HNHacker News
TopNewBestAskShowJobs

ephermata

221 karma · joined June 27, 2009

David Molnar

http://dmolnar.com

I previously posted under the account "dmolnar" as well, but this is my main account.

Please consider donating bone marrow. You could help someone beat leukemia. Learn more & join here: http://join.marrow.org/

submissionscomments
ephermata··on Ask PG & HN: Hardware VCs
You might talk with Mitch Altman, founder of Cornfield Electronics. Mitch has a successful hardware business now, selling the TV-B-Gone remote among other things. http://www.tvbgone.com/cfe_main.php

Mitch did it without needing VC. Mitch is also incredibly generous and approachable; we've worked together in creating and running a hacker space. He recently led a trip to China for people who wanted to learn more about manufacturing there. If you sent him a short blurb + video he might be able to offer some useful advice.

ephermata··on The future of UI will be boring
Dictating letters is still niche and will likely remain so, except for people who have severe RSI or other reasons to go through the training required. If I'm spending enough time at one sitting to do a letter, and if I have the investment in learning how to touch type, then I'm a lot faster at typing than dictating. Even if the dictation is perfect.

On the other hand, voice recognition for specialized vocabularies is happening. I've noticed a progressive replacement of "push 1 for X, push 2 for Y" with voice interfaces in support calls. My car has voice recognition for the sound system that actually works, e.g. "Play Track Shelter" plays a song of that name. While there are still freaky behaviors, these work well in that they don't waste my time when I'd rather be doing something else -- like having my problem fixed or listening to Icon of Coil. Because they are for specific narrow domains, these applications of voice recognition coexist with my typing, touch, et al.

Put another way, new UIs and input devices seem to work best when they have a path for incremental adoption. It's rare to see a complete overnight change. Even the iPhone, as amazing and revolutionary as it is, built on pre-existing experience with touch screen phones. (Maybe "reacted against" is a better term than "built on.") Plus the iPhone still has a keyboard, even if it's a neat touchscreen one instead of a hardware one.

ephermata··on The Original Tablet
There are a few pre-CES shipping devices with 12" capacitive multi-touch screens. The HP TouchSmart TX2 for example. These are more traditional tablet/notebook PCs, but with Win7 they do allow for "iPhone-like" gestures. The TouchSmart in particular is a mix between awesome and frustrating -- the screen is great and the multi-touch is fabulous, but it's too heavy, its battery life is awful, and mine just died for no apparent reason.
ephermata··on Google prefers to acquire startups that are pre-revenue
Dodgeball is one example where the team did not view it that way. The founders were excited at the time of acquisition with the directions they thought they could go with Google. After the acquisition, they found it difficult to get resources or traction within Google to improve Dodgeball. This led to them leaving.

Maybe today's teams coming in feel differently, since the Dodgeball story seems to be well known. Plus articles like this one.

That being said, there are plenty of examples where an acquisition by Google did take off. Android is the most high profile right now.

ephermata··on Ask HN:Places to apply for machine learning?
This is true, but there are plenty of other teams within Microsoft (or Yahoo or Google) that do applied machine learning. You probably won't end up with a publication in a peer-reviewed conference, but you will learn about the state of the art. Of course if the goal is improving your CV for application to a PhD program, that may not help much.
ephermata··on Stroustrup: What Should We Teach New Software Developers?
What do you have in mind when you say "computer science theoretician" ? Do you mean someone who publishes in the ACM Symposium on the Theory of Computing or do you mean something else?
ephermata··on All 32mil RockYou accounts hacked. Passwords were stored in plaintext.
What do you think about halting password puzzles?

http://crypto.stanford.edu/~xb/security07/index.html

ephermata··on All 32mil RockYou accounts hacked. Passwords were stored in plaintext.
Train our runtimes to detect and reject attempts to store user passwords in plaintext?
ephermata··on Twitter finally in the money with Google link
No. Google (and MS) can afford it, and the deal is worth it for the PR plus the head start on figuring out how to make money from the firehose. We have a few examples of people making money from Twitter -- the question now is how to follow that up so the platform itself can make money, and how its preferred partners can make money. This head start is worth the money, even if Twitter decides to open its firehose for free to all later.

