7 ms·

More on Lithium Batteries

90 points by jayhoon 13y ago

jared314 13y ago
> their theories (surprise, surprise) provided support for these hand waving models > These folk "knew" what they needed to obtain If I hadn't seen the resemblance between the username and the first name on the linked publications, I would have closed the page and written it off as a pseudo-science blogger fighting big corporate interests. I just wish the user info page had his real name on it, for confirmation.
- Dylan16807 13y ago
  What sounds pseudo-science about the first main paragraph, "It turns out that...electrode."? I don't understand your reaction to some complaints about non-tested models.
  - jared314 13y ago
    After reading too many blogs trying to substantiate their claims by quoting science out of context, I never consider the first paragraph as proof of anything. The second paragraph turned from promoting an idea to almost accusing others of bias or misconduct, in a way that almost mimicked a conspiracy theory. I can understand the complaints about inadequately tested models, because that happens. (And, as there are a lot of russian language posts in the blog, perhaps it is a cultural language style I am unfamiliar with.) But, if it wasn't for the username looking similar to the name on the paper, I would have written it off.
    - Dylan16807 13y ago
      Still, I think pseudo-science would use a quote as a springboard into something nonsensical, it wouldn't just say "these guys didn't do their job properly", it would do something like accuse them of actual conspiracy, or say that the answer was obvious because [inane blather].
      - jared314 13y ago
        "Nonsensical" and "inane blather" are relative to your personal knowledge and experience. Science and technology, to people outside of those fields, sounds like a string of nonsense jargon and technobabble. (That is one of the reasons why phys.org, university press departments, Congressman[1], and Fox News[2] can get things so very wrong.) My lack of a background in the detailed inner workings of Li-Ion batteries, or chemistry in general, prevents me from directly validating his statements. So, if it wasn't for the username looking similar to the first name on the paper, I would have written it off. [1] http://en.wikipedia.org/wiki/Series_of_tubes http://en.wikipedia.org/wiki/Series_of_tubes [2] https://news.ycombinator.com/item?id=6378495 https://news.ycombinator.com/item?id=6378495
        Dylan16807 13y ago
        The point is that he didn't make any statement there. So you don't need any field knowledge to realize he didn't springboard into blather, because he didn't springboard into anything at all.
  - ajuc 13y ago
    The fact that abstract says nothing about the results, and only complains about "established science".
- anatoly 13y ago
  I've been reading his blog for many years and can confirm that the resemblance is no coincidence and it is indeed that person.
- demallien 13y ago
  There was a honking big difference between pseudo-science and this blog post, and it was right at the start. Specifically: "So I decided to find out - and I did! " Followed up by a link to a published scientific article. Seriously, I don't know what your complaint is. There's a nice dry scientific paper, and a blog post with a more accessible tone. This seems like a pretty good mix of the different types of media (putting aside for now the issue of the paper being behind a bloody paywall).
  - jared314 13y ago
    See my response to Dylan16807 [1], as he had a similar question. [1] https://news.ycombinator.com/item?id=6388337 https://news.ycombinator.com/item?id=6388337
    - demallien 13y ago
      Your reply to Jared is talking about quoting out of context. There is no quoting in this case. The guy is simply introducing the research paper that he himself wrote and got published. Seriously I have no idea what on earth you're going on about.
MrMeker 13y ago
"It is all here http://pubs.acs.org/doi/abs/10.1021/jp406274e http://pubs.acs.org/doi/abs/10.1021/jp406274e http://pubs.acs.org/doi/abs/10.1021/jp406273p" http://pubs.acs.org/doi/abs/10.1021/jp406273p" (click) "Your current credentials do not allow retrieval of the full text. Purchase This Content Choose from the following options: $35.00 for 48 hours of access Members, log in with your ACS ID to see your reduced price." Every. Single. Time. I have ever tried to read a scientific paper, this is what happens. It's complete crap. The research was done in a national laboratory, under public funding. I should be able to read this. Fortunately, I am now a university student so I will see if the library can help me with this.
- dnautics 13y ago
  only NIH has open access requirements, and they only kick in one year after publication. Could ask Ilya to authorchoice it, which is ACS's open-access option (costs $800 a pop).
