The difficulty with the generation study is that there is no way to selectively opt in/out; your child is sequenced and then the data is retained until you opt out (at which point it is still retained as part of historical releases). It isnt ringfenced for medical research and can be accessed by pharmaceutical companies. It isnt even kept by the NHS but rather by a private arms-length body of the government which could be privatised under a change of leadership. We've seen failures with UK Biobank data security, why would this be any different?
Unsurprising. There's a poor economic argument for funding healthcare in the UK, other than it being used as a political football to win votes. The result is decades of underfunding and then massive tranches of funding to make headlines which has to be spent immediately and therefore gets squandered on "transformation initiatives" without a focus on health outcomes.
The public are being told that the UK can't afford to pay healthcare staff market rates, and so they are all leaving the NHS. The result is this dip in GDP, which IMO is more expensive.
Came here to say this. Both Seveneves and Fall have long unending psuedo-epilogues which add little to the main story, but also don't really wrap anything up satisfactorily either.
Just finding this now after noticing the traffic spike (thanks!). Yes we are playing with doing this, however to be honest it's a very challenging task. You can extract the 'key information' from a paper, but abstractly summarising concepts is a bit more challenging. Probably the way forward will be a tool-assisted approach to aid humans to write better.
Really disappointing to see a whole host of inaccuracies in this article, even beyond the artistic license you'd expect for the New Yorker. My own area of knowledge is in neuroscience so I'd focus on this.
For instance, while its somewhat correct that nerve signals are binary, insofar as they are all-or-nothing, and yet activity is highly analogue, with specific groups of neurons firing together in pre-programmed bursts & rhythms. This has been one of the challenges in producing BCIs
Sure, but its the corner of two rooms on two floors. Depending on what your house looks like, that's quite an undertaking to fit, it's destructive to the house, and doesn't replace your need for stairs.
I can see the use of this for neurological disease, but far and away the most common reason for a stairlift is frailty (or sarcopenia, to be more precise) along with other musculoskeletal conditions like osteoarthritis. Generally in sarcopenia you lose much more quadriceps strength, such that you find it difficult to stand up out of a chair, which is precisely the motion of how this is propelled. I'd therefore find this difficult to justify.
Secondly, stairlifts work because they are retrofitted into people's homes. This would require quite considerable changes to fit, and possibly loss of an upstairs room to make it work.
Everyone in the comments discussing the UX of scrolling in this setting- this is exactly how you look through a CT scan! And in the different planes, as in this page. Sure a slider is useful, but I think this is really authentic.
Hey HN, I've been sharing some of the articles about this website that I've set up for a little while, and been developing it based on your feedback.
The aim is to simplify/explain/debullshit papers about AI in healthcare so that patients and doctors can better understand, use and trust them.
I think it's now at the point of being feature-complete, and we've got about 75 papers summarised on there now, but would love any thoughts on this.
Yes this is why I'm not sure how it's been tested. In the scientific literature in many cases the data is anonymised and made available publicly or to those interested, but you can't always anonymise the data adequately, so an audit process might be necessary
Completely agree. I've not seen it tested legally, but the EU now has a 'right to explanation' where automated decisions are made about people. This would prohibit closed ML from most arenas.
That's fair- I've edited the title to make it more specific. I've not come across Bayesian but I'm going to add their research on the site to summarise down the line.
One of the problems with closed models is that any model can be found to train on the 'wrong' data point. So e.g. a chest x ray reader determines that images taken with the machine in ICU indicate sicker patients than elsewhere- that's not useful. If you can't inspect the model to check that, they might claim superior performance, but then the model doesn't work as well as advertised when it's tried out. Other biases might occur as well- for instance you can imagine a 'Greyball for healthcare' with the wrong incentives which recommends a certain drug/therapy more often than it should.
Granted this is the review of a single model, but given that it's by one of the big players in the EMR space, it shows how challenging this problem is. Sepsis is easy to under-recognise, and difficult to identify reliably without lots (and lots) of false alarms.
I can't comment on every use case, but I've recently been using htmx[1] with Django templates to provide sort-of async behaviour. For me it's the best of both worlds; I can specify my template and view as though it's a normal page load, and then use some hx- attributes to make it async.
Anecdotally, my university has been paying for fewer and fewer journal subscriptions. I was looking through the literature the other day and about half of the articles I needed to access weren't subscribed to by my (large, redbrick) university. I believe this is a shift recognising that academics will just skirt the access via sci-hub as a way for universities to trim their library budgets.
Most journals are hybrid now, and offer a choice when publishing- either it's free for the author but kept behind a paywall, or the author pays up front. The article processing charges are really expensive- BMJ is £3,500, Lancet $5000, Cell $5,200 etc, although some major funders have arranged 'site licenses' with publishers so that all their funded research can be published open access without charge to the scientist.
I'm very aware that I'm a HN novice, but can I ask why my post title was edited? The new title is much less descriptive, and x-rays are different from medical images, after all.