LinkedInยทFriday, 24 July 2026ยท24 Jul 2026
๐ช๐ต๐ฎ๐ ๐ฃ๐ฎ๐น๐ฎ๐ป๐๐ถ๐ฟ ๐๐ผ๐ ๐ฅ๐ถ๐ด๐ต๐ ๐ช๐ถ๐๐ต ๐ข๐ป๐๐ผ๐น๐ผ๐ด๐ถ๐ฒ๐ They put the #ontology at the centre and the model at the edge โโฆ
Nathan Bijnens
CEO at Scrydon | xMSFT | Your Data, Your AI, Your Control
๐ช๐ต๐ฎ๐ ๐ฃ๐ฎ๐น๐ฎ๐ป๐๐ถ๐ฟ ๐๐ผ๐ ๐ฅ๐ถ๐ด๐ต๐ ๐ช๐ถ๐๐ต ๐ข๐ป๐๐ผ๐น๐ผ๐ด๐ถ๐ฒ๐
They put the #ontology at the centre and the model at the edge โ years before it was fashionable.
While most of the industry was bolting LLMs onto existing data stacks, Palantir inverted the architecture. The language model became a replaceable reasoning component. The durable intelligence โ what the business actually is, how it works, and what happened last time โ lives in the ontology.
That architectural bet is being proven right.
Two things have changed since then.
First, #AI itself has collapsed the cost of building a proper ontology. What once required an army of forward-deployed engineers can now be done far faster, with domain experts validating rather than narrating.
Second, the more valuable the ontology becomes, the more critical the question of who owns it โ and under whose jurisdiction it sits.
The model is becoming a commodity.
The semantic model of your organisation is not.
Full post: https://lnkd.in/eduTQtr7
Is the industry still over-indexing on models while under-investing in the only part that actually compounds?
What Palantir Got Right With Ontologies - Scrydon
https://scrydon.com/insights/2026/07/24/what-palantir-got-right-with-ontologies/
๐ฌ 12โป 5
View on LinkedIn Cross-referenced
Related on the wire
Every time my house decided to charge its battery to dodge Flandersโ capacity tariff, that command travelled from Belgium to aโฆ
How a version update can cause major impact... in a good way!
Five people decide whether AI actually happens in a European public body. ๐ง๐ต๐ฒ๐ ๐ฎ๐น๐บ๐ผ๐๐ ๐ป๐ฒ๐๐ฒ๐ฟ ๐๐ถ๐ ๐ถ๐ป ๐๐ต๐ฒ ๐๐ฎ๐บ๐ฒโฆ
July has been interesting for #CyberSec & #AI, frontier AI labs disclosed that models under evaluation escaped their sandboxes and reachedโฆ
Governance as kernel, Context as memory, Process & Workflows as scheduler, Integrations as drivers โ and Humans, deliberately last, as theโฆ
๐ช๐ต๐ผ ๐ฎ๐ฐ๐๐๐ฎ๐น๐น๐ ๐ป๐ฒ๐ฒ๐ฑ๐ ๐ฎ๐ป ๐๐ ๐ข๐ฝ๐ฒ๐ฟ๐ฎ๐๐ถ๐ป๐ด ๐ฆ๐๐๐๐ฒ๐บ? Most companies are still treating AI like a fancy personalโฆ