LifeIsBio

Had a math prof in undergrad that once said, “this guy” 61 times in a 50 minute lecture!

This line stuck out to me as well, but my follow up thought was different.

I’ve had friends who have been on cocktails like these, and one of them once said something like, “I’ve been depressed before, and this is not that. I’m not depressed. I don’t have the emotional capacity to be depressed. This is more like a total emotional blank slate.”

She was basically a robot for a few months. Incapable of really any emotions, including sadness, anxiety, frustration, etc. Suffice to say, she also didn’t have the emotional drive to push her towards positive things like deciding on how to spend her weekend free time.

Thankfully she’s changed her meds and is feeling overall better (if, admittedly, at the price of some emotional stability).

There's really not much to share. I rewrite the prompt each time, but here's was a recent one:

I have an image of a hand drawn workflow diagram. I’d like to turn it into a mermaid.js file.

(with the image attached)

That's an interesting idea. A lot of times what I'm drawing is a blend of what the code is versus what I want it to be post-refactor.

One of my favorite applications of multimodal LLMs thus far is the ability to:

1. Draw a DAG of whatever pipeline I’m working on with pen and paper.

2. Take a photo of the graph, mistakes and all.

3. Ask ChatGPT to translate the image into mermaid.js

Given how complicated the pipelines are that I’m working with and the sloppiness of the hand drawn image, it’s truly amazing how well this workflow works.

I'm usually pretty verbose with prompts, so I wanted to see what chatgpt would do with just a "Schumpeter's Creative Destruction"

https://chatgpt.com/share/66ff28e2-ea74-800b-a230-86d562f60f...

Nonzero, for sure.

Yep, I'm in the rare disease space. "impossible" is pretty appropriate.

It's tricky. On the one hand, it's obviously not appropriate to be flippant about patient privacy. On the other, it's clearly that advancements in human health are being hindered by our current approach to (dis)allowing researchers access to data.

I want to second this. It seems like document chunking is the most difficult part of the pipeline at this point.

You gave the example of unstructured PDF, but there are challenges with structured docs as well. We’ve run into docs that are hard to chunk because of this deeply nested and repeated structure. For example, there might be a long experimental protocol with multiple steps; at the end of each step, there’s a table “Debugging” for troubleshooting anything that might have gone wrong in that step. The debugging table is a natural chunk, except that once chunked there are a dozen such tables that are semantically similar when decoupled from their original context and position in the tree structure of the document.

This is one example, but there are many other cases where key context for a chunk is nearby in a structured sense, but far away in the flattened document, and therefore completely lost when chunking.

Just to add to the list of this Jim Simons did and funded, he also established the Simons Foundation Autism Research Initiative (SFARI).

"SFARI’s mission is to improve the understanding, diagnosis and treatment of autism spectrum disorders by funding innovative research of the highest quality and relevance."

SFARI in turn funds a lot of foundational neurological and rare disease research, since autism is such a common phenotype.

The paper kinda leaves you hanging on the "alternatives" front, even though they have a section dedicated to it.

In addition to the _quality_ of any proposed alternative(s), computational speed also has to be a consideration. I've run into multiple situations where you want to measure similarities on the order of millions/billions of times. Especially for realtime applications (like RAG?) speed may even out weight quality.

I read this article when I was in grad school 5 years ago. Absolutely love it and talk about it to this day.

It really makes me frustrated about the ways I was introduced to statistics: brute force memorization of seeming arbitrary formulas.

Hey, HN! Maybe not your typical startup announcement here, but I recently left my job as a bioinformatics engineer to start a company called Lodestar Bio.

We are addressing challenges faced by families of children with rare diseases who are seeking a diagnosis, and our solution is a two-sided marketplace for rare disease genomic insights.

On one side, we will offer children who have a rare disease—and an inconclusive whole genome assay—another chance at a diagnosis. A majority of families who order a whole genome test do not receive their much needed diagnosis and are rarely provided with clear followup options. On the other side of the market, we will use the genomic data we collect to identify orphan drug leads, which we will sell to biopharma clients who are creating personalized medicines.

I'm happy to chat about any questions or comments you have!

About a year ago I did some work collecting interesting blogs from HN users and shared it here:

https://news.ycombinator.com/item?id=32291993

It is exciting! As others in the thread are saying, the cost of individuals with severe rare diseases are also very high.

Here’s a recent attempt at quantifying the costs across all rare disease:

https://chiesirarediseases.com/assets/pdf/chiesiglobalraredi...

That’s exactly what happened. :)

This is a reference to: https://news.ycombinator.com/item?id=36012360

Here's a thread where I fed all of his questions to ChatGPT-4.

https://news.ycombinator.com/item?id=36014796

It seems like his graduate student did him a great disservice by feeding the questions to 3.5

The game “20 questions” is probably the hardest I’ve seen chatGPT fail.

What’s interesting about the game is that, at first pass, there’s no ambiguity. All questions need to be answered with “Yes” or “No”. But many questions asked during the game actually have answers of “it depends”.

For example, I was thinking of “peanut butter” and chatGPT asked me “Does it fit in your hand?” as well as “Is it used in the kitchen?”. Given my answers, chatGPT spent the back half of its questions on different kitchen utensils. It never once considered backing up and verifying that there wasn’t some misunderstanding.

I played three games with it, and it made the same mistake each time.

Of course, playing the game via text loses a lot of information relative to playing IRL with your friends. In person, the answerer would pause, hum, and otherwise demonstrate that the question asked was ambiguous given the restrictions of the game.

Regardless, it was clear that chatGPT wasn’t accounting for ambiguity.

I haven't seen anyone mention Anvil[1] yet, but it lets you "Build web apps with nothing but Python." and is lovely tool that I've successfully used for a handful of side projects.

But as someone who feels most at home with Python, I always love to see new competition in this space.

[1] https://anvil.works/

I’ve run into this exact situation multiple times. Searching the whole history would be revolutionary.

I've found that people have wildly different definitions of "systems engineering", but this one is lovely. And also, very close to my own. ;)

https://jessimekirk.com/blog/whats_a_systems_engineer/

I've been using Resh[0] for the past 6 months or so. A rich and queryable shell history is a massive boost in day-to-day productivity. The syncing described here is a pretty cool feature.

[0] https://github.com/curusarn/resh

I recently came across the concept of Gambler’s Ruin in Taleb’s “Skin in the Game”. It’s… definitely written in Taleb’s style, but I’ve noticed myself relating all sorts of things back to the book recent. Worthwhile, overall.

https://en.m.wikipedia.org/wiki/Skin_in_the_Game_(book)

You and 4 other people have been added :)

I've got all the data pulled down now, so if you have any specific ideas for additional questions, feel free to share!

Some of the links are definitely down. If one link in a profile was active and "interesting", I just included all their links.

In [1]: html.unescape(users_table.find_one(name="graderjs")["about"])

Out [2]: 'https://github.com/crisdosyago<p>cris@dosycorp.com'

It looks like your about changed more recently than my BQ data pull?

I actually tried to do that about two years ago, and ran out of steam about 20% of the way through. It was a lot. Even this shorten list took a surprising amount of time!

If at least one site was active, I included all the sites scraped from the bio.