David Petersen

Building companies in NYC

[David Petersen @typesfaster]

For people asking about the correct Epitalon dosage - no one knows.

When Khavinson started his studies in Russia, he was testing the hypothesis that pineal gland extracts worked in humans. He used a specific extraction process and obtained a mixture of amino acids. He called that mixture Epithalimin.

He administered 10 mg of Epitalimin for 10 days to different patients etc, and those studies lasted 6–15 years.

During the course of those studies, his team isolated the active ingredient of epitalimin, which is epitalon.

And that was administered at 1000 times less.

So 10 ug per day for 10 days of epitalon was equivalent to the 10 mg epitalimin studies.

But his multi year studies were still ongoing with the pineal gland extracts, so he didn’t change those. But regardless, the dosing for peptides is not really known.

[David Petersen @typesfaster]

@pitdesi i've never seen this once

[David Petersen @typesfaster]

@pitdesi i assume this is one of the things the labs need to spend $10-50 million in tokens for and then publish their results as something researchers should look at. I haven't seen anything like that.

[David Petersen @typesfaster]

Why haven’t we seen any useful breakthroughs in everyday health and nutrition from LLMs/AI?

AI has made great progress in math and essentially solved programming.

Nutrition /health could benefits from wet labs and real-world research. But there’s already a huge body of published research. You’d think LLMs could connect findings across studies to identify potential treatments and propose new experiments.

My guesses:

Too much health and nutrition research is garbage: bad data, weak methods or outright fraud. Maybe the LLMs struggle to tell what’s reliable?

The AI labs may not be putting money into this kind of work because they are run by young people who care more about programming and math?

[David Petersen @typesfaster]

@pitdesi There is 100 years of data from nutrition + health research, the AI could take data from these various studies and extract ideas for things that are likely to work from there. It has barely done any of that.

[David Petersen @typesfaster]

@ferenstein we do have a lot of studies like the predimed study (https://t.co/iTA90uXflh), and the various types of studies we used to find out that lead is toxic, etc. but i agree with your point - it may just all be too small for AI

[David Petersen @typesfaster]

@nikitabier X should really have a way to click on an image like this and make it full size/full screen

[David Petersen @typesfaster]

They are incredibly useful for giving people actionable day-to-day help tips like talking after every meal or this acne infographic chat made:

[David Petersen @typesfaster]

@idea_catalogue lol it was an incoherent post i don't know how it got 200k views

[David Petersen @typesfaster]

I still find it kind of staggering that 90% of all coding work done the last 50 years was kind of effort because a machine could write it now in 1/10,000th the time.

Surprised more engineers aren't having an existential crisis.

[David Petersen @typesfaster]

@realEstateTrent he's about to be fired lol. was he smart?

[David Petersen @typesfaster]

@typesfast No young people even 100 years ago in NM

[David Petersen @typesfaster]

@Nelson81_

[David Petersen @typesfaster]

@EvanZhao6 what is it

[David Petersen @typesfaster]

@t_blom correct UK still has all the cookie shit hahaha

[David Petersen @typesfaster]

@antoniogm Neither europe nor the USA has ever really been able to give the middle class dream to 3rd world immigrants/migrants.

[David Petersen @typesfaster]

@luigifoschini I realize that now

[David Petersen @typesfaster]

@josiezayner So interesting

[David Petersen @typesfaster]

@arcad100 good point I should view all my work as a training set for what is happening now hahaha

[David Petersen @typesfaster]

@w2k_ building cool stuff now is awesome

thinking about all the time I wasted over the last 20 years building software is a bit painful

[David Petersen @typesfaster]

@_observer_123 4 generations of not having enough children

[David Petersen @typesfaster]

@duocpro1 i think because engineers are happy they can build more cool stuff faster