Fantastic article, thanks. Can you please be more specific when you say “NBIM got their data house in order first.” ? What data needed to be moved where, why and how?
Great question, and I wish I had a more detailed answer. NBIM hasn’t published the full technical breakdown. But here’s what we know from the seminar and their public materials.
Bryne described three phases that all happened before AI entered the picture. First, they insourced operations they’d been outsourcing to external vendors, bringing front-office processes and tech infrastructure back in-house. Second, they moved everything to public cloud (they were actually a first mover in Norway on this). And third, they did the database migration and data cleanup work, which is where the hard deadline came in: old systems go dark on January 31st, figure it out or you’re sitting there with nothing. The specifics on which databases went where and what “clean” meant for their data, they haven’t shared publicly. My read from Bryne’s comments is that the big issue was the usual one: data spread across systems that don’t talk to each other, inconsistent formats, stuff living in places it shouldn’t. The kind of work nobody wants to do but everything downstream depends on. I’d point you to the full seminar video since Bryne walks through it herself, and her delivery makes the point better than I can in a comment.
Fantastic article, thanks. Can you please be more specific when you say “NBIM got their data house in order first.” ? What data needed to be moved where, why and how?
Great question, and I wish I had a more detailed answer. NBIM hasn’t published the full technical breakdown. But here’s what we know from the seminar and their public materials.
Bryne described three phases that all happened before AI entered the picture. First, they insourced operations they’d been outsourcing to external vendors, bringing front-office processes and tech infrastructure back in-house. Second, they moved everything to public cloud (they were actually a first mover in Norway on this). And third, they did the database migration and data cleanup work, which is where the hard deadline came in: old systems go dark on January 31st, figure it out or you’re sitting there with nothing. The specifics on which databases went where and what “clean” meant for their data, they haven’t shared publicly. My read from Bryne’s comments is that the big issue was the usual one: data spread across systems that don’t talk to each other, inconsistent formats, stuff living in places it shouldn’t. The kind of work nobody wants to do but everything downstream depends on. I’d point you to the full seminar video since Bryne walks through it herself, and her delivery makes the point better than I can in a comment.
https://youtu.be/hBGd3DCgRkM
What do you tell them about data protection and cyber security when building AI ‘solutions’?
Tell them not to worry
Where do I find NBIM’s playbook?