Collaborative Intelligence Week in Review - 30Aug2024
Best business article(s) I read this week…
How Human-Informed AI Leads to More Accurate Digital Twins
This article describes how human expertise and AI can enhance the accuracy and effectiveness of digital twins. This approach reduces costs, accelerates innovation, and improves safety, making it a valuable strategy for mission-critical applications. The article provides examples and insights, highlighting a hybrid methodology beneficial for industries where data is limited but precision is essential.
Do Not Use LLM or Generative AI For These Use Cases
I have written at length about the need for executives to think more broadly about AI because Generative AI is often inappropriate for many applications. This article summarizes twelve use cases and six AI techniques and concludes with a matrix showing where different AI techniques are most appropriate.
Best technical article(s) I read this week…
Leveraging Large Language Models for Enhanced Process
The authors describe how they trained an LLM to understand and generate responses to specific questions related to business process models. I found this interesting because it is another valuable generative AI business application. Businesses can use the approach described (and the open-source software provided) to improve their understanding of their business processes and identify prospective optimizations.
Other items I found valuable/interesting…
Digital Doubles Vs. Digital Twins: Unraveling The Distinctions
This podcast from Forrester describes the differences between digital doubles and digital twins. Then, it discusses the use cases for each technology and some challenges in deploying and using each effectively.
Engineers design tiny batteries for powering cell-sized robots
I included this because I think it might be a very weak signal of an emerging trend: microscopic robot swarms that can sense and respond to their environments. This technology has a wide range of applications, including healthcare, agriculture, and the military.
20 Generative AI Tools For Creating Synthetic Data
I generate a lot of synthetic data for testing on confidential or HIPAA-restricted client data. While I was aware of several tools on the list, some were new to me.
Coolest thing I saw…
DeepMind’s AlphaProof and AlphaGeometry 2
Google DeepMind released a video explaining how its new AlphaProof and AlphaGeometry models can solve Math Olympiad problems. Although the technical papers were released earlier this year, this video is much easier for a layperson to understand. I’m sort of a math geek, so I thought this was cool.
A company that caught my eye…
The company, which automates software development via generative AI, announced it had closed a $320M investment from Eric Schmidt, Sequoia, and Atlassian (among others). It also announced a partnership with Google Cloud to build two supercomputers that can scale to “tens of thousands of [NVIDIA] Blackwell GPUs.”