  - antihero 13y ago
    It's just a bunch of text and diagrams. Can't they stick it on a $5 DO VM or even Dropbox? Why do people bother with these crappy hosting services that require people to pay to see your content? I am honestly asking what these people actually offer other than hosting. I'm totally up for setting up a little web app for hosting scientific papers, but what would be preventing me from competing with these people?
    - testbro 13y ago
      Scientific publishing isn't just about putting stuff somewhere for people to read it. Publishers, through their parasitic relationship with academia, have kept themselves relevant through the critical mass of people continuing to use them. The purpose served by publishers isn't really the hosting of articles, it's the names of journals the publishers own which carries an implication that top-tier journals get better reviewers and therefore higher quality articles. As long as academia values impact factor and journal rankings, the parasite can subsist. Most academics tend to self-host preprints, or stick them on arXiv so the publisher as a hosting service isn't really the "value-add" offered.
frabcus 13y ago
This is great stuff, and I want more! Any other good blog posts about it? Or overviews? How are people reacting?
mmagin 13y ago
So, in essence, most scientists today are busy writing papers or pushing bits around in computer simulations rather than doing real experiments?
- marcosdumay 13y ago
  Well, that's true. Experiments are usually so expensive that you want to get the most out of them, so most people work discovering what experiments they should do, and what they do with the results, instead of just jumping in the lab and doing the experiments. Then, sometimes, people with that culture face a problem just like this one, where blind experiments are the fastest, simplest, cheapest route to understanding something. Except that nobody knows it beforehand.
Xcelerate 13y ago
> When computational chemists got involved, their theories (surprise, surprise) provided support for these hand waving models. These folk "knew" what they needed to obtain - and they tweaked their models until they got it: the models are complex; there are lot of knobs to tweak. Incredibly, even on rare occasion when they got it right they dismissed their own results, ascribing discrepancies to model imperfections... This is stupid. My research is molecular dynamics and ab initio studies of novel carbon anodes in lithium batteries. We don't "tweak" the model until it fits; if you did that you would be 1) an idiot 2) not a scientist. All models describe reality only to an approximation. How well this approximation works depends on how good your model is. In computational science, we have algorithms that can almost perfectly match experimental carbon anodes. The only problem? These techniques (short of someone inventing a quantum computer) take essentially infinite time to run on a classical computer. So you cut down the computational complexity of your model by making approximations (e.g. removing electron-electron correlations) until you get something that can run on today's best supercomputers. But you don't stop there! You have to actually assess these models and figure out how well they work! For instance, there's probably 20 or so MD models that describe water (TIP3P, TIP4P, etc...) Each model can accurately predict certain properties of real water, but is also very poor at predicting other properties. In other words, your model needs to fit what it was designed for and you will get good results. It seems to me that most people who bash simulations nowadays are experimentalists that aren't up to date on the state-of-the-art algorithms that provide very good agreement with reality (for some situations at least).
- andrewcooke 13y ago
  so why were they wrong? (or is the post incorrect?)
- koffiezet 13y ago
  To be honest, what you're saying is that to get a 'good model' - you have to 'tweak' these approximations it so they work well "to fit what it was designed for". Maybe you don't like the word 'tweaking' - but your entire post sounds that it is exactly that, make your model match (certain properties of) reality as close as possible. I'm not saying that in a dismissive/judgmental way - it is probably the most realistic way of doing a lot of things - but it pretty much is exactly what the original post says (although he is dismissive/negative about this). I don't know enough about chemistry in general to comment on this, but we do have batteries and they do improve over time - so some people must be doing "something" right. What surprises me more are his claims that we don't even understand (or understood) even the basics of lithium on a chemical level? Is this even true? What could be the impact of this paper on future battery development? Could computational models be improved using this data?

