Do you have any evidence for that?

The modems that were affected in this attack were maintained by a company called SkyLogic that was still responsible for them, even though the property had been acquired by Viasat.

The Viasat deployed-and-managed modems were not affected by the attack and actually have been operating continuously in Ukraine since 2014.

I'm not so sure about that. A GEO satellite costs around $500-$600 MM including construction, launch, and insurance to cover an area like the US, whereas Starlink has so far cost around $800 MM (about $1 MM per sat) and still cannot provide continuous coverage over the US. By the time that happens, it'll have cost in the neighborhood of $1.4-$1.5 B.

Then, you've got to factor in that each SL terminal currently prices for around $500, and that's thought to be the "aspirational" price, not the current cost to manufacture.

All this for satellites that have around 7 Gbps of throughput, each, with beams the size of Texas.

Except it's really nowhere close to 20 Gbps, and also, the 610 Mbps demonstration used a Ball Aerospace (not Starlink) antenna and an entire satellite was dedicated to it.

Starlink's current capability is closer to 3 Gbps per satellite if using their terminal and each one costs about $1 MM to make and launch. Actually, their terminals use a bit under 250 MHz of bandwidth, so, their theoretical top speed is a couple hundred Mbps, which has to be shared between uplink and downlink. Not saying you can't connect to a remote datacenter with a couple hundred Mbps link, but also, it'll have cost SpaceX more than $1 MM to do it (once you take into account the ground segment).

This is mostly a PR move on SpaceX's part to distract from the performance of their network. Azure Orbital's first (and primary) connectivity partner is O3B.

If those measurements came in that order in time, then, yeah, that looks pretty suspicious.

So far as I can tell, this dump contains only the SHA-1 hashes of passwords and no one has figured out how to invert SHA-1.

The SHA-1 hashes of common, already-known passwords are available, so it's possible to invert hashes for these passwords. But, claiming that you can recover any of the passwords is wholly different from being able to confirm that a few well-known passwords were used by a segment of the population. Case in point: Of the ~420 million passwords in the leak, only about 7 million are in the top 55 board on leakedsource.com/blog/myspace, i.e., 1.6%.

It would appear that, if anything, this is really a list of email addresses from circa 2013. It could also be interesting to look at the distribution of passwords by looking at frequencies of specific hashes.

* Soul required for activation not included.

Hey, that's really helpful. Thank you!

Could you expand on this? In what parts of the world are you seeing this kind of work? What companies?

Why do you think ODE modeling is on the way out? Can you point me somewhere to start understanding "non-linear mixed effects"?

Thanks, buy you a beer sometime when I'm in the UK. ;-)

To be fair, the timeline proposes to get to a human brain not in a decade, but about 2030 -- or two decades.

I think it's impossible to say right now that the project will fail. Lab-based neuroscientists talk a lot about what we don't know, and how this is all pie-in-the-sky, but so what? The people in the labs will never by themselves figure it all out because they don't have a unifying architecture for understanding their data, generating new hypotheses and testing them. In short, the theory component is lacking.

For example, there are research papers where the authors have disabled the ability of an enzyme to become phosphorylated at a particular site (CaMKII at Thr 286) and then had these rats perform a standard test for finding their way through a maze. The result was that they had a harder time remembering where they were. Now, what can you really learn from that? CaMKII's ability to become autophosphorylated is important for forming memories..., and that's about it. There are so many layers in between cause and effect: The success rate of the gene transfer, gene expression, unknown variability in the formation of rat brains, and maybe, hey, maybe the rat was just having a shitty day and that's why it couldn't find the pad under the water. How does this help you generate new ideas for experiments? I don't know.

At any rate, when people say that there's so much that we don't know, this is partly what they're referring to: We have tons and tons of information, but no framework to hold it together and make sense of it.

The vision is that HBP will fill that role as a site for the assimilation of data into models that are constantly being refined and tested. They've spent years building the infrastructure necessary to gather standardized* data, build the models, run them, and analyze them. The idea now is to start taking in even more data, and just build things up one little bit at a time. Build a better cortical column, build several and hook them up together, build a rat cortex, a rat brain, a cat brain, a macaque brain, and then a human brain, learning along the way how to get it right and what the patterns are.

Make no mistake: The people working there are some of the brightest in the world. I see no technical reason why they won't succeed; I think it's only a matter of time and effort.

* Standardized data is very important because every lab tends to do things a little differently and there are so many variables that have to be pinned down. You might be amazed at what a difference 5 deg C can make in protein phosphorylation rates (about 20%, actually.)

I wanted to add on to this. I've had pretty painless experiences using the nvidia drivers under linux for several years, now. If you don't want or need the most up-to-date driver, Ubuntu has the older ones in their repository. 3D, window acceleration, and various display-, TV-, and HDMI-out work with no issues.

I wanted to especially point out that the latest driver release from nvidia fixes the Youtube blue-tint video problem, so you can again enjoy hardware-accelerated Youtube videos. Nvidia did this despite having essentially no financial incentive to do it (so far as I can tell.) Kudos to them, and I'm happy to say I'm a satisfied nvidia customer and Linux user.

I think that your reasons for wanting to be in a relationship align perfectly well with hers: You want to be in a relationship, in part to gain access to sex.

Now, don't get me wrong: As a male, I know this is a fundamental power imbalance. We want pussy more than women want dick, and the market price of pussy and dick accurately reflects that. But, at the same time, she did not say that sex should be a tool of manipulation. She said it was a "short-term hook" until deeper emotions develop, which I interpreted to mean that it keeps bringing you back to her.

What's wrong with that? If you're spending time with her and getting what you want, you've won. If you feel like she's manipulating you or using you, you always have the option to leave. I get the feeling that what you're afraid of is the emotional vulnerability that comes with investing in something risky. Yeah, good luck solving that one. At 31, I have not found an easy way.

Last but not least, I want to point out that women typically develop much stronger emotions much more quickly than men when sex comes into play. It's usually the men who walk away with fewer scars.

Thanks for your post. Your approach to getting more attention from women is generally in-line with mine, so I think you're right.

A small correction: There doesn't appear to be evidence that abstaining from sex increases testosterone levels. http://wiki.answers.com/Q/Does_masturbating_increase_testosterone_levels

Nevertheless, I end up feeling the same way after about a week. I'm more outgoing, a little bit more aggressive, and a bit more upbeat. I don't know what combination of hormones/neurotransmitters/receptor density makes that happen.

It's a cute story at face value, but I think it misses the fact that in every economic situation, there is a context which determines what is cheapest / most efficient to do. Even today, I would bet it's cheaper for the Chinese to use manpower rather than gasoline to accomplish the same thing.