Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
kefka
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
80 ms
·
661.
▲
by
kefka
11y ago
> but you can't argue that college students are more likely to pay back loans Well actually, yes you can. Student loans are not dischargable in bankruptcy proceedings. That distinction is only shared with criminal court ordered pena
662.
▲
by
kefka
11y ago
My wife was subject to very confusing papers that resulted in a federal loan many years ago. To her (she is dyslexic), it appeared to be a scholarship or a grant. 1. She has no degree, yet owes money. 2. Cannot get a better job to
663.
▲
by
kefka
11y ago
Better question: Why are Federally subsidized educational loans at 8% interest, whilst the Big Banks get their money at 0%, or thereabouts? One invests in the future, while the other breaks it. You can probably guess which does which.
664.
▲
by
kefka
11y ago
The claim I hear when I mention this to feminists is that "Every other day is a men's day". Or I get the "But they are intimidating and I'm scared to go."
665.
▲
by
kefka
11y ago
Really? " The word 'password' on a computer screen is magnified with a magnifying glass in this picture illustration taken in Berlin May 21, 2013. " Because that somehow represents a data transfer pact between the EC and
666.
▲
by
kefka
11y ago
That kind of a situation seems pretty simple: request data upon review, and actively deny the paper until data has been submitted. Or shame them academically. You'd be surprised how far that goes.
667.
▲
by
kefka
11y ago
That's because all women from prior generations were "men": or that's what their pen names and ghostwriters were! http://www.webdesignschoolsguide.com/library/10-famous-femal... http://ne
668.
▲
by
kefka
11y ago
That's correct. They are a bit different. However, with many of the node-red-contrib modules, I can do things like "watch twitter for certain #tags, send them to IRC along with a semantic analysis. Also log to MongoDB. And any @ s
669.
▲
by
kefka
11y ago
I fail to see how this is different than Node-Red. Node-red already has a nice ecosystem, and extra modules allow usage of any npmjs modules to be used (in excess of 200k).
670.
▲
by
kefka
11y ago
Well, yes. Or at least until they shape up and no longer introduce an unknown failure mode to all products using those chips. The cost for a company to use those chips are too great now. And that result is purely FTDI's response. They
671.
▲
by
kefka
11y ago
No, having a cheap VPS handle automatable tasks isn't ubergeek. Setting up Node-red on a VPS, and logging a dozen IRC rooms you're interested in, storing in a MongoDB, and creating a search DB... over TOR is being an "uber-ge
672.
▲
by
kefka
11y ago
That's why nobody is addressing "FTDI USB serial compatible". That too is a thing, and the Chinese use 16 bit arm chips, and still get them out for about $1/ea. and then, if I use "compatible", am I still a cou
673.
▲
by
kefka
11y ago
That's the wrong question to ask. Instead, 'How do I know that the devices I bought up to 3 years ago don't have a "detected counterfeit" chip and get nuked by a future FTDI driver update?' Or, 'How do I k
674.
▲
by
kefka
11y ago
> Yes, they should have. And tested the function. The applicable standards actually demand it. I agree. I don't want to put counterfeits in my product. Now, show me FTDI's standards documents that indicate how to identify and
675.
▲
by
kefka
11y ago
Or better yet, you just stay away from FTDI, knowing that their capriciousness can lead to customers' devices being bricked or disabled in various ways. The CH340G is really a nice and cheap chip. And I'm starting to see it in mor
676.
▲
by
kefka
11y ago
But that's the rub: FTDI doesn't have any good process to identify fake chips, other than "install it in your board and wait and see". And if you've tried ordering from FTDI themselves, they routinely have shortages
677.
▲
by
kefka
11y ago
There's a term we in random ee blogs call this: being ftdicked If you run a windows platform and you introduce FTDI chips in your design schematic, you run afoul of using accidentally counterfeit chips. Digikey and Mouser both have had
678.
▲
by
kefka
11y ago
The answer is that Pluto isn't some sort of classification issue. The proper answer is that Pluto is a planet because of historical reasons (by fiat). By definition, it is a planet, and the asteroids in the asteroid belt are simply ast
679.
▲
by
kefka
11y ago
I'm doing a bit of tricks to speed up the performance. And I'm perfectly find that you question the performance :) I encourage you to try it out. My first problem/observation is that Haar cascades looove running on a GPU due
680.
▲
by
kefka
11y ago
Good point. You can say that the progress to each of those things is increasing, and that they are. But when we look at access to medicine, it is abysmal in this country. And when we had a decent reform (PPACA), the Republicans shit a brick
681.
▲
by
kefka
11y ago
One discussion on that topic is that Ted Cruz wishes to initiate a Constitutional Convention over various perceived issues. Given the recent contentious passed bills and judgements (PPACA, Gay Marriage, et al.), a convention of this type wo
682.
▲
by
kefka
11y ago
If the means of production were in the hands of the citizens, then there's no reason to make "profit". And if the production was completely automated, then everything would be easy to obtain and at very low cost. Now, we don&
683.
▲
by
kefka
11y ago
Poor people don't make enough money for companies to target effectively. Poor people also do not have the skills to make them employable- as in the minimum requirements for a human life are not equal to what their economic output is. A
684.
▲
by
kefka
11y ago
Absolutely. My problem was that of how to market something like this. I have no clue how to do so, and the 2 people that contacted me fizzled after initial contact. I know I did get it in Hackaday, where there was a a bit of a spat between
685.
▲
by
kefka
11y ago
That's what my github.com/jwcrawley/uWho project does: Facial recognition on an old computer with no GPGPU processing. Now , any of the CVDazzle "exploits" screwing up a LBP cascade on face detection works to subv
686.
▲
by
kefka
11y ago
You don't actually need a computer with my software. There's a button that you can easily load a video file and do the same classification on the video file. Just record a video of the front doors, and load it later on. You don&#x
687.
▲
by
kefka
11y ago
Sorry, I wasn't angry at all. I was at a stoplight and read what you wrote, and responded with my voice to text (android). :) And they are indeed different projects. I would dare say that they have much better quality, but I've no
688.
▲
by
kefka
11y ago
you'll get no argument out of being that I use OpenCV. It is the tool that I had, and works effectively the way I programmed it. Also, I do not have a general purpose GPU so I cannot do gpu-accelerated calculations. this model is indee
689.
▲
by
kefka
11y ago
I also built an Open Source (GPL3) facial recognition program as well, called uWho (github.com/jwcrawley/uWho). Mine doesn't need anything like CUDA or OpenCL, and runs on a 6 year old laptop at 1280x720@15fps. I ran it at ou
690.
▲
by
kefka
11y ago
My understanding was that the connection prior saw the cleartext of the final destination's packets. SSL was the way to make sure that end to end was encrypted so that nobody could read the data (mainly login/password info). I thi
More ›