IBM smashes Moore's Law, cuts bit size to 12 atoms | Hacker News Reader

IBM smashes Moore's Law, cuts bit size to 12 atoms

computerworld.com

39 pointsthenicepostr17 comments

Data storage density doesn't really have anything to do with Moore's law.

It's difficult to take this article seriously given the terrible grammar and obvious technical errors, often in the same sentence, e.g. "The researchers then combined 96 bits make [sic] one byte of data..."; presumably "bits" should be "atoms". Such errors make me question whether anything in the article is accurate or being interpreted correctly.

My understanding was that Moore's law states that the number of transistors that you get for a given cost doubles every 18 months. Hence I don't think any result shown by researchers in a lab can be said to "smash" it until a similar technique is used in production for chips you can actually buy.

Moore's law is one of the most misinterpreted laws :)

It states that the number of transistors per INGOT SLICE will double each 18 months.

Ingot (in semiconductor manufacturing) is a slice of silicon monocrystalline suitable for photolithographic and chemical processing. The main point is that while transistors are getting smaller, ingot slices are getting bigger.

Edit: I mean misquoted (misinterpretation is a result).

Yes, I agree it's often misquoted. I don't see where it says anything about ingot slices in his original statement (see eg. http://en.wikipedia.org/wiki/Moore%27s_law#History) though; I was thinking of the statement "The complexity for minimum component costs has increased at a rate of roughly a factor of two per year", etc.

"The ultimate end of Moore's Law is a single atom."

That's a bit depressing. Why a single atom, and not some smaller particle?

You may be able to store more than one bit in each atom.

From what I remember from physics lessons, going with more elementary particles might be possible eventually, but not in any foreseeable future - (most/some) atoms are fairly stable by themselves, while keeping a hadron-based structure stable will probably be very difficult.

Nice response. As for ultimate limits, see https://en.wikipedia.org/wiki/Bekenstein_bound. Of course, this bound doesn't really give bits per quark/particle limits - just an overall limit based on the spherical radius of the information-storing system.

I'm not quite sure I understand. Wouldn't a single atom as it is, the smallest particle possible?

I think conanite means subparticles, e.g. from an atom to a neutron.

Also bits and quarks etc.

This seems an outlier at best; Moore's Law is an approximation, and this falls outside it (as have many). I would say "smashing" the law would be to find some way of manufacturing that would affect the rate of change the 18 months figure to ... 3 months, or something radically different, in perpetuity.

FTA: The experiment was performed at low temperature: about 1 degree Kelvin, which corresponds to about -272 °C (-458 °F). The byte starts switching randomly about once a minute due to thermal energy (heat) at about 5 degrees Kelvin.

"low" indeed.

They proved such bits may exist, but have definitely not brought anything to market, nor even demonstrated a process to do so. They used a scanning electron microscope to fiddle atoms around, creating one bit. A science experiment, nothing more.

I would say that is more of delaying, rather than smashing it outright.

Moore's Law says we should be able to fit transistors into a certain amount of space at a certain time. Getting well ahead of the curve would count as breaking Moor's Law, I'd have thought.

Reply on news.ycombinator.com