AiThority Interview with Chris Fisher, Founder & Managing Partner of Myriad Venture Partners

Hi Chris, welcome to our AiThority Interview Series. Please tell us about your AI journey so far.

Thank you! Very excited to share our insights and ongoing developments. The constant evolution of AI is extremely exciting, holding the promise of democratizing its usage along with the various tools and functions it can either support or replace. Over the years, like many others, we’ve closely observed this maturation and consistently sought the best use cases for integrating AI into our partners’ business infrastructure and processes. AI is now a transformative force across industries, transitioning from a niche technology to being intentionally embedded into all types of products and services. 

What is Myriad Ventures and how does it fit into the current technology landscape?

Myriad Venture Partners is an early-stage venture firm focused on AI, clean technology and B2B software. Operating under a traditional GP-LP structure, Myriad is launching a $200 million independent venture capital fund to build on its founders’ previous success at Xerox Ventures. Our investment is anchored by Xerox and other limited partners. Myriad leverages its network of corporate titans and top-tier venture capitalists to build mutually beneficial partnerships with cutting-edge businesses and founders. This collaborative approach enables companies that have faced challenges in effective R&D and M&A to partner with best-in-class startups to generate new offerings and profitable revenue.

You’ve recently made investments in Sentra, Vartana, and Quadric. Could you elucidate the strategies underlying these investments, particularly in the context of automation?

Sentra presents tremendous potential with its advanced cloud-based security solutions for enterprises. Leveraging automation, it detects and prevents cyberattacks, optimizes security operations and ensures compliance. This results in an impressive reduction in security costs by up to 80% and enhances security posture by up to 90%.  Vartana, on the other hand, stands out for its expertise in complex, sales-person-driven, B2B transaction markets. Offering a sales closing and financing platform for B2B enterprises, Vartana streamlines the closing process, provides payment flexibility and generates real-time quotes through automation. Vartana also integrates with CRM systems and provides analytics and insights, increasing conversion rates, decreasing time to finance and improving cash flow. 

Quadric’s general-purpose neural processing unit (GPNPU) holds great promise for on-device artificial intelligence computing. It uses automation to simplify the design and programming of chips that can run any kind of machine-learning model. Quadric also provides a developer studio that enables easy simulation and visualization of AI software and SoC design choices. This combined platform is designed from the ground up to deliver high performance, low power, and low latency in edge AI applications. 

In summary, these companies showcase diverse applications of automation, highlighting how automation can enhance efficiency, reliability, speed, and innovation in various sectors and scenarios. 

Are there security and privacy concerns that should be considered when utilizing generative tools?

The rise of large language models like ChatGPT has enabled new possibilities for generating human-like text, images, code, and more. While these advanced generative AI systems offer significant capabilities, they also bring forth notable security and privacy risks that enterprises must carefully evaluate before deployment. 

Responsible adoption requires proactively addressing challenges around data sourcing, biases, misinformation and legal compliance. A core concern is ensuring training data is collected ethically and does not unintentionally expose private information. Generative models are typically trained on massive datasets scraped from the public web or proprietary corpora. Even de-identified data runs the risk of encoding sensitive attributes. Rigorous auditing and testing of datasets are crucial to avoid violations of user privacy expectations or rights. Applying protocols like differential privacy and secure computation can help mitigate exposure risks.

Another major challenge is the potential propagation of biases through generative models, amplifying harmful stereotypes. Rigorous testing on diverse input prompts is imperative to surface any skew (hallucinations) in outputs across factors like race, gender, or ethnicity. Mitigating bias requires proactive data filtering, aggressively augmenting underrepresented groups in training data, and preference conditioning during generation. Achieving fair and inclusive AI remains an ongoing research problem. 

Generative models also carry the potential of spreading misinformation if deployed without appropriate safeguards. Ensuring quality control processes for fact-checking outputs before dissemination is vital, particularly in specialized domains like healthcare and finance where erroneous content could cause tangible harm. Even well-intentioned models often generate plausible-sounding but incorrect assertions. Monitoring and evaluating veracity is an ongoing responsibility. 

Finally, legal and ethical hurdles surrounding intellectual property, data protection and accountability must be addressed. Improper deployment of generative models could lead to violation of copyright or privacy laws, and the liability in cases when AI systems cause harm remains unclear. Proactive governance and oversight are pivotal in navigating these complex challenges. 

Around 57 % of North American and European organizations currently use IoT technology within their business operations. What are your comments?

This reflects the maturation of IoT from early proof-of-concept experiments into mission-critical deployments delivering real business value. Industries like manufacturing, energy, and transportation/logistics are undergoing digital transformations centered around IoT platforms and applications. IoT facilitates the connection of previously siloed operational data sources, the application of analytics, and the facilitation of automation. 

Enterprises have drawn insights from early IoT projects and now embrace more contemporary IT approaches including hybrid cloud and edge computing. This enables efficient handling of massive amounts of IoT data across distributed environments. IoT security has advanced with solutions and managed services focused on operational technology (OT.) As operational networks become interconnected, ensuring their security becomes a prerequisite for widespread adoption. 

Advancements in technologies like low-power WANs and 5G networks have improved enterprise IoT connectivity. While 5G has not yet fully delivered on its promise, continued enhancements will make it more viable and affordable.

How do you prepare for the AI-led disruptions?

