An impressive demonstration can hide review costs, integration work and recurring bills. The right project earns its place by improving a meaningful outcome under realistic conditions.
The following are published positions brought together by AIStrategy. They are not answers given to this page’s question or a transcript of a shared panel. Our synthesis and next questions follow below.
THOUGHTLEADERS.GR / THE PERSPECTIVES
Start with what has actually been said.
Microsoft Research
Microsoft Research
Dimitris Papailiopoulos
Machine learning, optimisation & systems
PUBLISHED POSITION / EDITORIAL SUMMARY
More output is not automatically better reasoning: efficiency and accuracy need to be measured together.
Read the source context
Editorial synthesis of the coauthored GFPO paper, August 2025. The results concern specific models and benchmarks; they do not predict a company’s savings.
Innovative Greeks · portrait via ThoughtLeaders.gr
Big Pi Ventures · Upstream
Markos Veremis
Partner, Big Pi Ventures · Co-founder, Upstream
PUBLISHED POSITION / EDITORIAL SUMMARY
AI adoption depends on teams that keep learning and renewing their capabilities; innovation has to become part of everyday operations.
Read the source context
The Era of We: Rethinking Leadership in the Age of AI, Andersen Legal interview, 18 January 2026. Veremis describes collective leadership, continuous learning and operational innovation. This is a leadership argument, not measured evidence of financial returns.
Follow a published position through its original source, the Greek question, a practical case and a next step. This collection reads the central ThoughtLeaders feed.
Connecting to the public collection
Checking source availability…
Published perspectives will appear here when the source is available.
Independent public statements, connected editorially. Source dates describe the original publication; they do not imply attendance at a common event.
Lower unit cost matters only if quality remains acceptable and someone can use the released capacity. Hardware, cloud, software and human review should be compared as one operating cost.
What evidence would move this forward?
Compare the full cost per accepted task, correction rate and service quality against the current process. Include a lower-benefit scenario and a clear date to scale, revise or stop.