Hacker Newsnew | past | comments | ask | show | jobs | submit | blix's commentslogin

I inspect and modify my own weights literally all the time. I just do it on a more abstract level than individual neurons.

I call this process "learning"


Fukuoka's desalination plant treats about 16400 m^3 of water per day. Assuming 3kWh per m^3 of water, this works out to a time-averaged power consuption of ~2000kW.

The osmotic power plant generates about 100kW, so it's about 5% of the total desalination energy requirement.


Ah, so a slightly more efficient desalination plant then.


Slightly more efficient, with less waste.


Pretty solid win-win


Depends on the CAPEX and OPEX requirements. If it is cheap to do, it could be a solid win, but if the plant requires a lot of capital, it might be cheaper to just take the hit on efficiency


Yes the brine could just be diluted wih gray water to reduce the environnemental impact without the energy recovery of the osmotic plant and the capital can be invested in other renewable with better efficiency.

That being said it's a first so it's a pilot project needed to have feedback on a real plant in operation and not just back of the enveloppe calculations and suppositions. Sometime you need to just build the thing to encounter problems, issues or non-issues.


According to this delightful overview [1] of the desalination plant, the capacity overall is 12000kW so that's definitely close enough.

1. https://www.niph.go.jp/soshiki/suido/pdf/h21JPUS/abstract/r9...


Why are we assuming 3kWh per cubic meter of water?


I did some cursory research and that seems to be a common estimate for modern osmosis-based desalination energy costs.

If you have a better estimate, feel free to supply it.


I wish this comment was more representative of my personal experience in science.

Instead I got PIs happy to say that weak evidence "proved" their theory and to try suppress evidence that negatively impacted "fundablity". The most successful scientists I worked with were the ones who always talked like a PR puff piece.


What field, may I ask?


Applied physics. I'd prefer not to get too specific. Most of my peers are working for the US DoD or DoE now.


I wonder if what you described is due to the money incentive?

I did theoretical physics (no money) and my experience totally matches what the other person described.


"Shocking" carries a meaning of a being sudden, surprising, or startling.

It's this aspect that is being challenged, not the emotional reaction.


Obama also created ICE as we know it today. And normalized drone strikes.


How are drone strikes any different than a pilot, a warplane, and advanced precision bombs/missiles?

Except they're cheaper to run and don't physically risk a pilot.


The issue is not really with the difference in impact between drone attacks and other types of aerial attacks, but with the dramatic increase in scale, resulting from reduced cost and risk.

It probably would have been more accurate to say something like "mass extra-judicial assasination/execution of individuals opaquely labelled as 'militants,' including US citizens, in foreign jurisdictions" instead of "drone strikes," but the latter is shorter and I thought would be understood as implying the former.


That appears to be an issue of policy not one of technology.

Because they'd more than likely target those same individuals with less precise weapons if not for the given alternative.


The technology enables the policy. If the cost and risk were higher, there would be fewer strikes.


They invaded two countries simultaneously (one landlocked). Then used secret stealth helicopters to fly a hit squad into an allied nations territory for one particular individual.

I don't think this is a fruitful debate but I doubt risk & cost are as much a determining factor as you'd like.


Using something similar to a benzene ring with spokes sticking out of it is absolutely a reasonable choice for depicting sodium hexametaphosphate in a schematic. This is actually a pretty common choice in scientific literature regarding this molecule.


If you look at it more carefully, it isn't really a benzene ring, though. It's got multiple layers to it.

And the neural net crossover is just wrong. Really, really wrong.

This is nominally educational content. Any way in which it is wrong is something that people can and will pick up on and form incorrect assumptions about which they will carry forward. It's not enough that it is vaguely sort of, if we're generous, isn't entirely wrong. That's not the bar for human work either.


experienced players who know their teammates well can reliably get 3-4s. if you only go for safe 2s against these opponents you will lose every time.


I've seen almost no one outside during the day all summer lol. Phoenix this summer has the deadest daytime sidewalks of any place I've ever lived.


My doctor told me after an intestinal surgery that I would never be able to eat vegetables again.


Do you really think the engineers building a robot to go into Fukishima have never once looked at the first handful of google results for "radiation hardened camera"?


A more generous reading of that comment would be that it’s other things not the camera’s that are causing the problem.

A CCD image sensor that’s lost 20% of its pixels could still be providing useful information, especially if you’re just trying to get the robot out. Other systems may inherently have issues long before that point.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: