skwb

Very interesting! Is there any consideration to making a navidrome plugin?

Since getting laid off in May and failing to find any jobs for ML in healthcare, I am working with a friend I met during my MPH to start a boutique consultancy to help hospitals deploy AI / health technology.

The HackerNews hug of death

Yes but you have: 1. specific explicit training and certifications 2. someone to yell at and who can be fired for non-performance

It's hard to describe, but it's felt like LLMs have completely sucked the entire energy out of computer vision. Like... I know CVPR still happens and there's great research that comes out of it, but almost every single job posting in ML is about LLMs to do this and that to the detriment of computer vision.

I'm a avid (hobbyist) photographer and I've noticed a TON of genuinely good 3rd party lenses (primarily Sigma and Tamron) and even 'fine' lenses at rock bottom prices (Viltrox, 7Artisans, TTArtisans, etc) for like $250. The conventional wisdom I've heard is that computer aided design has totally revolutionized this field.

I can only hope that projects like these help build better lenses for the future.

The fact that courts do strike down admins on violating the APA does, in fact, make it illegal.

Also, what would be illegal about the change?

At the *very* least you should be following the administrative rules act (requiring you to solicit 45 days for comments by effected parties) before making such a dramatic change in policy.

Courts absolutely love striking down EOs (of both Dems and Reps Admins) when they should have been following the administrative rules act.

Completely off topic to the actual article - but...

I really get irked a lot by seeing (overly obvious) ai generated images being used for stock photos. One of the reasons my wife and I still subscribe to a couple of newspapers is that the photography helps bring to life the story being told. Why where these photos taken and how do they impact the story. You don't get that same visceral emotional reaction with low quality cartoon images.

I don't disagree with anything you're saying necessarily, but a lot of people are conflating my statement of IRB exemption with having an explicit IRB authorization. I guess more to my point is people seem to be failing to understand the role of retrospective research (and I will easily concede that different institutions have different legal interpretations) and how it's an important part of research. Don't get me wrong, you absolutely still have regulations about what you can and cannot do with that data, but saying you need an IRB for everything doesn't match the reality that I've seen first hand.

That said, there's plenty of buying and selling of radiological images for industry development on the second hand market. Now where the line of "research" vs. "industrial" work is, well that's something I would leave to legal council. But as you said any sort of "altering" of clinical outcomes is a clear IRB is required zone like DL based recon.

I have published multiple papers in peer review journals where we were not covered by IRBs. That is not to say other research ethics standards don't exist, but that IRBs are simply one portion of the healthcare research regulations.

I've worked in MRI AI both in academic and non-academic centers, including technology that has received 510K clearance.

I admit it sounds pedantic, but I'm not discussing IRB *exemptions* that are sometimes required by an institution nor am I discussing BAAs. I was specifically talking about the specific IRB applications (which I've submitted and signed before) that the blog author was talking about. Yes, HIPAA and other state and local regulations also govern what you can and cannot do with the data, but that's not what I argued.

Sorta off topic but the FDA doesn't care so much about SAR unless you're directly programming the MR machine's pulse sequence. If you're just doing quantification of some brain structure for monitoring a biomarker, they're primarily concerned if your product 1) matches an existing prerequisite and 2) functionality that your product achieves performance that you say it does. That is why the marketing around most of the early DL / AI based radiology startups were focused on language for "study prioritization" rather than more specific claims.

I just pointed out the numerous exemptions to IRBs for research. People can literally make a research career and never touch IRBs.

The point I'm trying to make is that the blog is so narrowly defined that it doesn't match the reality of clinical research.

I mean tons of retrospective studies are literally "IRB exempt" (and highlighted as such in their methods) where no explicit consent is required or needed. Physicians doing case reviews don't need consent of the patient this work. Doing retrospective analyses for defining clinical phenotypes on patient data that has been aggregated isn't needed either. Collecting clinical MRIs to do deep learning doesn't require an IRB.

IRBs are only REALLY required when you are intervening in patient care or pose some theoretical risk to a human.

Some institutions still want you to submit approval for institutional data, but as a non-lawyer it seems that's much more of a CYA policy.

My wife and I went to Tofino (on Vancouver) this last summer where you can rent a boat for a tour of the coastal black bears. Very highly recommend it.

Use python and write my results in a CSV that I quickly import into R and do my fancy stats.

Tbf python's stats implementations can be garbage; the last time I checked you can't do multiple levels for hierarchical regression.

In one of their articles they claim that the microbiome is stable over 15 years, but this is categorized as "unpublished work" which leaves me with a large sense of doubt. I don't mind research that is in preliminary stages (as in conference abstracts, technical papers, etc), but the fact they are very naïve in how they approach scientific evidence makes me highly doubtful.

I take very significant issue with the fact their appendix A and B do not link to the original literature. When I was skimming the pitch deck, I thought they had done the experiments themselves, but it just looks like they copy pasted the results without proper citation. At best this appears as miscommunication, at worst misrepresentation of the work they have accomplished.

I work in medical startup space and the evidence we develop to prove our technology works typically has a higher threshold than what I've seen presented. I would want to see significantly more work done to demonstrate the scientific validity of this research. Do they intend to carry out any additional pre-clinical research (in vitro/in vivo)? Do they intend to do a clinical trial? It appears from their slide deck that they don't?

