2024-05-12
The @GOVUK's AI Safety Institute just released a GitHub repo, inspect_ai, to engineer and evaluate LLMs. It's on the account belonging to the UK Department of Business, Energy & Industrial Strategy. Docs: https://ukgovernmentbeis.github.io/ ... GitHub: https://github.com/... [image]
TechCrunch
The UK government's new AI safety body releases Inspect, an evaluation tool for AI model capabilities, including models' core knowledge and ability to reason
The U.K. Safety Institute, the U.K.'s recently established AI safety body, has released a toolset designed to “strengthen AI safety” …
2023-05-01
Chatbots can't make docs empathetic, but they can help frame empathetic responses. Interesting findings bc physicians here weren't burned out at work but were voluntarily engaging w/ pts on Reddit! Nice work by @DaveySmithMD @CALonghurst and team. https://jamanetwork.com/... https://twitter.com/...
Wall Street Journal
Some US hospitals test if GPT-3 can cut the time staff spend replying to online queries; a study claims the first ChatGPT version replied better than doctors
Pilot program aims to see if AI will cut time that medical staff spend replying to online inquiries
2023-04-19
I have deep respect for AI experts saying Epic shouldn't put LLMs into EHRs, but I disagree. I think this is the right move by Epic. Documentation burden is extensive & templates/scribes/dictation aren't enough. But here are 4 things that are missing: https://arstechnica.com/...
Ars Technica
Epic Systems partners with Microsoft to use Azure OpenAI Service, including GPT-4, to help automate some processes in Epic's electronic health record software
Benj Edwards / Ars Technica :
2021-06-28
They used billing codes to decide who had sepsis, then used the time of clinical intervention (6 hrs prior to sepsis order set, antibiotics, lactate order) to decide the time of onset. This still has some issues but is nowhere near as egregious as what was reported by @verge. https://twitter.com/...
Wired
Study: an algorithm developed by EHR provider Epic to predict sepsis infections in patients missed two-thirds of sepsis cases and frequently issued false alarms
Thank you for folks who have shared or commented on our paper. I know the paper is being used by some to dunk on Epic. Rather than piling on, I want to provide a clear-eyed view of what we found, what it means, and what I would suggest to Epic (& other model devs) going forward. https://twitter.com/...
Wired
Study: an algorithm developed by EHR provider Epic to predict sepsis infections in patients missed two-thirds of sepsis cases and frequently issued false alarms
Since more people have probably read the @verge article than the paper itself, I do want to correct something in the way they paraphrased my comments. Epic *did not* define the time of sepsis based on the time that sepsis was billed. https://www.theverge.com/... https://twitter.com/...
Wired
Study: an algorithm developed by EHR provider Epic to predict sepsis infections in patients missed two-thirds of sepsis cases and frequently issued false alarms
2021-06-27
They used billing codes to decide who had sepsis, then used the time of clinical intervention (6 hrs prior to sepsis order set, antibiotics, lactate order) to decide the time of onset. This still has some issues but is nowhere near as egregious as what was reported by @verge. https://twitter.com/...
Wired
Study: an algorithm developed by EHR provider Epic to predict sepsis infections in patients missed two-thirds of sepsis cases and frequently issued false alarms
A study found that a system used to identify cases of sepsis missed most instances and frequently issued false alarms. Source: JAMA Network .
Thank you for folks who have shared or commented on our paper. I know the paper is being used by some to dunk on Epic. Rather than piling on, I want to provide a clear-eyed view of what we found, what it means, and what I would suggest to Epic (& other model devs) going forward. https://twitter.com/...
Wired
Study: an algorithm developed by EHR provider Epic to predict sepsis infections in patients missed two-thirds of sepsis cases and frequently issued false alarms
A study found that a system used to identify cases of sepsis missed most instances and frequently issued false alarms. Source: JAMA Network .
Since more people have probably read the @verge article than the paper itself, I do want to correct something in the way they paraphrased my comments. Epic *did not* define the time of sepsis based on the time that sepsis was billed. https://www.theverge.com/... https://twitter.com/...
Wired
Study: an algorithm developed by EHR provider Epic to predict sepsis infections in patients missed two-thirds of sepsis cases and frequently issued false alarms
A study found that a system used to identify cases of sepsis missed most instances and frequently issued false alarms. Source: JAMA Network .
2020-10-12
A really nice description of the difference between functional and object-oriented programming. https://arstechnica.com/... https://twitter.com/...
Ars Technica