2024 marks the second full year of blogging and sharing my insights, analysis, perspectives, and general thoughts with the world about Artificial Intelligence and the businesses that both power them and utilize their capabilities. One of the biggest milestones, in my opinion is my thesis on Artificial Intelligence which summarizes, relates, and interweaves the common themes, frameworks, and concepts discussed in my blog.
Derived from that thesis are major frameworks I’ve developed as constructs to understand the ever-evolving landscape of Artificial Intelligence around us:
- The Council of Experts Leader-Agent Model — An idealistic agentic architecture where the leader agent breaks down a problem and utilizes industry-specific or use case specific expert agents to provide input that facilitates the completion of a task. The framework also interwove the parties involved from foundation model makers, hyper scaler platforms, industry specific model creators, and solution providers who integrate the entire stack. This idealized future remains my hope for a truly revolutionary AI-powered world that scales human intelligence.
- A Roadmap to Trust — as large enterprises and consumers alike starts to adopt Generative AI capabilities, even agentic architectures, there is a level of comfort that is required with slowly relinquishing control to both autonomy and probabilistic responses rather than static. Designers can also utilize design to aid consumers in adopting Generative capabilities by abstracting out complex prompts and creating wow moments that follow intuitive, simplistic, yet powerful interfaces.
Despite my personal favorites in the frameworks, readership statistics showed otherwise. The top 5 blog posts read on Medium (Linkedin did not provide article specific analytics or comprehensive rankings that were easily accessible) included (in order):
- Opening Pandoras Box — the Music Genome Project and LLMs — a fascinating history behind one of the earliest Artificial Intelligence players I’ve interacted with, centering around building a genomic dataset of music to create the Pandora Radio’s recommendation engine
- A Quantum Leap in AI — How Quantum Computing Could Remodel AI — foresight into the realm of quantum computers and the massive impact a breakthrough in quantum could mean for AI model training and inference
- A Path of Productization — AI Pathways for Business Implementation — A decision model on whether enterprises seeking to associate with AI should either build, bolster buy, or bail and the criteria for each
- The Council of Experts Leader-Agent Model — A Proposed Framework for General Intelligence — as noted above, my idealistic framework for an agentic, industry specific, competitive AI ecosystem
- Assessing the Landscape of Adaptive Testing Modalities with AI — A perspective on how Artificial Intelligence could augment adaptive learning assessments and bolster that modality of assessment
Most importantly, I am extremely humbled by the opportunities over 2024 to engage with the community and share my perspectives as a subject matter expert in AI boasting a 10-year career to date:
Finally, I would be remiss without mentioning some of my foresights into 2025, building upon another year of experience and attempting to provide an insightful view into the future instead of hype-driven comments:
- Models that “think” — starting with OpenAI o1 and expanding through Google Gemini 2.0 Flash Thinking Experimental, machine learning specialist are navigating around the theoretical “limit of available data” towards utilizing chain of through prompting and researching breakthroughs in model “thinking” or breakdown which will only accelerate into this year. Continue to hear news about new models quarterly as foundation models continue an ever-competing cycle in a commodity space.
- Agents — Agentic architectures quickly reach the Gartner’s peak of inflated expectations as marketing spins claim agents are the future. In reality, they are AI Microservices and will truly make since in a leader-expert model. Combined with the point above into a single ecosystem or platform could be quite intriguing.
- Regulation — this topic, honestly, is a wild card with the incoming administration so I will leave this topic alone.
- AI Start-up Bust — Entering our 3rd year of generative AI, I anticipate that 2025 will start to shed light on start-ups with claims to AI or Generative AI rich capabilities will start to fail and lose traction and will surprisingly hinge on OpenAI’s ability to continue captivating a global audience with ChatGPT as a consumer product. The better the perception of consumer available “wow” products exist, the more traction these companies will gain, however, once OpenAI loses ground, anticipate reality will start to show as these early AI companies simply road the hype and did not have either a short-term design or long-term data advantage.
- Hardware shift — a gradual shift away from Nvidia and other chip manufacturers in favor of homegrown AI-specific chips for vertical integration will begin, hinging on a nuanced view of “American designed chips” but realizing that the fabrication of such will still be shifted abroad. Should that be stifled, that will only put more strain on the system. Fabs, in reality, take nearly a decade to realize benefits so I anticipate that view should be adopted and applied here.
In summary , I estimate that 2025 will shift the AI landscape once again with models that can decompose problems which could start to yield highly valuable use cases for the technology, especially once multimodal and industry-specific use cases get factored in. Combine that with lower regulation and that is a recipe for innovation, so long as AI practitioners do not get distracted by AGI.
In the coming year, I promise to continue to share my perspectives on the AI industry, tie them to personal anecdotes, and experiment more with AI solutions such as Napkin.ai or Google Notebook. My aim is to engage more with the community, both via comments (please write and share opinions, I’d love to have a conversation with you) and speaking events, both virtually and in person. I am deeply honored and humbled by my readership and those collages and friends of mine who have shared their experiences reading my blogs. Thank you for a wonderful 2024 and I look forward to the next year of innovation, progression, experimentation, and exploration into the world of AI.