If it doesn't work out, then the deal ends. I have no idea what the terms of the deal are, but suppose as a wild guess the deal costs Google $100 million over two or three years, then ends. That's not great, but Google can afford that for the chance of making lots more money. At the size Google is now, they need to find revenue opportunities and new markets measured in the billions to move their stock price, at which point $100 million sounds less of a big deal...yes, it's weird...

ephermata··on Ask HN: Examples where you make something people want, but are unable to profit?
There are a lot of broken bodies on the micropayment road. Millicent & Peppercoin come to mind to start. Maybe the world has changed since then, but if you were doing a company on micropayments and came to me, I'd want a clear answer as to what you would do that everyone else did not.
ephermata··on Tis the season for Grad School Advice
Yes, you're right, 2001 was a tough year to graduate. That being said, von Ahn's focusing on PhD programs, where the time to finish may be as short as 3 years, but more likely is in the 5-7 range. While these programs do pay a stipend, if you stay in the program for the entire time (instead of dropping in, waiting out the recession, and then leaving), the total compensation over those 5-7 years is likely much less than you would make in industry if you'd gone to industry straight out, even taking timing into account.

One key thing is that a "great" EECS grad student stipend, say the one offered by the National Science Foundation fellowship, is (or was) $25K. Most tech salaries are 2x or 3x that or more. A lot of grad students make less. So if you made $0 for the first two years pursuing industry, then landed a job at $50K, you've made up those two years. (Not really, of course, due to tax issues and the fact you had to defer the money, many grad students make less, etc.)

Another key thing is that if you stay in a PhD program for the full 5-7 years, the economy may change radically around you. So if you stick with it, you forego the higher-paying jobs available at that time. Of course, you could always just drop out and start a company or something at that time...plenty of people do.

ephermata··on Octopart (YC 07) Powers Product Search for EETimes
Congratulations!
ephermata··on Google Goggles
Speculation: Because those aren't as mobile as a phone, and because Android needs something to differentiate itself from the iPhone. First seamless turn by turn directions, now this.
ephermata··on Apple Bans Another Developer, 1000+ Apps Pulled
There's been some academic work on identifying these attacks recently. Most of the work I've seen crawls social networks directly. You could try contacting the authors to see if they will share their data sets.

Representative example: SybilLimit: A Near-Optimal Social Network Defense against Sybil Attacks Haifeng Yu, Phillip B. Gibbons, Michael Kaminsky,Feng Xiao http://www.comp.nus.edu.sg/~yuhf/yuh-sybillimit.pdf

A more recent paper looks at a data set from Digg to discover vote manipulation; the authors are interested in further applications and would probably appreciate your contact.

Sybil Resilient Content Rating D.M. Tran, B. Min, J. Li, L. Subramanian http://www.cs.nyu.edu/~lakshmi/nsdi09.pdf

ephermata··on Ask HN: Examples where you make something people want, but are unable to profit?
Visual Studio "Express" is free, but doesn't have many of the features of the for-pay versions. http://www.microsoft.com/exPress/download/ (Link also includes downloads for betas of Visual Studio 2010 & trial edition of for-pay versions.)
ephermata··on Ask HN: Examples where you make something people want, but are unable to profit?
Jury is still out, but free/freemium music services look hard. (People don't seem to want subscription services.) lala sold allegedly because they weren't getting to profitability fast enough, imeem struggled, and Pandora needed an act of the United States Congress to deal with sharp increases in streaming royalties.
ephermata··on Accessibility and the iPhone - a tale of woe
Trying to find the source of this: someone recently observed (maybe here on HN?) that a number of communications methods we take for granted in our everyday lives started as accessibility options. The example that struck me was text messaging, which allegedly began as a way for deaf people to use cell phones. Anyone know the reference?
ephermata··on Challenging the MLS Monopoly
I recently heard about the business of teaching people how to form syndicates to buy "short sale" homes and then flip those homes for a profit. (Short sale == sell house for less than the loan value.) If you pay $$ to a guru you can get standardized contracts, hand holding, etc. Example blog post from someone in this area: http://cawholesaledeals.com/blog/tag/short-sale-syndication/

Leaving aside the morality of these gurus...