Vista AI | Full Stack Developer | Hybrid in Palo Alto, but also open to remote (US) | Full Time

I am a ML engineer at Vista AI, helping out in the recruiting process.

Vista AI is a venture backed company dedicated to making advanced magnetic resonance (MR) imaging techniques more accessible to patients. Our first clinically approved (510K cleared) product, Heart Vista, is dedicated to automate the process of acquiring cardiac MR images, which have historically only been accessible to patients at advanced facilities such as Stanford. Using machine vision, we automatically recognize, plan, and acquire the standard cardiac MRI planes, which often requires a highly experienced technologist to accomplish.

We are looking to build on this initial success by hiring a Full Stack Engineer who has good front end and backend experience to help build internal annotation tools and manage our clinical imaging data. A successful candidate should have experience in Node.JS, Vue.JS (or other modern Javascript frontend), solid DevOps skills (Docker, kubernetes, cloud infrastructure, SSL, oauth/keycloak), and good understanding of modern databases (Postgres/Redis).

Prior experience in medical imaging or other healthcare setting is highly desired!

To apply, please see the following link or send a Resume to apply@vista.ai.

https://www.linkedin.com/hiring/jobs/3605622950/detail/

I live in the Seattle area now, and I find it hilarious that the public transportation infrastructure around Climate Pledge Arena (which Amazon bought the naming rights for) is horrendously bad. My wife and I looked into taking the subway + somehow transferring to the monorail, only to see that service was super limited by the time the concert finished.

We end up parking nearby since it's cheaper/easier to do so. Completely the worst set of incentives.

this is where being socially awkward helps :)

"cardiac MRI which is not always widely available". This is mainly because the views of the heart are super hard to get (exams are done mostly at specialized centers like Stanford)!

I actually ended up doing my PhD on automating the cardiac view planes using deep learning.

FWIW I have seen lots of trash in Nature and Science.

Nobel peace prize winners can been frankly weirdly wrong on these topics when they overextend from their own specific domain. Think back to the AIDs crisis, where Peter Duesberg, an award winning molecular biologist, thought that AIDs couldn't possibly be caused by HIV. He was outside of his peer group that came to realize that AIDs is in fact caused by HIV. Experts can be wrong, but that is why science is not just a process, but an institution. Peer review is how we deal with these discrepancies, and "doing your own research" without expert mentorship is not a substantial substitute.

I have a PhD in bioinformatics, and yet I know that my personal expertise (and ability to consume) information about viral engineering is limited. I therefore have to rely on experts, and make a substantial effort to tune my priors to ensure I'm listening to the right sources. A single medium blog from a reporter without any editorial supervision is not an adequate substitute. When I have questions about how to adjust my priors for this subject, I have been in communication with colleagues whose expertise and knowledge are qualified to answer questions. And from all of this, there has been a general consensus from these scientists that while they cannot specifically rule out lab origin hypothesis, there does not even begin to approximate the amount of evidence we need in order to "prove" it. Remember in science we are trying to make claims that are by nature testable- if you cannot test a hypothesis, it's then just pure speculation.

If you want to waste your mental effort on "doing your own research" and making baseless speculation, fine, waste your time. Go off the deep end and find amusement of the sort of baseless conspiracy theory folks that appear on Joe Rogan. But do not for a moment bring baseless speculation into the realm of science. Too many people have spent too much time to waste it on people who cannot intellectually appreciate the differences between testable scientific hypothesis and a baseless speculative claim.

Ralph Baric invented modern coronavirology. He was Zhengli Shi's mentor, and published frequently with her in the past. He signed Jesse Bloom's and Alina Chan's letter in Science calling for further investigation of the origins of SARS-CoV-2:

I read the Science article; the way it was drafted doesn't really specifically rule any particular thing out however. It's just calling for a further investigation, and says that both hypotheses remain "viable", which is an extremely low bar in science.

Combing Jasnah's thread with another by Andersen (https://twitter.com/K_G_Andersen/status/1391507230848032772), it paints a picture that any explicit engineering seems a bit far fetched. I have yet to seen any specific responses to these critiques of lab engineered FCN site hypothesis. Instead, I see mountains of people who are not in a position to critically evaluate these claims. Science is both an institution and a process, and not everyone is equally qualified to evaluate the evidence.

FWIW I'm going to defer to the virology experts on this one. I know when I am out of my depth, and when to seek expert advise.

https://twitter.com/angie_rasmussen/status/13942570285586391...

The "viable" in the "Theories of accidental release from a lab and zoonotic spillover both remain viable" is doing a lot of legwork there.

Did University of Minnesota have any type of IRB? I sorta hope this can be the "come to Jesus moment" for academic CS to implement safeguards (such as IRBs) to prevent harm.

100%. PPI needs to hire a tech policy director whose knowledge can expand outside of section 230.

When I first saw this going around before seeing it on here, my first thought that it's still a bad take because how else are you form your inference without data collection? You're not just going to be given your conditional probability; you're almost certainly going to have to go out to collect the data to construct this (and the data collection being the regulatory item covered).

I think that the hot take machines that are twitter/online communities have a certain amount of mindrot where poorly researched topics make headline news concerns me (which tbf is in itself a hot take I suppose).