Collaborative Intelligence Week in Review - 13Oct2023
Best business article(s) I read this week…
Fueling Interdisciplinary Innovation With AI: Volvo’s Anders Sjögren
This is an episode from MIT’s Me, Myself, and AI podcast (transcript available on linked page). It is an interview with Volvo’s senior technical leader, who is responsible for strategy and innovation. In the interview, Sjögren describes the different ways that Volvo is trying to use AI in their vehicles.
Best technical article(s) I read this week…
Formally Specifying the High-Level Behavior of LLM-Based Agents
This is a preprint of a forthcoming research paper from IBM. The authors present their Linear Temporal Logic (LTL) framework for designing and implementing task-specific LLM-based intelligent, autonomous agents. If formal logic and finite state machines aren’t your thing, skip down to Section 5 to read the results.
Other item(s) of note…
This article is a summary of a 130-page research paper produced by Man Group’s Academic Advisory Board. (The Man Group is an investment management firm with ~$143B in assets under management.) I thought the innovation timeline produced was a valuable summary of recent activity (2018-2023) in this space.
Gartner’s Hype Cycle for Generative AI
This press release accompanied Gartner’s latest publication regarding Generative AI. It includes the Hype Cycle graphic and some other interesting tidbits.
Can AI fix Wall Street’s ‘spaghetti code’ crisis?
Describes how large tech companies (such as IBM and Microsoft) are attempting to use AI to translate old Cobol programs into more modern languages. According to the article, somewhere between 220 billion and 800 billion lines of Cobol are still in use today, much of which is difficult to maintain due to a lack of documentation, poor architecture, and a lack of Cobol skills.
Overview of ChatGPT’s Vision Input Capability
Last week I noted that ChatGPT was releasing new capabilities for voice and image recognition. This Medium article provides a good overview of how the vision input works and how it can be utilized.
Harvard Business Review’s YouTube Channel
In general, I tend to get my information by reading rather than listening (podcasts) or viewing (videos). So, I just discovered HBR’s YouTube site this week, even though it’s been around a long time. This isn’t specific to CI, but I wanted to pass along the link in case you were not aware of it either. It has a lot of good content.
Coolest thing I saw…
Unfortunately, I didn’t discover anything particularly interesting or exciting this week.