Remember my blog on NASA, ESA, and AI? Or, Saab Gripen/E’s record-low operational cost, Software-First architecture, and top-class Electronic Warfare System EWS? Or, on deep ML in life sciences, in-silico research, mRNA vaccine design, or some similarities of MilTech defending a peaceful country against a (Russian) aggressor and MedTech defending patients against an aggressive virus? These are hardly the first ideas to cross your mind whenever talking about AI Factory.
Indeed in practice as well as in training & education, the AIF mainstream has been “traditional” industry such as manufacturing/automotive because of its practical experience in robotics, automation, and lights-out factories (the first robotic line assembling new robots was launched in Japan 23 years ago).
Having said that, let´s add that “new” industry and service sectors are applying the AIF tech&business pattern at an awesome pace.
The common denominator across most sectors is empowerment of users.
Flying a Gripen shall (and does) feel simple and smooth thank all the reasoning made by the pilot’s “tool” itself – the aircraft. This enables him/her to concentrate on tactics, to think a step ahead of the enemy, to jam and spoof enemy sensors; for example, to see a Russian fleet on radar without ever turning the nose eastward, because the EWS can make the radar waves turn sideways (during a NATO war game in the US, a Gripen scored 10 “kills” out of 10 air duels).
Designing a rocket engine at SpaceX shall (and does) feel intuitive and smooth, enabling the designers to concentrate on tasks rather than on tool detail; “rocket science” only literally, no longer figuratively – this demo by Elon Musk was made a decade ago.
I use to lunch with a business colleague since 15+ years back. We cooperated on UML/SysML modeling and training early on; he’s working for a European finance corporation. This spring, we talked mostly about AI-Ops, GPT-4, and generative AI in IT architecture, both ways:
- incorporating deep ML into architecture (Enterprise as well as IT)
- and how an architect can apply AI to his own work, to relieve it from the nitty-gritty of lengthy integration-script writing & tiny detail and to enable him to concentrate on architecture (including of course security, sustainability, and reliability of generated or externally sourced code) , because higher levels of abstraction are part of architects’ job description.
I mentioned the US Congress report by the National Security Commission stating that T. A. Edison’s words about electricity apply to AI: “A field of fields… it holds the secrets which will reorganize the life of the world.” My colleague replied it’s fascinating to watch the fast tech changes all around us, and simultaneously use that very tech daily at work.
A great shift indeed, at lightning speed. Long story short: Architecture for AI in the AI era, and AI-generated “companions” for architects. The same AI Factory pattern, yet pretty far from “Factory” in a narrow sense. Those are just three examples among thousands.
Specifically for big-data architects, I recommend the expert interview this month on innovation with Nathan Marz, in ODBMS Industry Watch, with several useful links plus an amusing one by Blue Origin Space and Amazon founder Jeff Bezos; two minutes on YouTube… Enjoy!
by Milan Kratochvil.
Recommended courses:
AI, Architecture, and Machine Learning
Agile Architechture Fundamentals
Avancerad objektmodellering med UML
(on demand: Modular Product Line Architecture )