90 points by jayhoon 13y ago

> their theories (surprise, surprise) provided support for these hand waving models > These folk "knew" what they needed to obtain If I hadn't seen the resemblance between the username and the first name on the linked publications, I would have closed the page and written it off as a pseudo-science blogger fighting big corporate interests. I just wish the user info page had his real name on it, for confirmation.

Dylan16807 13y ago
What sounds pseudo-science about the first main paragraph, "It turns out that...electrode."? I don't understand your reaction to some complaints about non-tested models.
- jared314 13y ago
  After reading too many blogs trying to substantiate their claims by quoting science out of context, I never consider the first paragraph as proof of anything. The second paragraph turned from promoting an idea to almost accusing others of bias or misconduct, in a way that almost mimicked a conspiracy theory. I can understand the complaints about inadequately tested models, because that happens. (And, as there are a lot of russian language posts in the blog, perhaps it is a cultural language style I am unfamiliar with.) But, if it wasn't for the username looking similar to the name on the paper, I would have written it off.
  - Dylan16807 13y ago
    Still, I think pseudo-science would use a quote as a springboard into something nonsensical, it wouldn't just say "these guys didn't do their job properly", it would do something like accuse them of actual conspiracy, or say that the answer was obvious because [inane blather].
    - jared314 13y ago
      "Nonsensical" and "inane blather" are relative to your personal knowledge and experience. Science and technology, to people outside of those fields, sounds like a string of nonsense jargon and technobabble. (That is one of the reasons why phys.org, university press departments, Congressman[1], and Fox News[2] can get things so very wrong.) My lack of a background in the detailed inner workings of Li-Ion batteries, or chemistry in general, prevents me from directly validating his statements. So, if it wasn't for the username looking similar to the first name on the paper, I would have written it off. [1] http://en.wikipedia.org/wiki/Series_of_tubes http://en.wikipedia.org/wiki/Series_of_tubes [2] https://news.ycombinator.com/item?id=6378495 https://news.ycombinator.com/item?id=6378495
      - Dylan16807 13y ago
        The point is that he didn't make any statement there. So you don't need any field knowledge to realize he didn't springboard into blather, because he didn't springboard into anything at all.
- ajuc 13y ago
  The fact that abstract says nothing about the results, and only complains about "established science".
anatoly 13y ago
I've been reading his blog for many years and can confirm that the resemblance is no coincidence and it is indeed that person.
demallien 13y ago
There was a honking big difference between pseudo-science and this blog post, and it was right at the start. Specifically: "So I decided to find out - and I did! " Followed up by a link to a published scientific article. Seriously, I don't know what your complaint is. There's a nice dry scientific paper, and a blog post with a more accessible tone. This seems like a pretty good mix of the different types of media (putting aside for now the issue of the paper being behind a bloody paywall).
- jared314 13y ago
  See my response to Dylan16807 [1], as he had a similar question. [1] https://news.ycombinator.com/item?id=6388337 https://news.ycombinator.com/item?id=6388337
  - demallien 13y ago
    Your reply to Jared is talking about quoting out of context. There is no quoting in this case. The guy is simply introducing the research paper that he himself wrote and got published. Seriously I have no idea what on earth you're going on about.

MrMeker 13y ago

"It is all here http://pubs.acs.org/doi/abs/10.1021/jp406274e http://pubs.acs.org/doi/abs/10.1021/jp406274e http://pubs.acs.org/doi/abs/10.1021/jp406273p" http://pubs.acs.org/doi/abs/10.1021/jp406273p" (click) "Your current credentials do not allow retrieval of the full text. Purchase This Content Choose from the following options: $35.00 for 48 hours of access Members, log in with your ACS ID to see your reduced price." Every. Single. Time. I have ever tried to read a scientific paper, this is what happens. It's complete crap. The research was done in a national laboratory, under public funding. I should be able to read this. Fortunately, I am now a university student so I will see if the library can help me with this.

