in the most junior roles. Ultimately, all of us will need to adjust to a new technology, a new way of thinking and a new set of tools. Investing in our education establishments from the early years through to university is a necessity. Investing in talent with the right skillsets and helping existing teams adapt to change is critical.

The path forward

As the technology matures and new tools emerge, we’ll move from adapting AI to our current workflows to adapting our work around this powerful software because of its promise to boost productivity and effectiveness, increase profits and enhance our lives. This behavioural change will drive new business models, and in a decade, our work and lives will be profoundly different. This change also presents a massive long-term opportunity for people, businesses and investors. To leverage and influence it we need to find allies to enrich our perspectives, work with adaptable talent, iterate and learn quickly to rapidly refine our vision of the future. Otherwise, we will miss the opportunity of a lifetime.

Avid Larizadeh Duggan, OBE, is a Senior Managing Director of Teachers’ Venture Growth (TVG). Based in our London office, Avid has been active in the startup ecosystem as a developer, product manager, founder and investor for over 20 years. Avid earned her B.S. and M.S. in engineering at Stanford University. She also holds an MBA from Harvard Business School.