Hacker News Reader

Comment by shanedanger | Hacker News Reader

"the author grossly does not understand the medical tests he had done and misinterprets the data"

As the author himself, I can corroborate your statement that I don't understand medical tests as well as a trained physician. :) However, the point of the post was to present the data so others could interpret / draw conclusions. I gave my observations and perspective, but I made every attempt to indicate the potential weaknesses in my experiment. I do hope people don't lose miss the forest for the trees here.

The EGFR thing is regrettable. Thanks for pointing it out!

View on HN

I'm sorry, but you've collected no data, only noise. Eating rice for 2 weeks would have produced similar effects. As to your subjective experience – it's probably 100% placebo.

If you want to call something an experiment, you should know how to conduct an experiment in that respective field. Your experiences only show how a product like Soylent induces magical thinking in its consumers.

If you want to produce a single data point with Soylent (that would hardly be relevant, as it would be a sampling size of one, but still, it would be one sampling of data rather than zero), eat nothing but Soylent for at least 6 months.

If you don't mind me asking, can you produce some food studies/experiments that are up to your standards, so I can get an idea of what one would look like?

Most of those published in the New England, for example. Of course, one can substitute a shorter duration for a larger sample if the results are ver significant with a large effect. But if you have a sample size of 1, and want it to be data, however insignificant, rather than noise, you better keep at it for a while.

So something is not an experiment unless you do it for 6 months? :)

I don't think you read the parent comment properly.

Of course it was "an experiment" but that doesn't mean you can automatically draw useful or even valid conclusions from it. If I ate nothing but peas for a day, my "experiment" wouldn't say much (if anything) about the effects of a pea-only diet on the human body, but if I did it for a year, I might get some useful data. Except that would still only be the effects of peas on my body, and may not generalise to other humans anyway.

You did not mention the placebo effect though. That's the largest weakness of such an experiment. Your brain may be convinced (without you being strictly conscious about it) that Soylent is good for you and trigger all kind of benefits that you assessed yourself.

That's why clinical trials are running in double-blind modes, to remove as much as possible the placebo effect. A self-assessed experiment is, and will always be, meaningless.

After reading the opening paragraph and seeing how excited the author was to try this I couldn't help to think about the placebo effext. With all the bogus diets out there they show results in 2 weeks and then fail miserably I can't help to think longer testing periods are needed.

Obviously you didn't read the whole thing; search for the word "placebo" and you'll find my discussion of it under the section called "Potential Weaknesses In The Data".

I read the whole thing but I did not notice it as being very prominent, while it's the number ONE "FAIL" point of your experiment.

Well, they can't use a placebo for that, you'll starve the placebo group.

You could provide a different meal replacement powder and compare the results (provided the subjects don't know which one looks like what)

We don't need to re-research nutrition starting from verification of macronutrients. The placebo could be an imbalanced mix of macronutrients. That way, you would expect malnutrition, not starvation.

> You could provide a different meal replacement powder and compare the results

Yeah, you could at least do that.

Reply on news.ycombinator.com