Thrive Holdings: measuring model ROI with evals on real daily workflows
samaysham · x · 2026-10-16
Thrive Holdings explains its approach: the firm's job is to turn model progress into useful work and measurable ROI. Running evals on the workflows its businesses perform daily reveals which tasks models can reliably take on, where people still need to step in, and what to build next—consistent with its newly launched evals based on real work from 75+ businesses.
Related event: Thrive Holdings Launches Real-Work Benchmark Across 75 Companies(4 posts)→
More from Research
- RECON paper: reusing REPA's projection head at inference makes DiT training 25x faster — udmrzn · 2026-10-17
- Princeton & UPenn's Recurrent Looped Transformer Carries State Across Tokens — TheTuringPost · 2026-10-17
- Zephon's TorchTitan integration resumes training across GPU counts with identical batches — josh_wills · 2026-10-17
- Mila Opens PhD Position on AI Evaluation from HCI and Technical Angles — windx0303 · 2026-10-17
- Nick Cammarata proposes multi-resolution, no-pruning attribution graphs to fix mech interp errors — inductionheads · 2026-10-17
- RUC open-sources EvoOntology: self-evolving ontology layer lifts data agents by up to 26.7 points — udmrzn · 2026-10-17