We can hand over someone’s age, diagnoses and NEWS and still not quite convey how unwell they are.
If you've ever thought for referrals, handovers, and lists that a common currency of "how sick is this pt" would be useful-
Then read our paper on FI-Lab:
journals.plos.org/plosmedici…
Weeks prior to rotation: new hospital emails rota, logins, induction materials needed for next job.
First day of new job: NHS mail has wiped your account.
Latest NHS mail changes are about to cause chaos to rotational members of staff.
Information governance strikes again.
There needs to be urgent reform.
Here's my rephrase without cybersecurity jargon:
"Our AI model tried really hard to hack out of its sandbox, a computer with no internet access, in order to find the answer to a test problem it had been given. To do this, it found previously unknown software bugs that allowed it to reach an OpenAI computer it wasn't supposed to be able to access. Then it started hacking other computers on OpenAI's networks until it found one that had Internet access.
After gaining Internet access, the AI model thought about where it could find the answers to the test question and figured the AI platform Hugging Face might have the data it was looking for. It then found ways to hack Hugging Face to steal the information it could use to cheat the test. The AI model used several hacking techniques together, including using a stolen password and finding several totally new security bugs in Hugging Face's computers, allowing the AI model to take control of those computers."
Here's exactly what happened, from the blog post:
"While operating in our sandboxed testing environment, our models spent a substantial amount of inference compute finding a way to obtain open Internet access, in pursuit of solving the evaluation problem. To gain access, the models identified and exploited a zero-day vulnerability (which we’ve now responsibly disclosed to the vendor) in the package registry cache proxy. With this access, our models performed a series of privilege escalation and lateral movement actions in our research testing environment until the models reached a node with Internet access.
After gaining Internet access, the models inferred that Hugging Face potentially hosted models, datasets and solutions for ExploitGym. Knowing this, the model searched for and successfully found ways to gain access to secret information that it could use to cheat the evaluation. In one example, the model chained together multiple attack vectors, including using stolen credentials and zero-day vulnerabilities to find a remote code execution path on the Hugging Face servers."
If you break out of your isolation env, get onto the Internet, crack into Huggingface, and steal the answer sheet for your cybersecurity exam, I, for one, would say that you have passed.
7. However no new lives have been saved.
When I commented on Louis Moseley's original post to ask about the sepsis diagnosis rates, which would help clarify, he didn't reply
5. It could just be that many people who previously wouldn't have been called septic were now labelled by lowering the threshold. These people would have survived under either system
2. They did several things at once, they made a dedicated team to go to review patients, and they used palantir software to flag patients who needed review and who were labelled as having sepsis
1. This was not a trial. There was no control group. They compared before and after. The results aren't published in a journal. There's been no peer review. If I were reviewing I'd have several concerns:
The 68% reduction in early sepsis mortality reported in the times does not prove that "palantir tech saves lives"
Short thread - my PhD thesis was about using hospital data to do this sort of thing
This is absurd. Palantir tech saves lives.
If the NHS can’t interpret its own data well it will continue to deliver substandard care.
In Florida, Palantir software has more than halved sepsis deaths. Down 68%. It spots trends early that humans can’t. Abandoning that for ideological reasons is *insane*.
I'm a loyal listener, but this was disappointing. Goddard even admits on air he knows the consultant situation better. Here are 16 reasons we're angry, by the excellent @DevanSinha
Doctors pay dispute is only one, though totemic & overwhelming, of many serious issues affecting working conditions & unattractiveness of the NHS to doctors.
Once started listing issues to express problems doctors face:
XVI Theses on UK Postgraduate Doctors’ Discontent😬
"Why are early career doctors angry?" asks @DAaronovitch on BBC Briefing Room.
Two consultants who qualified in 1979 and 1990 are on to explain.
Part of it? Feeling that consultants don't understand their struggles. Funnily enough, that didn't come up.
Here's a chart of the marginal tax rate for someone with two kids. They pay 53% tax on every £ they earn between £60k and £80k. 62% on every £ they earn between £100k and £125k.
9% more if they've a student loan
(See all the detail in our calculator, pinned to my profile)