Wonder if this would be a thin edge of the wedge area to start with aggregating data on house sales? People are already meeting in person and swapping information about where are the sales, which ones do they want to syndicate, so you have less of a bootstrapping problem. You could also provide ways for people to signal interest in a particular home deal, or provide a commission to people who bring in data on a home that is later successfully syndicated/sold even if the reporter isn't in on the deal. Maybe there are other approaches. Then grow from that segment to more real estate data if the model proves out.

ephermata··on On Testosterone: Driving a Porsche vs. Camry
In that case, a confounding factor is that the people in the club may be able to tell the difference between the outfits. So it isn't just the second man's confidence, it's that the people there know he's willing to blow a lot of money on looks. That may conceivably affect their reactions regardless of the man's confidence.
ephermata··on YC RFS5: An Accidental Case Study
Pocket-carry is qualitatively different. I have ideas all the time but don't want to carry around a backpack with my laptop. Once I need a backpack or bag, there isn't that much difference for me between a netbook and a "real" laptop.

From your description, the n900, a decent BT keyboard, and video display glasses work well. All those pieces can be carried in coat pockets. Might be a little hard with a rigid BT keyboard. I regularly carry the cables etc. for display glasses, a folding keyboard, and an iphone or touch pro 2 in my coat pocket without problems. Unfortunately, this doesn't work for the development I'm doing now -- I use it mainly for displaying and responding to e-mail & twitter.

ephermata··on Has sci-fi run out of steam?
Ian McDonald's _Brasyl_ is up there. Hard to find but worth it. Does suffer from the "quantum effects are magical pixie dust" disease but this can be overlooked.

Charlie Stross, _Glasshouse_. Yes, _Halting State_ is good too, but in _Glasshouse_ Stross manages to get his temptation to wink at the reader under control and as a result writes a classic. Has a funny throwaway bit about DRM, too: "For some reason the ancients encrypted all their records and threw away all the keys, that's why we don't know anything about them..."

Ken MacLeod, _The Star Fraction_/_The Cassini Division_/_The Stone Canal_/_The Sky Road_ (Fall Revolution Series). Yes, there sure are a lot of Trotskyists! in! Spaaaaaace! but there's also serious exploration of how technological change affects social structures on the personal and political level. Overview from Reason magazine http://reason.com/archives/2000/11/01/anarchies-states-and-u...

and that's just things that unambiguously classify as SF. If you make things a little broader, then you can include novels like the 2008 Hugo winner, _The Yiddish Policeman's Union_, which is classic Michael Chabon applied to alt history.

ephermata··on Has sci-fi run out of steam?
I actually thought "Eumenides in the Fourth Floor Lavatory" was "worse" than Unaccompanied Sonata. Still I wish he'd write more short stories like that...they're more engaging than the novels I've seen recently from him.
ephermata··on YouTube pulls a Hulu -- yanking API access from TV-connected devices
Why couldn't it be copyright concerns? YouTube has negotiated deals with different copyright holders. We don't know all the provisions of those deals. One provision might be "no playing videos on set top boxes." Instead of trying to explicitly mark which videos have this restriction and which don't, they could opt to just disallow TV viewing entirely -- and try to drive views to the XL product as a replacement for people who still really want to see YouTube on the TV.

That being said, the points made elsewhere in the thread about the line being blurry are valid. My phone (HTC Touch Pro 2) has TV out and explicitly lists YouTube as a video source for that TV out. Does that count?

ephermata··on Ask HN: Who's hiring (take 3)?
Plus, don't you have a product which may benefit from developers, even if they don't start working on that project?
ephermata··on RFS 5: Development on Handhelds
You could also just carry around a small foldable keyboard. I've been using this: http://www.wetkeys.com/BlueTooth-keyboard-Flexible-HID-Inter...
ephermata··on RFS 5: Development on Handhelds
For display, you can hook up video display glasses to an iPhone today, then use ScreenSplittr to mirror the display to the glasses. With some Windows Mobile phones, like the Touch Pro or Touch Pro 2, you get TV out "out of the box" with no extra software (do need a special cable). That gives you a 640x480 screen for ~$300 with the myvu crystal glasses (myvu.com). Does anyone know of an Android phone that supports this? There were reports that the G1 has the appropriate hardware but I haven't seen anyone get the tv out to work.