The rapid advances in artificial intelligence represent both massive opportunities and significant disruptions for businesses across every industry. To harness AI’s capabilities while managing risks, our portfolio companies help organizations take a strategic and proactive approach to integration. 

We prioritize solutions that enhance internal awareness at all levels regarding AI’s emerging applications and potential benefits. Cross-functional training is key to empowering employees to visualize how AI can enhance products, services, and workflows. It is also critical to foster an understanding of AI’s limitations and challenges, such as algorithmic bias, to maintain a realistic perspective and temper unrealistic expectations. 

In collaboration with our partner organizations, we conduct in-depth assessments of where and how AI could impact their operations, employees, and customers. Scenario planning helps reveal blind spots and knowledge gaps, and assessing operational readiness across parameters like data, infrastructure, and skills is crucial. Developing robust risk assessment practices also helps anticipate downsides like job losses. 

With foundational knowledge in place, pilot projects help build critical hands-on experience. Starting with limited deployments enables testing, learning, and adaptation before broader rollouts. Interdisciplinary teams with both technical and domain experts fuel creative exploration. Feedback loops ensure user needs and concerns guide development. With a deliberate strategy, companies can lead rather than follow AI’s emergence. Change inevitably brings both turbulence and opportunity — organizations that actively shape their integration of AI will best position themselves to maximize its benefits while navigating its risks.

Please tell us about the major milestones you managed to achieve in 2023.

We successfully secured initial funding of $100M, while building an initial portfolio of 14 dynamic companies, including Seurat & Mojave. In addition, we have been able to build out our corporate advisory council, starting with HCL and our anchor investor Xerox. 

Could you shed some light on how you’re helping your portfolio companies address innovation challenges?

Myriad’s investing strategy solves the “innovator’s dilemma,” meaning mature businesses must innovate or get disrupted by innovative startups, and Myriad extends the lifespan of today’s big industries, by creating partnership opportunities with the big ideas and industries of tomorrow. The firm focuses on building long-term partnerships with the most groundbreaking startups, creating a runway for them to grow into leaders in their markets.

In addition, Myriad is helping its corporate partners run more effective R&D and technology sourcing models, whether that be partnering with best-in-class startups vs internally developed assets, or monetizing and spinning out internally developed assets to create startups funded by the venture ecosystem – we help our corporate partners achieve both.

For instance, we helped Xerox spin out Novity and Mojave, both are operating as venture-backed companies today, with Mojave securing a funding round led by Fifth Wall and At-One.

How are GenAI capabilities sparking an exciting revolution in the field of AI?

The advent of large language models like ChatGPT represents a significant product-focused shift for artificial intelligence, unlocking new generative capabilities that are fueling tremendous innovation. While AI has made strides in analytical applications like computer vision and predictive modeling, generative AI allows for free-form synthesis of content, interactions, and even code. This paradigm change is pushing the boundaries of what is possible with AI. 

One important aspect is the ability to generate highly realistic and coherent text that is responsive to natural language inputs. Whether composing emails, reports, or code, few-shot learning approaches like prompt engineering open endless possibilities for customizable text generation. And text is just the beginning — similar techniques now enable the creation of synthetic media like images, videos, and audio. 

Generative AI also shows promise for automating software development through code generation and paraphrasing. Instead of manual coding, developers can describe desired functions in plain language, allowing models to translate these descriptions into executable scripts and applications. This has the potential to significantly accelerate programming while making it more accessible to non-experts. 

More broadly, generative models excel at inferring patterns from massive datasets to produce useful outputs, unlocking applications from personalized recommendations to advanced question-answering systems. Their flexibility and proficiency in various language-related tasks differentiate them from previous narrow AI solutions. While generative AI does face challenges around data bias, security, and transparency, its sheer versatility foreshadows a new paradigm for harnessing artificial intelligence. Much as assembly lines revolutionized manufacturing, this technology signals a shift toward mass-producing intelligence. The ramifications for how AI gets built, deployed and used will be profound. We have only begun tapping into generative AI’s transformative potential.

Where do you see AI/Machine Learning and other smart technologies heading beyond 2024?

One significant trend is the broadening scope of generative AI to evolve into multipurpose intelligent agents. Models like ChatGPT already display strong language processing skills, and continuous improvement in areas like computer vision, reasoning, and planning is anticipated. Bringing together strengths across modalities will enable increasingly versatile systems for knowledge work and digital assistance. 

We also expect to see greater advances in unsupervised and self-supervised learning, allowing AI systems to learn from unlabeled datasets with less human oversight. This will expand the types of tasks AI can take on. The growth of on-device AI processing, exemplified by efforts such as those by our portfolio company Quadric, involves more computation occurring locally on devices like phones rather than in the cloud. This shift offers advantages in terms of latency, privacy, and reliability. 

We’ll also be looking at the expansion of AI into more specialized vertical domains such as healthcare, manufacturing, and finance. Pre-trained models can be adapted and fine-tuned for industry-specific applications, enhancing their effectiveness in diverse sectors. Lastly, we’re excited about more advanced robotics and embodied AI systems that can physically interact with the world. Research areas like simulated learning environments and high-fidelity 3D models are expected to contribute to the realization of more capable real-world AI.

Thank you, Chris! That was fun and we hope to see you back on AiThority.com soon.