Construction Technology Podcast
Oct 24, 2026 · 50 min · 9 segments
Josh Whisenant of Roger joins Luigi La Corte, Mike Powers, and Andrew Zukoski the day after Procore's Groundbreak conference in Orlando. The group recaps Procore's AI announcements, including the…
well they did launch a model maybe we can do a bit of like a ground break retro yeah sure because josh was there with us too but um they did release a model magpie magpie 3.0 I wasn't aware that there were previous versions of Magpie, but...
If you just release a model with like, yeah, this is provision intelligence 8.0, guys, where have you been? You're like, well, maybe we've missed it.
The problem with these benchmarks is if it's not a third-party benchmark, it's just so manipulatable.
On this one very discrete task, are we better? It doesn't mean you're frontier unless it's this very small task.
But anyway, so Magpie 3.0, it's a reasoning model to choose between their taxonomy of document types, what actions to take, I mean, there were so many words thrown around, like model, ontology.
But I guess what a customer would care about is Magpie helps get tasks within Procore done better than GPT-6 Astra with API access.
How does it work? Do you get tokens of the customer? And you can use 10 tokens a month? How does it work?
well they did launch a model maybe we can do a bit of like a ground break retro yeah sure because josh was there with us too but um they did release a model magpie magpie 3.0 I wasn't aware that there were previous versions of Magpie, but...
If you just release a model with like, yeah, this is provision intelligence 8.0, guys, where have you been? You're like, well, maybe we've missed it.
The problem with these benchmarks is if it's not a third-party benchmark, it's just so manipulatable.
On this one very discrete task, are we better? It doesn't mean you're frontier unless it's this very small task.
But anyway, so Magpie 3.0, it's a reasoning model to choose between their taxonomy of document types, what actions to take, I mean, there were so many words thrown around, like model, ontology.
But I guess what a customer would care about is Magpie helps get tasks within Procore done better than GPT-6 Astra with API access.
How does it work? Do you get tokens of the customer? And you can use 10 tokens a month? How does it work?
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.
