Technical leaders trying to integrate LLMs into their systems often describe a familiar tension: the desire for capability, managing costs and the need for control. For the past few years, the architectural pattern has heavily favored capability; we've routed our most complex workflows to proprietary APIs, trading data sovereignty and fine-grained control for the sheer power of closed-weight frontier models. However, things are changing, not least because of growing concerns about token costs.
Richard Gall
Verified
As seen in:
Thoughtworks,
Medium,
betanews,
The Code Project,
DZone,
Packt Hub,
KDnuggets,
Singularity Hub,
The New Stack,
Security Boulevard,
Training Industry
and
milquetoast of human kindness | occasional poet | podcast @_TalkAboutTech | @UTAW_uk member
Get in touch with Richard
Contact Richard, search articles and posts on X, monitor coverage, and track replies from one place.
Learn more about Muck Rack