aantix

Pull requests welcomed.

Doctors are humans too. They don’t know the entire solution space.

Any diagnosis and action set of steps found, you’ll have to work with your doctor to implement.

Have you ever lived with a chronic condition for decades for which multiple doctors literally have zero answers for?

I do have the resources to validate.

I am smart. I am capable of understanding deeply technical subjects.

I have a team of smart doctors to present my findings to.

I’m looking for plausibility. Not the full diagnosis.

Look at the pricing page. The cost has plummeted. Compute is cheap.

https://dnacomplete.com/tier-selection/

Nebula offered specials for full genome sequencing.

They’re now owned by DNA complete.

DNA Complete prices it at 495. I paid half that for 30x.

“By analyzing your DNA multiple times, we can significantly increase accuracy. 30x Whole Genome Sequencing means that 100% of your DNA is sequenced an average of 30 times, thus providing even higher accuracy. Typical DNA Ancestry tests sequence less than 1% of your DNA.”

Feel free to read the original comment again.

Do not rely “solely”..

I had high blood pressure for 30 years. ER visits. Multiple cardiologists in SF.

I had a nodule on my left adrenal gland. Hyperaldosteronism.

I diagnosed this. Not the doctors.

The doctors missed this for decades.

No. Maybe if you abandon all logic.

With an LLM, we’re trying to efficiently search the entire space using as many granular, personal attributes as possible.

We’re trying to find plausible hypotheses. Especially those that haven’t been considered.

The final determination and action is still a partnership between you and your doctor.

It’s possible to now get plausible diagnoses through an LLM. For which you can further discuss with your doctor.

1. Get your DNA fully sequenced for $200 from Nebula.

2. Download your .cram file. 50gb+

2a. Or if you have a 23andme profile, just request to download your raw data. Use that output file.

3. Use WGS Extract to generate a datafile from the cram file of the SNPs and genotypes.

4. Zip that output file. Upload to a ChatGPT session.

5. This enables you to ask questions about your specific genetic profile.

Bring your findings to your doctor. Discuss their plausibility. It may surface a solution that they hadn’t thought of.

A new YouTube app/player, for my kids.

It allows us to control the algorithm. It’s all LLM translating to YouTube search queries under the hood.

Visually it looks the same.

The suggested videos come from predefined buckets on topics they love.

E.g. 33% fun math, 33% DIY engineering, 33% creative activities.

Video recommendations that have a banned word in the title/desc don't get displayed e.g. MrBeast, anything with Minecraft in it, never gets surfaced.

For anyone interested in using it, send me an email.

I'll put you on my list. And you can contribute ideas to our community Google Doc.

jim.jones1@gmail.com

The Rails support for multi-model, nested form updates is superb.

Separate entities on the backend - a unified update view if that’s what’s desired.

No need for any outside dependencies.

Curious, who do you guys consider is the leader in the space with real-time voice interactions, interruptions, etc?

MyChart only allows me to download a single pdf with all tests, records and observations in a single pdf.

Otherwise to download each individually, I have to go through each record, one by one, and hit download.

There are way too many. I hope this supports uploading a single, all inclusive pdf, of all visits/tests/etc?

Was it consistently misdiagnosed by dermatologists?

With the Apple Health app, I've set up imports from most of my the major lab providers that I've used. I get notified when a new result comes in.

I can export all of the records in to a single pdf - I airdrop that PDF to my Mac.

Is that single pdf importable? And does the system know enough that this pdf represents multiple tests/labs/blood pressure readings/letters/etc?

Is this based on census data?

I genuinely wish there was a map where you could see how many kids and their age groups are in a given neighborhood.

Our current house is great, but there aren't many kids in the neighborhood.

I understand this is sensitive information, so it probably doesn't exist.

But choosing a neighborhood with other families that are in a similar life experience is kind of hard..

Especially considering it seems that kids play outside much less, so that's less of a signal.

Thinking out loud here..

If the basic hypothesis is that “don’t bet against the U.S.” and that the U.S. long term, always go up, and I’m assuming most of us buy in to this hypothesis because most of us are probably holding an index fund for the S&P or QQQQ long term..

Looking at my portfolio, I just weathered the 2022-2023 storm without even looking. It could have been down 50%, it could have been 90%, I wasn’t selling. I’m all in for another 20 years.

Given that stance, why wouldn’t I just buy and hold a leveraged asset like TQQQ?

Fascinating.

What are the black swan events for those holdings - I assume they can get margin called?

Is TQQQ a long-term holding?

You could stomach the 80% draw down?

The only LLM agent I've seen who asked any sort of clarifying questions about design was Devin.

Feels so strange that the average investor can't outperform an index fund with low management fees.

A small investor is so agile - they can move in and out of positions. Why that agility can't be utilized to outperform a slow moving index fund, long-term?

How does this compare to Cline or Cursor's Composer agent?

I have a friend that has a Devin subscription, working on a large scale Rails app.

So far, he says it’s a net win. Saves him time.

From the sessions I observed, Devin even came back with some really good follow up questions for clarification.

Sample size 1.

Can it be combined with scheduled tasks?

E.g.

"Every month, log in to LES.com and pay the current balance. If the balance exceeds $500, alert me before paying."

Hyperkalemia is most commonly encountered in patients with decreased kidney function.

Purely from a cost perspective - imagine a 79 year old grandma.

Heavily overweight. She is already partially immobile. Pre-diabetic. She may have other conditions, further complicated by her weight. She's on a fixed income.

Which is more probable -

1) A dietary intervention that she attends once a week that revamps her entire daily consumption (but remember, she's on a fixed income) along with some intense exercise?

or

2) put her on a single medication that changes her tastes for sugary and starchy foods, reduces her cravings, reduces inflammation, and in turn, will make her lighter and more mobile.

It is a no-brainer for Medicare. This will save so many downstream costs.

You might as well not even publish the report if you're not listing the offending protein powders.

Not sure about the Go ecosystem.

I would say that Avo is by far the best Rails admin-like solution out there. It's beautiful.

https://avohq.io/

If you're looking to build a SaaS app, start with a professionally curated Rails template. It will save you months of development time. Payments/Auth/etc are built in and there are solid patterns in place to extend from.

Jumpstart Pro is great. https://jumpstartrails.com/

So is Bullet Train. https://bullettrain.co/