Hacker News Reader

Comment by tbrownaw | Hacker News Reader

So what are they using this software for, what's the proposed alternative, and what makes that alternative work better for their use case?

View on HN

When foundry came out the demo looked next level. I'm sure you could do much of the same thing with FOSS but IMO the problem that the NHS has been struggling with for multiple decades really looks like a lack of Technical Leadership due to the complexity of the environment and care needed when dealing with patient data.

Hopefully Palantir has the necessary skillset to navigate the political environment which involves developing a platform that: 1. protects patient privacy 2. supports needs of providers (e.g. hospitals, gps, specialists, DoH) 3. allows providers to use data to support their operations 4. allows NHS to use the data to improve patient outcomes and efficiency

This Foundry demo impressed me at the time but its a bit dated now: https://www.youtube.com/watch?v=uF-GSj-Exms

Actual data analyst from a hospital talking about what the platform achieves: https://www.youtube.com/watch?v=Ps47Azr2Jz0

That second video was fascinating, thanks for posting.

I’ve heard Foundry is not only insanely expensive, but to actually accomplish something comparable to the demos in your domain requires a huge amount of integration work to build it out in a way that locks you in.

They are currently trying to screw us for a 14% year on year increase (over 4 years)

and screwing us for licenses to run apps in "production"

> requires a huge amount of integration work

Oh no!..

Data integration is literally Palantir's business.

It's totally reasonable to be skeptical of palantir without knowing the exact product in question, given their record.

Healthcare systems often use Foundry for organizing data. It's a complex problem and Foundry has a good toolset for the job.

https://www.palantir.com/offerings/health/

As somebody that works in the biotech/health space, this page is not that exciting?

The bottleneck in drug development is not discovery; we have to test more hypotheses more efficiently, not generate more hypotheses. You don't need a product like foundry to have reproducibility or share pipeline templates; there are already free, scripting-language-agnostic workflow tools.

Healthcare use of foundry != biotech/bioinformatics use of foundry?

A former work colleague works in health ontologies. They are complicated and include EMT and ward staff using terms of art with inverse meaning.

Perhaps I misread your intent, belittling complexity in somebody else's information space (eg a function of multiple parallel legacy systems and organisational change) seems unhelpful. You weren't excited, maybe people on the management and health economics side were?

I think you misread it a bit. On the bioinformatics point, I was addressing some text in the link. I don’t mean to belittle people that work in healthcare, just suspicious how Palantir secured contracts.

Yeah, ontology is a key word. Its like the best concepts from things like Informatica and Pachyderm and Snowflake, structured in a semi-sane data security model and certified with all the fedramp whatnot you need to make cloud saas work for enterprise.

Yeah, it is dead boring until you hit the real world of trying to make a data request from an IRB which doesn't talk to the IT folks who don't really know the ins and outs of the data they steward but have their own forms and approvals and bespoke deidentification process, . . . once your org is in that state then Foundry starts looking really good. Taking months to sort out getting a csv into your free agnostic workflow tools wasn't fine before, but c19 really made the tool's strengths clear and people want to retain that increased velocity.

Like the iPod, if you are Cdr Taco it is "No wireless. Less space than a nomad. Lame." If you are a normal person struggling with meaningful data in the enterprise and see all the things packaged together tidily, then iPod economics happen.

Its possible I was too dismissive. This problem is hard because of the huge gap between the health/biology domain and the tech domain. Is there a reason you feel like Palantir is better equipped to close this gap? I think I'm jaded/cynical because there is an endless graveyard of bad software tools in this world.

> This problem is hard because of the huge gap between the health/biology domain and the tech domain. Is there a reason you feel like Palantir is better equipped to close this gap?

I'm not certain that it is better equipped than any hypothetical or specifically focused system. Given any part of it, I see a lot of products that can be composed into a similar offering. That misses the point though because the problems are socio-political in nature, not technological. It is expensive, which means that if an organization adopts it, everyone from the top down better get into alignment or you will waste a lot of cash. Internal alignment like that can be achieved without spending a lot of money probably maybe, but not likely.

It is also externally aligned a little better than IBM/Oracle (saw Watson, Deloitte "data democracy" etc) as a SaaS with training and consulting.

They were already contracted by NHS to monitor vaccine distribution and covid data in 2020, that contract was terminated and moved to Mozaic Services after public outcry over data privacy concerns. https://www.cnbc.com/2021/09/10/uk-ends-one-of-its-data-shar...

Did you read the article? Adult social care dashboard not vaccine data. Mozaic just handled migration and is not the host.

NHs FDP (Foundry) still has the vaccine data last time I checked.

Reply on news.ycombinator.com