The sticking point is the rest of the interface. There's the BlueInput drivers for Windows Mobile, that lets you hook up a bluetooth mouse and keyboard. You could then use RDP to connect to a machine elsewhere and have it host your IDE and compiler. Almost certainly not the "right" way to do it but it's a start.

Multi-touch interfaces turn out to be hard with video glasses, at least on current phones -- you don't see where your finger is. Of course you could ignore the video glasses and just look at the phone screen, but then you have a small screen.

May be worth thinking through what kinds of development are best suited for a person walking around. Are you planning to camp out at a coffee shop and work for six hours straight, are you planning to make quick changes to an existing app "on the fly" as you think about them, or something else?

ephermata··on Mendeley, the-Last.fm-of-research
last.fm uses a plugin for your music player to record to which songs you listen, and for how long. You can then elect to expose this to your friends on last.fm or use this with last.fm's collaborative filtering to suggest new songs.

The idea of doing this for research papers isn't bad, but I wonder how many people will pick it up. Depending on the subfield, I already see people act secretively about the current direction on which they are working, even with close colleagues. Sometimes it's about fear of being scooped, other times it's about trying a different direction. That would work against doing a stream of papers read, even if you could elect to share it only with some people.

Maybe more of an issue is the network effect. To really show value, I think they will need to capture enough people in specific sub sub areas that new ideas result, leading to publishable papers. Getting those people on to the service is a chicken and egg problem.

ephermata··on My Nightmare Interviews With Google
The "how many X are in Y" is also known as a "case study question." From what I understand it's a standard practice in interviews for consulting companies, such as Bain or McKinsey. I've never seen one in an interview, but friends of mine who took that route obsessed over them.

As your interviewer said, these questions are asked just to see how you will come up with an answer. The actual number hardly matters, so long as you can sell it to the interviewer. Pushing you is part of the process, as well, to see how you hold up "under pressure."

If you want to be uncharitable, this is much like what a management consultant does -- come in, get some problem from a troubled company ("How can we improve sales by 20% in six months?"), develop a strategy, and then convince the client that they should do it. The interview technique makes sense for that context because it selects for people who can pull this off.

Actually, for a PR role I don't think this question is such a bad fit. You will have to think quickly on your feet, answer hostile people, and come up with diplomatic, plausible answers to crazy questions ("why does your product not work with Slackware?"). The precise answer is of course important, since if you're wrong you will be caught, but how you react is more important.

ephermata··on Ask HN: What's the deal with mailing lists?
Oh also, if the project happens to have an IRC channel, those can be excellent places to ask these kinds of one-off questions. Depends on the project of course but instant messaging is pretty good for the quick help kind of thing.
ephermata··on Ask HN: What's the deal with mailing lists?
Mailing lists are a way to collect high volume discussions on a topic in a way that can be quickly scanned and searched. I can route everything through GMail or other mail client, then use a single set of tools to manage my conversations and "tasks" that arise from those conversations. In my case, I've set up my mail clients to let me do most everything without taking my hands off the keyboard, just "shoveling" mail through via shortcuts. I also have mail filters that shunt different lists into different folders for searching later.

Web forums are fine for many people, but they drive me nuts. Too much clicking and too many pageloads, at least for the forums I have seen. This breaks my communication "flow" and leaves me feeling unwilling to spend any more time at the forum. It _is_ true that it is easier to drop in and ask a quick help question in a forum, as you point out, but that doesn't outweigh the fact I usually never want to visit the forum again. More fundamentally, my async store/forward communication happens in my email client, I don't like having to switch to other programs.

Now, don't get me wrong. Mailing list discussions are far from perfect. In particular it's easy for the entire discussion to be hijacked by a few hot topics. (Not necessarily trolls, although that works too.) This inflates volume and drives away people who would otherwise make great contributions. I've seen this repeatedly.

Really what I'd like is something that is to mailman what git was to version control -- a way for me to opt in to different "versions" of a mailing list potentially managed by different people but sharing the same core message flows. Something that would let me avoid needing to keep my own set of killfiles updated for each list all the time.

The Cypherpunks Distributed Remailer sort of tried this, but their focus wasn't on different filters. They never had that many nodes anyway. The other thing I've seen is people self-tagging posts with [<keyword>] in the title, but that is brittle and still requires me to create new filters on the fly.

← PreviousPage 3 of 4Next →