glimcat

Tweaked it a bit for better readability. Text size, spacing, width, a few colors, background texture...maybe a few other things.

http://pastebin.com/DMURgDsv

Yale students - not enough coffee, or actual zombies?

Be explicit about what forms of (re)distribution are allowed instead of going for a blanket license. Where appropriate, also be explicit about what is not allowed.

It shouldn't say much more than "you give us the right to use your content to fulfill the services you ask us to provide, and you have the right to remove your content at any time."

Conditions inside Amazon don't necessarily reflect the state of every business in the industry.

My only real qualm with this is whether to bill them my consulting rate or to put up with something well below the salary I would expect for ongoing stuff.

But if it's short and interesting, it usually falls into "I don't care" and stays there.

Premature optimization.

It's not worth worrying much about what will happen when shit hits the fan if the fan isn't spinning yet. First, figure out how to get some power to the thing.

Labs, office hours, networking, research opportunities, independent study, structural constraints, etc.

I love the whole movement of making lecture recordings available online for free. Ready and open access to learning has long been an ideal of academia. But it's not likely to topple the ivory tower any time soon, there's just too much that it can't do.

There's also plenty that it can do but isn't yet, leaving many areas to target if you want to do an EDU startup. "Video lectures plus computer-administered homework" is not the endpoint.

More important if you're going through HR. Less important if you're networking.

Also, I know and know of gobs of people in CS who were educated in physics.

> Make sure to have some kind of online portfolio (github is fine), a place to write about your experiences (a simple blog), and make sure to talk about the difficulties you had (plus how you solved them).

I think the critical thing is to document your experiences for your own use. It doesn't have to be in any particular form.

Online portfolios, open source, blogs, etc. can work very well - but they're just modes of expression. Some people take to them very well and use them very effectively, and that's great. But not everyone needs to have a coding blog. It's just one arrow in the quiver, and you should focus your time and attention wherever it will give optimal results for you.

"Death by ironic circumstances. Time of death chosen from within the allowed range such that hazards to bystanders are minimized, then such that irony is maximized, then randomly."

Solving for maxima in the Law of Universal Perversity should do it.

This for me is one of the hardest parts. The stereotypical obsession with "The Idea" is usually devoted to finding one which will make it magically rain money. Finding something that will keep me interested and motivated for more than a year - that's HARD.

If you don't have a viable product, it doesn't matter.

If you have a viable product, you can sort it out.

Starting a company is about 1% interesting stuff, with the rest going to grunt work, marketing, meetings, etc. ad nauseum.

Staying within academia is up to about 25% interesting stuff, sometimes more, with the rest going to teaching / outreach / departmental involvement / begging for grant money. If you enjoy teaching, interesting stuff can be 40 or 50 percent, sometimes more.

If anything, the IP workflow for collaborations bridging the gap out of academia needs to be improved. Take the horror show that is the US patent system and convolve that with university bureaucracy - it's not pretty.

Why before what.

It's usually inappropriate. For a blog, the feature I want to see is interesting or useful information presented clearly and efficiently. Adding other stuff usually just elevates the noise floor for no benefit. Ditto for unfocused writing.

The initial "why" is often "I want them to come back to my site" (often contributing to other goals like conversion). You don't need some ridiculous badge system for this. The best and first callback for a blog is good, relevant content and the promise of more.

Much of the time, the trick is to avoid tripping the primate anti-cheating instincts. "We are doing X." "Oh, okay then."

Yes. "Premature optimization is the root of all evil." Get it working first, then worry about whether it works well enough.

Before someone gets to the "effective autodidacticism" regime and for a not insignificant period afterwards, they are not likely to be very productive. Meanwhile, helping them make effective progress is a major time sink for more experienced personnel.

In a big company, you have a very big load with a lot of momentum behind it and a lot of people are responsible for pulling the load. If this guy pulls really hard and that guy can barely pull at all or gets tangled and pushes it in the opposite direction...it keeps moving.

In a startup, you're trying to juggle an ever-heavier load while making everything up as you go and trying to figure out what route your on, while everyone is riding full gallop and trying not to break their necks. Throwing someone into that who can't handle it can be quite disruptive. At best, it can be inefficient.

Maybe you'd learn fast and that period wouldn't last long - but if so, you should just do it and then try to join a startup (or whatever).

Finding a mentor will greatly reduce the potential gap between where you are and a useful level of skill, but that should be something you look for independent of involvement at a startup. As a stopgap, look into attending developer meetups in your area and look for organizations like ACM chapters. This can give you people to turn to when you get stuck. Eventually, you can help tutor others and learn even faster that way. Maybe it's what leads to developing a good mentoring relationship too.

Congratulations, Nest!

You're officially successful enough to be sued by entrenched non-innovators who took out overly broad patents so that they would be able to milk anyone who successfully built and marketed a product.

Honeywell makes damn near every wall thermostat I've ever seen, so it's not like they haven't had time to think about this application area.

The question was overly vague, which is why the first thing I said was "it depends on your definition." The broad estimate is on the first line.

It depends on your definition. Something like one in a couple thousand is probably reasonable, or a few million.

If we're talking competent and capable of independently producing professional-grade output, you'll see another reducing factor of around 1:100 or more...so already down to a few tens of thousands. Once we start breaking that down into subfields like data mining or computer vision, the number of people in the field can get rather small.

Feature tracking on an object which stands out clearly against the background? Using morphing filters on the segmented object? Not that innovative.

Now I do like the particle trick. You see this in some video games to highlight objects of interest (e.g. World of Warcraft). It makes it much easier for players to locate them and zero in on them against background clutter. Probably useful for AR if you have a good application case.

However, the potential is heavily constrained until you get displays capable of "non-interfering annotation" - which is to say, you need a good HMD and not a phone. Not that you can go out and buy "a good HMD" by this measure, but that's a big thorny mess out to the side of the software concerns.

I think we can dismiss that as rounding error.

Yes. The obstacles are primarily cultural, not technical. You need to impedance match both simultaneously in order to break in, and ??? to seriously disrupt the entrenched players.

Good thing it's not called Hackerbook then.

That's also a rather problematic conclusion to draw due to (a) sample size, (b) overgeneralization, and (c) the survey was framed as "will you buy"...so maybe they just don't play stocks, or have to pay off medical bills, or what have you.

I suppose it remains to be seen whether they'll sell as well at $249 and as a separate device vs. $149 and readily available as $100 and only needing one for both purposes. I suspect they'll sell, but at reduced volume and excluding major potential demographics.

They might do better to work towards building such an ecosystem for Windows, particularly given the wide range of uses and high level of interest which were demonstrated. "Warranty support" has not been high on the consumer values list in the existing ecosystem, so much as "cheap accessible cameras" and "does cool stuff."

Close the office door and disable the NIC.

Sounds like a problem which can be solved via CSS hacking.

I like a little (cheap) abstraction in my device development. I can always go back and bitbang in Assembly later once I'm certain what I want it to do.

I also think that reprogrammability is a significant feature for some of these IOT applications, so it may make sense to maintain compatibility with the Arduino stack longer-term if possible - and since it's open, you can do so while using your own board.

Thanks. I knew they included Bluetooth 4.0, but Low Energy was left out of the specs page. I actually still can't find any mention of it except in CoreBluetooth - their site organization is a bit painful if you're not there to buy.

There's also a demo implementation for Android by one of the radio manufacturers - not great, but maybe usable for initial development. This is much more significant on iPhone because it diffuses into the market more rapidly than a single model of Android handheld usually will.