pfisherman

I was around for the dot com boom and bust, and this does not feel similar. The issue with the internet was that much of the value came from network effects that were not there in the late 90s and early aughts when personal computing was desktop boxes with 56k dial up connections. Very much, “if you build it, they will come.” It was the mass rollout of cable modems and then smart phones that changed the math.

There is no cart before the horse here. AI is coming for you, not the other way around. The pessimistic takes are underestimating the impact by at least a couple orders of magnitude. Think smart phones as a lower bound.

I have no idea what capacity people here work with AI, but given my view and experience the pessimistic takes I commonly see on here do not seem realistic.

Could this also be attributed to rising interest rates, a giant tax increase (tariffs), and the highly uncertain - err I mean “dynamic” — operating environment caused by the current administration?

From my viewpoint, companies are in a soft hiring freeze so that they can maintain a cash cushion to deal with volatility.

Key question is the license they attach to model and weights. I have been seeing an increasing amount of releases in this space under non-commercial licenses.

I think companies in the space should either totally open source or not publish at all.

I can see publishing like this as achieving one (or more) of a several objectives:

1. Marketing software to for sales / licensing

2. Marketing startup to investors

3. Crowdsourcing use cases or product features from academia

Now here are the problems with those:

1. Selling software (exclusively) to drug companies is a terrible business model. Very low ceiling there. You can make more from one drug.

2. Indicates company focus is producing models and not drugs. See point one.

3. Computational labs want to release open source, so not viable to build on restricted tooling. Experimental labs may just be using to algo-wash prior hypotheses / biases.

Now weigh against disadvantage of letting competitors know what you are working on, how far you have progressed, as well as your methods.

You would not need much more than 2 megabases. The genome is not one contiguous sequence. It is organized (physically segregated) into chromosomes and topologically associated domains. IIRC 2 megabases is like the 3 sd threshold for interactions between cis regulatory elements / variants and their effector genes.

You mean Tesla will merge with Twitter / XAI, and market cap will soar to $50T on Elon’s promises of “AGI next year”.

They decided to try to be a pharma company instead of a diagnostics company. The microarray tech was obsolete, but they did have a lot of data. The problem is that being a pharmaceutical company is very difficult. I always question why they wanted to play in early pipeline rather than using their data to have an edge in later stage acquisitions. Perhaps they felt that running later stages of development and commercialization was too difficult, risky, and capital intensive; but if so then they probably had no business trying to sell drugs in the first place.

From my uninformed perspective this seem like a rather big and sensible lever toward rebalancing US national finances. The actuarial math must have changed significantly since the inception of the program, no?

See my comment above. Getting approval for rare diseases and expanding the indication to the common form of the disease is a well established strategy in pharma.

Also rare genetic diseases give insight into the underlying mechanisms and pathology of common sporadic diseases, which can be leveraged to develop new and better therapies.

Getting a new drug or therapy approved for a rare form of a disease and then expanding the indication to the common disease patient population is a well established strategy.

So what is the difference between moral ambition and hubris?

To me limiting oneself to “idealistic but not ambitious” with a focus on fulfilling relationships and positive impact on family, friends, and local community seems like a good way to go.

“To do evil a human being must first of all believe that what he's doing is good, or else that it's a well-considered act in conformity with natural law. Fortunately, it is in the nature of the human being to seek a justification for his actions... Ideology—that is what gives the evildoing its long-sought justification and gives the evildoer the necessary steadfastness and determination.” - Aleksandr Solzhenitsyn

Complexity is not just variation in chord progression, key, or melody.

Dark Side of the Moon is basically the same chord progression repeated over and over; but with different rhythm, tempo, arrangement for each song. The variation within the scope of the repetition and call backs to various melodic and rhythmic motifs at various points throughout is part of what makes the album such an epic and thematically cohesive listening experience.

Agreed on chord numbers and progression being the analysis that should have been done. For example, blues is mostly defined by a 1-4-5 progression and the ol 2-5-1 is pretty ubiquitous across time and genre.

Also, I think disappearance of 7th chords - major, minor, or dominant - is vastly overstated. Keep in mind that these are from guitar tabs so likely ignoring chord inversion / voicing / substitution taking placw to simplify notation. For example a B minor triad can be substituted for a Gmaj7.

Bm triad = B,D,F#

Gmaj7 = G,B,D,F#

Or if you want to be fancy a Bb/Gm can work as either Bbmaj7 or C7 depending on where you put it in a progression.

Running mechanics also greatly impact injury risk. One of the best suggestions I got was to drastically increase my cadence, which forced me to go from longer (more powerful) strides to shorter (lower impact) strides. My observation is it also shifted much of the load from my lower legs (knees and ankles) to my core (hips and glutes).

> I would never do such a thing, why are you?

LOL! Because I wrote a research proposal that was approved, got an IRB, and worked with a team of clinical staff to execute the study. That’s generally what goes on at major academic research hospitals.

> If Boston Scientific has an issue with a lead in one of their neurostimulators, they need to know who needs surgury to replace it and who doesn't.

This is an assumption you are making that is not valid. Please do yourself a favor and look at some of the links that I have posted in this thread. What I am telling you is not a secret. Many people have been writing and talking about for many years.

It all ends up in the EHR, which is a legal document and considered THE source of truth about a patient’s care. Without seeing you hospital IT system I can guess with pretty high certainty that you are getting manufacturer info and model number, and very sparse coverage off lot number.

