lukko

The cloth sim is GPU. It’s not that heavy currently, around 40k quads. You could go much heavier, easily 500k, but it also changed the cloth behaviour.

The limiting factor is actually not the sim, but the initial load time for the procedural geometry. I might try adding some subdivisions though..

haha, yes - it’s definitely possible, just may be bit tricky.

You can get detect planes and the mesh geo of the room.

It's because of the simultaneous recording / screen capture - it's smoother on device with less lag.

Thanks!

It's been fun. It's not too dissimilar to iOS - a lot of spatial capabilities are linked very closely with RealityKit, it's worth looking at the API if you're interested. I was thinking we'd use Metal for rendering, but as I think there are privacy issues with accessing the raw camera data, it's only supported for 'full' spaces, so not the mixed / camera feed + overlay.

Yes, that's basically it. If the gesture is targeted to something, either by looking at it or touching it directly in space then it responds. Direct touching takes priority, so I think you could look away but still touch in space.

Haha, I wish - that would be great!

Good point - I should have probably shown the direct interaction too (touching the actual fabric in space) - there's a gif here: https://jmp.sh/s/lHqJm6NEvqMqkXMPyUpZ. It was a little but laggy when also screen recording on device.

In the video I am looking at where I want to interact and then using a pinch gesture and the sounds are mapped to different cells of the cloth. By either looking and tapping, or playing directly you can hopefully play your intended sound.

This is cool - not sure if it's possible, but it would great if you could determine other things, like is the needle weight correct, or is the cartridge aligned properly (maybe left / right channel would be louder / softer??).

I found the copy on the site a bit gross - I'm not sure something like this needs a critique of modern copywriting, and for me it gives that same weird sense of over familiarity ('we get you') that I think it's trying to criticise.

why did you stop playing?

Hmm, why does the hydrogen atom look like a d orbital?

Isn't it 1s1 in the ground state so the probability distribution would look like a sphere.

This is cool - for Metal, the debugging tools are pretty useful in Xcode - you can access buffers and read values directly.

Related - a while I made a process to convert DICOM / MRI files to voxels for printing at the native resolution of 3D printers [0]. It means you can preserve all the fine detail of the scans, rather than converting to a mesh (segmentation).

On the CD question - it's probably because there is little incentive to build a secure / cross platform solution for patients to access their scans. The CD model is very outdated, but does work, and there is no need for HIPAA compliance even though a CD isn't very secure.

[0] https://www.lukehale.com/voxelprinting

There are a few other reasons - doctors document every interaction partly for medico-legal reasons – just in case something happens. The notes become especially long and defensive in any situations that have a possibility of being misinterpreted. If it's not written down, it didn't happen. It's obviously also a record for other clinicians / healthcare professionals to read through and see what happened during the admission.

But yep I do worry about any kind of generative AI in this context.

One of my old registrars co-founded this company: https://tortus.ai. They are doing a trial at Great Ormond Street at the moment - I haven't tried what they're building but it's an AI assistant that reduces some of the admin burden.

I am really hopeful that systems like this will take off – the reality of being a junior doctor in the UK is that most of your time will be used on quite tedious admin tasks (documenting every patient interaction, filling forms, booking clinics etc.) using very & slow outdated computer systems. I don't think anyone expects this when they apply to medical school, and it can be quite demoralising when you start your first job.

Thanks for sending! Yep, I have tried it before - it's cool and definitely in the same ballpark.

This is amazing! It's very satisfying to use and the combination of transcription + intent seems like it has huge potential.

I would love to use this in healthcare for dictating patient letters etc. I guess a local model / HIPAA compliance is some way off?

Working on Lungy:

https://www.lungy.app

It's an interactive breathing app (responds when you breathe). Initially for stress & anxiety, we're developing a medical device (SaMD) for asthma + COPD.

Well done for shipping! I'm really impressed by the tool - it's intuitive and like a very light-weight After Effects.

I think just re-designing the landing page would make a huge difference, very quickly. I think currently it lets down what is a well-crafted design / animation tool. Get rid of the white and teal, and the default system fonts. The looom site is a lovely example (https://www.iorama.studio/looom).

Hey, thanks for downloading and for the feedback! The Terms of Service is a tricky one - I know most people don't read it, but we did try to summarise the most important points, rather than just having a tiny link that will no one will click. Thanks for pointing out the typos.

On open source / free - I would love to be able to give access to the whole thing for free, but unfortunately it's just not sustainable. The free version is very good though, with unlimited use and learning modules. Even with a universal basic income + health insurance, there's no way to fund, grow and release a medical device without some kind of monetisation - and grant companies and investors would definitely expect some kind of business plan. I'm not sure there would be any benefit from open source in this case? It just means it wouldn't get to market as a medical device.

Hey HN!

I’m an NHS doctor and the founder of Pia (https://www.piahealth.co) which developed Lungy (https://www.lungy.app). Lungy is a mindful breathing app that uses the camera +/- microphone to respond to breathing, generating real-time visuals and tracking your breathing progress over the exercise. It’s initially a wellness app (for stress & anxiety), but we’re very close to having a medical device version ready for asthma & COPD - the two most common lung conditions.

It’s been one year since Lungy launched (here’s the original ShowHN: https://news.ycombinator.com/item?id=34534615) - I was delighted with the response. One year on, we have re-designed and rebuilt the whole interface, and added a real-time, 3D soft body solver which gives some really cool interactions like blobs / objects that expand and inflate as you breathe.