dnautics 13y ago
only NIH has open access requirements, and they only kick in one year after publication. Could ask Ilya to authorchoice it, which is ACS's open-access option (costs $800 a pop).
- antihero 13y ago
  It's just a bunch of text and diagrams. Can't they stick it on a $5 DO VM or even Dropbox? Why do people bother with these crappy hosting services that require people to pay to see your content? I am honestly asking what these people actually offer other than hosting. I'm totally up for setting up a little web app for hosting scientific papers, but what would be preventing me from competing with these people?
  - testbro 13y ago
    Scientific publishing isn't just about putting stuff somewhere for people to read it. Publishers, through their parasitic relationship with academia, have kept themselves relevant through the critical mass of people continuing to use them. The purpose served by publishers isn't really the hosting of articles, it's the names of journals the publishers own which carries an implication that top-tier journals get better reviewers and therefore higher quality articles. As long as academia values impact factor and journal rankings, the parasite can subsist. Most academics tend to self-host preprints, or stick them on arXiv so the publisher as a hosting service isn't really the "value-add" offered.

frabcus 13y ago

This is great stuff, and I want more! Any other good blog posts about it? Or overviews? How are people reacting?

mmagin 13y ago

So, in essence, most scientists today are busy writing papers or pushing bits around in computer simulations rather than doing real experiments?

marcosdumay 13y ago
Well, that's true. Experiments are usually so expensive that you want to get the most out of them, so most people work discovering what experiments they should do, and what they do with the results, instead of just jumping in the lab and doing the experiments. Then, sometimes, people with that culture face a problem just like this one, where blind experiments are the fastest, simplest, cheapest route to understanding something. Except that nobody knows it beforehand.

Xcelerate 13y ago

> When computational chemists got involved, their theories (surprise, surprise) provided support for these hand waving models. These folk "knew" what they needed to obtain - and they tweaked their models until they got it: the models are complex; there are lot of knobs to tweak. Incredibly, even on rare occasion when they got it right they dismissed their own results, ascribing discrepancies to model imperfections... This is stupid. My research is molecular dynamics and ab initio studies of novel carbon anodes in lithium batteries. We don't "tweak" the model until it fits; if you did that you would be 1) an idiot 2) not a scientist. All models describe reality only to an approximation. How well this approximation works depends on how good your model is. In computational science, we have algorithms that can almost perfectly match experimental carbon anodes. The only problem? These techniques (short of someone inventing a quantum computer) take essentially infinite time to run on a classical computer. So you cut down the computational complexity of your model by making approximations (e.g. removing electron-electron correlations) until you get something that can run on today's best supercomputers. But you don't stop there! You have to actually assess these models and figure out how well they work! For instance, there's probably 20 or so MD models that describe water (TIP3P, TIP4P, etc...) Each model can accurately predict certain properties of real water, but is also very poor at predicting other properties. In other words, your model needs to fit what it was designed for and you will get good results. It seems to me that most people who bash simulations nowadays are experimentalists that aren't up to date on the state-of-the-art algorithms that provide very good agreement with reality (for some situations at least).

andrewcooke 13y ago
so why were they wrong? (or is the post incorrect?)
koffiezet 13y ago
To be honest, what you're saying is that to get a 'good model' - you have to 'tweak' these approximations it so they work well "to fit what it was designed for". Maybe you don't like the word 'tweaking' - but your entire post sounds that it is exactly that, make your model match (certain properties of) reality as close as possible. I'm not saying that in a dismissive/judgmental way - it is probably the most realistic way of doing a lot of things - but it pretty much is exactly what the original post says (although he is dismissive/negative about this). I don't know enough about chemistry in general to comment on this, but we do have batteries and they do improve over time - so some people must be doing "something" right. What surprises me more are his claims that we don't even understand (or understood) even the basics of lithium on a chemical level? Is this even true? What could be the impact of this paper on future battery development? Could computational models be improved using this data?