The problem is that this is not enough to unambiguously resolve the identity of devices across databases, or even within the same database at different time points.

For example, mapping between model number and UDI is often one to many. Model numbers might be omitted or mutated across databases and time. Manufacturer and device names are most definitely mutated. Overall it is a nightmare of an entity resolution problem.

Which barcode are you scanning? Are you sure it the right one? Which field is the data from the barcode scan going into? What data is actually deposited into the field?

I invite you to do the exercise of tracking implant data of a case you observe (i.e. you have the MRN and first hand knowledge) through your clinic / hospital IT system. I think you might be surprised.

Anyway, these details like training clinicians to scan the right barcode, checking integrity, routing the right info into the right fields, etc are what I was referring to when I mentioned up front expenditures on IT and operations.

Manufactures are required to have UDIs on the label. FDA does not regulate the practice of medicine. Thus they cannot compel hospitals to record UDIs in electronic health records and they do not.

Source: I have published articles, organized symposia, and lobbied the US federal government about this subject.

Here is a relevant article (not one of mine :)

https://jamanetwork.com/journals/jamainternalmedicine/articl...

> At present, UDIs are not routinely or comprehensively captured, integrated as structured data within patients’ electronic health records, or available for clinicians, patients, and others to use.

UDI information is not captured at the point of care. Like if you try to go into EPIC at a hospital and query encounters or implantations by UDI it will not return anything.

I think that maybe you are having trouble believing that such simple and basic information is omitted from medical records. Like it is obvious from the perspective of anyone who touches data that we should be capturing this info. But I can assure that we are not.

Here is some reading material about the issue.

https://pmc.ncbi.nlm.nih.gov/articles/PMC9872505/

https://nestcc.org/nestcc-udi-playbook/

It is. FDA requires manufactures to display UDIs in packaging (or devices themselves) in human and machine readable format. But hospital IT systems are not set up to capture this information and they have no compelling (i.e. financial) incentive to invest in the IT system upgrades and operational changes needed to capture this information. You would think better inventory management would be enough of a selling point, but apparently not. I am convinced that it will take government interventions - like CMS requiring UDIs on reimbursement forms - to get it done.

No they don’t. I helped lead a failed effort to implement UDI capture at a major academic research hospital. To date only Duke and Mercy health systems have implemented UDI capture in a limited capacity.

The big problem is that implementation requires significant up front capital investment and operational changes; for nebulous / speculative ROI in terms of dollars. So organizational incentives are just not there.

Ultimately this is the type of problem that will require government intervention to solve. And that intervention will require the creation of incentives for organizations to implement UDI capture - similar to how it took an act of congress (Obamacare) to incentivize healthcare systems to adopt electronic health care records and protocols for communicating / sharing health care information.

Recalls get issued with reasonable frequency. The crazy part is that clinics have no way of identifying exposures (patients) or inventory impacted by recalls because US hospitals do not capture unique device identifiers (UDIs) in electronic health records or their supply chain systems.

Simple explanation. Musk has lost it. His long term drug (ketamine) use has resulted in a psychotic break, which we are watching play out publicly.

Another problem with measuring effectiveness that you are relying on patient (or caregiver) reported outcomes. There are no “objective” measurable criteria for depression, such as a biomarker on some lab test or survival at 5 years. Depression is assessed by asking people questions about how they feel, which is prone to all kinds of noise and bias.

A lot of the problem with studying psychiatric disorders is that you can’t just go around biopsying people’s brains; so we don’t have much of any idea what is going on at a molecular or cellular level. We just kind of know the behavioral tics that doctors can observe and what patients can tell us.

You can argue until you are blue in the face that people should not have interpreted what he wrote as offensive. But in fact a very large number of people did.

I dunno. Could be that the hiring practices at the time, or the team that interviewed him vetted algorithmic problem solving skill and did not run a vibe check on social or emotional intelligence. Because blasting out a company wide post / email disparaging and entire class of people (who are like 50 percent of the population), and then getting indignant about the entirely foreseeable backlash is indicative of extremely low EQ.

See the “Rooney rule” in the NFL. Hiring female sports casters, referees, etc.

Basketball is not football. It is a five man game where one player can have an outsized impact.

Superstars not only impact the makeup of their teams roster, they impact the composition of rosters across the league. When Shaq was in his prime teams used to carry extra big men to have more fouls to give.

The most recent iteration of the league was teams going for small ball with three point shooting wings to compete with the Warriors. Right now you see serious contenders in the western conference going for size to make it past Jokic. The next iteration will be the Wemby stoppers.

Before Wilt and Dr J you had Oscar Robertson, Pistol Pete Maravich, Jerry West, Rick Barry, George Milan, etc. The league has always revolved around stars.

This article was written by someone who has very little knowledge or appreciation of the history of the game. The NBA has always been a league of specialists revolving around a few superstars with a couple of “glue guys” thrown into the mix.

This year has been a first hand education in the benefits / efficacy of vaccines for my child. Like the differences in outcomes are so stark that the effect of vaccination is obvious to even a young child.

I have also heard anecdotes of very bad flu / respiratory illness on other continents from some of my coworkers.

There was no official shut down here. But the teachers and about 45% of the kids @ child’s grade level were out sick last week. So was definitely not business as usual either.