Hacker News Reader

Hacker News

Comment by aaron695 | Hacker News Reader

Full thread

aaron695

Pretty sure this is 'fake' and a old article.

Why is DNA better than silicon? Is it magic?

View on HN

Have you read the article, or are you just responding to the headline/title?

EDIT: Seeing as I got a downvote I thought I'd expand on my question. You asked:

  > Why is DNA better than silicon?
  > Is it magic?

The article says clearly what the advantage is of DNA over silicon - it's at the very top of the second page where it says:

  > ... biological computations could potentially
  > have vastly more parallelism than conventional
  > ones.

Secondly, the article isn't "fake" in any sense of the term - it's written and published by an eminent researcher in the field of computational complexity. It's a very clear exposition of the techniques used to solve NP problems with physical systems, including an analysis of why this doesn't really work, and why it's unlikely ever to work.

Yes, it's "old", although for some of us it's not that old, and given comparatively recent excitement about using slime-molds to plan "optimal" transportation networks, clearly it's still relevant, and not well enough known.

So my question still stands - you seem to have dismissed this quickly, and I wonder if you've actually read and understood it, or are you simply reacting to the headline/title?

> ... biological computations could potentially > have vastly more parallelism than conventional > ones.

I think it's all the same. There is no magic to biological?

> you seem to have dismissed this quickly, and I wonder if you've actually read and understood it

I think I did read it first time quite well, my problem is it's tiring to re-talk about old articles , but to be honest rereading things that logically don't make sense I don't do, so maybe I'm lazy.

There is no magic, and the article makes that clear. It might win on the constant because of the potential parallelism, but even that's not clear. But if it did win on the constant, the amount of potential parallelism is huge! It's just that we never seem to be able to take advantage of it "properly."

But the thing is that this idea keeps getting re-discovered, with people getting excited all over again. This article is a clear debunking of the hype, and has value in being able to point people at it and say "Look, it's been done, it doesn't work, and here's the explanation. If you've done something really new, then you need to make it clear why it's different from this 20 year old work."

Having said all that, I'm not sure what it is you're claiming doesn't make logical sense.

Lipton is legit. In page 1 though

> However, it does _not_ mean that all instances of NP problems can be solved in a _feasible_ sense.

Basically, this says nothing about the P vs. NP problem (as I'm sure many are thinking of when NP is mentioned). This is more about a non-trivial utilization of the massively parallel properties of DNA computers in order to speed up brute force computation.

Reply on news.ycombinator.com