mng2

Yes, it seems much more like a high-level overview than an in-depth tutorial.

HSBC has this really odd system where they only ask for the (e.g.) 1st, 6th, and 7th characters of your password. That implies that they store plaintext or something reversible...

I'd be very skeptical about that claim without substantial evidence of it surviving in the field. Especially if you use ROHS solder, which is more brittle than leaded solder.

There is an ESD diode in the design but it only protects certain pathways -- a zap could come from anywhere if the thing is bare.

There are 8 high-speed serial links on the front side of that board and probably 8 on the back. That's quite a bit of bandwidth.

Electronic CAD is tedium, honestly. Even the expensive stuff I use at work is tedious, which is why we have a PCB layout department.

I haven't used KiCAD for several years, but in the past I used Rohrbacher's Quick Library Builder [0] to generate parts. Otherwise making parts is indeed a pain.

I enjoy KiCAD's single-key commands, though.

0: http://kicad.rohrbacher.net/quicklib.php

Comes from jazz idiom. Cool people, basically.

This is not my field (just an enthusiast) but consider that a typical fusion plasma has a density of 10^20 particles per cubic meter. Even if your plasma is 1 cubic centimeter, that's still 10^14 particles! Thus a full-on EM simulation, where every charged particle interacts with every charged particle, is not practical. A full-scale simulation is still unrealistic even if you typically cheat by breaking the problem up into cells and ignoring the Coulomb interaction beyond cell boundaries.

Fusion machine design is a combination of hard-won experimental knowledge and theoretical guesswork. I'm skeptical of the current wave of fusion startups simply because they don't have the weight of the decades of experimental work that have gone into tokamak research. But hey, they might make a breakthough, who knows.

What I heard from a friend of a friend is that QA was not handled as well as it could've been.

Unfortunately hardware updates can't be delivered wirelessly.

We already had Loma Prieta in 1989. You don't see very many bricks in the Bay Area, compared to back east.

I don't think it's just the UK. At Berkeley, for example, the Math minor requires 4 lower division courses and 5 upper div.

https://math.berkeley.edu/programs/undergraduate/minoring-ma...

Four LUTs per slice, not six, right?

If you'd like to be even more disappointed, take a look at the GRE Math section.

>we're potentially only a few years away

The press loves to hype this sort of thing but I think if you talk to most scientists in the field they'll be much more circumspect about getting to commercial fusion. Personally I think it's at least twenty years out, and that's if everything goes well, which it will not.

No, it happens with my parents. My sister and I are many years apart though, and once I went to college there was only one kid in the house, so I can kinda understand.

Yeah, Kodachrome processing stopped about 5 years ago, since the chemicals were too nasty. I think Kodak still sells B&W reversal though.

I wouldn't be so trusting. In the best case they are real, or ghost shift production. In the worst case they are fakes that have all the safety features cut out. There's no way to know unless you open them up.

Nothing too exciting. Having RAM closer to the die is nice, but it's still DRAM. Incremental progress rather than a quantum leap, as usual. I see some advantage in customers not needing to lay out high-speed memory on their PCB, which can require multiple design cycles to get right.

They're achieving this using Intel's EMIB silicon bridge technology, but that was apparently already available to them as a fab customer.

I used to live in Berkeley and did the walk/BART/bus/carshare thing, and it was wonderful. I miss it a lot.

If there were, then fusion would be much easier.

Consider a reactor that consists of two colliding beams. Should work great, right? Well, the reaction rate is minuscule, so you're wasting a whole bunch of energy in the ones that don't collide. It's a numbers game; in the sun there's a process that has a reaction rate of something like 1 in billions of years, but it's okay because there are orders of magnitude more protons in the sun.

The point of tokamaks and stellarators is that the field lines are closed, and you won't have leakage out of cusps like in mirror machines. You prop up the energy distribution with techniques like neutral beam injection and electron cyclotron resonance heating, since the ions tend to dump their energy to the electrons.

In other machines, mostly inertial confinement devices, the claim is that a pulsed mode of operation will be effective. However this remains to be shown, and some work (see Todd Rider's PhD thesis) claims to prove that such non-equilibrium schemes cannot reach breakeven.

Yeah, this business tends not to be that transparent about where they source their whiskey. It ranges from the big boys (Diageo is doing it with Orphan Barrel) to the small labels. Most of the new brands just source MGP; which is not necessarily bad but it's not fun to buy bottles from two different brands that taste pretty much the same.

At that point you just call it a supercomputer. Private companies already buy time on government/academic supercomputers, so I don't know if there's a compelling opportunity there since your startup costs would be immense.

Definitely the first step is to get a cheap USB dongle. There is a RTL-SDR compatibility list[0] that you may find useful. Once you play with a cheap one you'll have a better idea of the parameters you can improve by upgrading to more sophisticated/expensive hardware.

0: https://www.reddit.com/r/RTLSDR/wiki/compatibility

Yeah, I've been using Jupyter with inline plotting and it's really quite nice. A notebook that's not tied to proprietary software (i.e. Mathematica) and which can run a full-fledged programming language is a marvelous tool.

I forgot to address this in my earlier comment, but you don't worry about the plasma escaping -- it is very delicate, generally orders of magnitude less dense than air. Rather, keeping it alive is the difficult part. If the plasma hits the wall of a fusion device in an uncontrolled manner, it will dump all its energy into it. If the plasma picks up too many impurities, it will radiate energy away in the form of bremsstrahlung. It is hard to keep the ions hot enough to fuse while there are many phenomena conspiring against you.

Quite right. The figure of merit for a fusion plasma is called the 'triple product' since it's the temperature, density and confinement time multiplied together.

The problem in general with pulsed systems is that they take a lot of energy to get them from idle to active, and the research machines have miniscule efficiency. The claim is often that "we'll scale this up and it'll be more efficient" but there are always unknown unknowns. Nature is a cruel mistress, especially in the fusion business.

My experience is similar. If I have coffee or tea after lunch, I have trouble falling asleep at night. However soda or iced tea seems to have no effect sleepwise.

Yeah, yikes. Jensen is really piling it higher and deeper.

Not with official workflows at least, and even then, you would somehow have to keep the PCIe block up during reconfiguration for compliance. I think it would be difficult to reverse engineer to that level.

The easier thing to do would be to delay enumeration.

I don't think it's supported for Spartan-6 though.