My background is as a junior surgical trainee and I started building Lungy in 2020 during the first COVID lockdown in London. During COVID, there were huge numbers of patients coming off ventilators and patients are often given breathing exercises on a worksheet and disposable plastic devices called incentive spirometers to encourage deep breathing. This is intended to prevent chest infections and strengthen breathing muscles that have weakened. I noticed often the incentive spirometer would sit by the bedside, whilst the patient would be on their phone – this was the spark that lead to Lungy!

Since making the first version we’ve made exercises fully customisable (you can dial in exact timings for each breath phase), added new breathing indicators, learning modules, e.g. self-care for anxiety symptoms, and lots of new visuals. The UI is now much nicer. The free version gives you access to a new breathing exercise each day, whilst premium ($14.99 per year, $39.99 unlimited) unlocks the full library of exercises.

Thanks for reading - would love to hear any feedback!

Version 2.0 here: https://apps.apple.com/app/apple-store/id1545223887

I feel very strongly about this too. True art is a struggle to bring it into reality - you have to almost expand the space of possibility. It's easy to make derivative work, all the arguments about artists just interpolating other people's work are very weak - some artists do that, but good ones are truly original, even if they steal. It sounds a bit grandiose, but art can cut through the human experience to reveal some kind of truth. AI cannot do that because it does not understand the world or what it is to be human.

I think AI art is a complete misnomer and there is no net benefit of being able to generate images so easily, it just means there is more noise in the world.

Hey, thanks so much. Currently breathing doesn’t influence the sound, as it would also influence the detection of breathing which might set up a feedback loop if that makes sense. So we’d need to cancel out the generated audio - I’d really like to have that interaction though.

The medical device regulation process is long and quite difficult. I don’t think that’s necessarily a bad thing, but it is also an opaque process with expensive consultancies charging a lot for any guidance - this makes it very difficult for small companies without a lot of funding to even know where to start. I think just the prospect of having to go through registration does exclude quite a lot of innovative companies / products, which means patients are left with a limited selection of fairly mediocre products - the registration is very boring and difficult, but obviously necessary.

Pia (https://www.piahealth.co) | Remote / London, UK | Full-time (Short-term contract)

Pia are developing beautiful, interactive & evidence-based digital health technology. We're working towards creating a next-generation breathing platform for patients with asthma, COPD and long-COVID (https://www.lungy.app).

We're looking for a machine learning engineer with experience in computer vision, audio and multimodal ML models. Please get in touch with CV and/or links to projects you've worked on: hello [at] lungy [dot] app

Amazing - thanks! And for respiratory rate, are you just taking the chest rise and fall from the accelerometer? Would it be obscured by general x,y,z body movement or are you filtering it in some way?

It actually looks much smoother than the trace we get from respiratory inductance plethysmography...

This is really great! Thank you.

I developed a breathing app (https://www.lungy.app) last year - we're actually doing a study at the moment looking at how heart rate variability changes with breathing exercises. We're using a polysomnograph for gathering the data at the moment - using a Polar H10 may be much simpler. Is it possible to get the raw ECG lead data, or is it just BPM at each time step?

Exactly. Also, what is the intervention if the coaches cannot provide any advice or guidance[1] or manage the condition. I really think services like this should be regulated - I'm not sure why ADHD seems like a soft diagnosis and always slips through the net.

[1] (a) you acknowledge and understand that Coaches are not trained or licensed as medical professionals, and you agree to not request that Coaches provide you with any medical or healthcare advice or guidance, and (b) you agree to not rely upon any information provided to you by Coaches, content received through the Services or other Members when making important lifestyle, health, financial or relationship decisions.

I always wonder how this can be offered to patients when it seems like it should be classed as a medical device / service and regulated. It's intended for use in a named medical disorder and the guidance is very clear [1] (UK guidelines).

It is presented as having an evidence base with trained clinicians, but then the T+Cs state: 'Coaches are not trained or licensed medical professionals and are not qualified to diagnose, treat or manage any healthcare conditions, including behavioural health conditions.'

It makes me really sad to see so many companies like this - the services are always hugely expensive and target vulnerable people.

[1] https://assets.publishing.service.gov.uk/government/uploads/...

I agree, I think it's awful. It's a waste of everyone's time and insulting to people who have a problem and feel slightly vulnerable. People aren't just numbers or conversion rates, and especially in digital health they should be treated respectfully.

Well done for launching! I'm just very conscious of over-claiming and pseudoscience in this space...

Is this classed as a medical device / SaMD? Does it require FDA / MHRA approval or CE marking? Why is there information on your website about Fatal Familial Insomnia when it is vanishingly rare and has only been reported a few dozen times worldwide? Are those self-reported improvements in sleep and can you measure more objectively with polysomnography?

In the UK there's an app called Sleepio which is NICE-approved and prescribed to patients - is your app similar? One last question - how did you come up with the pricing ($63 per month)?

In the larger diameter blood vessels, it’s also due to relieving back pressure and finding a path of least resistance. This causes existing pathways to dilate and eventually join together to form a new channel / anastomosis (aortic coarctation is quite an extreme example - because of severe narrowing at the aorta existing blood vessels widen so much they cause notching of the ribs, which is visible on chest x-ray). Things like liver cirrhosis can cause back pressure and dilatation in a similar way.

Quite a while ago now, but I used to work at a tissue engineering lab in London. I was interested in how techniques from procedural design and architecture could be brought across, as actually the design of many of the TE scaffolds seemed very crude.

Blood vessels grow according to growth factors that are released in response to hypoxia, I wanted to try and model them as a space colonisation. I never quite got around to producing the actual structures, but some more details here:

https://www.rcseng.ac.uk/-/media/files/rcs/standards-and-res...