Disqus - Latest Comments for jdunck

Re: 30x500 | Challenge - Name your audience | 30x500 "Read their minds"

Jeremy Dunck — Wed, 13 May 2015 18:58:43 -0000

Ideas:

django developers
fitness newbies
power users without admin access

Going w/ the first one since that most clearly fits, despite some misgivings as to market size.

Re: 30x500 | Learn to Make a Method - Jess Mecham, YNAB | 30x500 Pioneers

Jeremy Dunck — Fri, 13 Mar 2015 05:05:56 -0000

Cliff notes: 3 Rules
1) Find big pain, fix it to death ("definitive guide to X")
2) Pick something people are talking about
3) Pick something related to your product.

...and then promote it, unless you have a built-in audience. Suggests subscribing to product-related email digest lists, then suggesting your related content to the digest's manager.

After you have at least a couple good ebombs, try to get guest posts on other blogs.

Re: Girls Who Code

Jeremy Dunck — Mon, 30 Dec 2013 02:50:47 -0000

The claimed result is "So if you want to invest two years in something that will help you succeed in business, the evidence suggests you'd do better to learn how to hack than get an MBA."

It's quite a stretch to suppose that the Forbes top whatever is a good stand-in for "succeed in business".

Re: Girls Who Code

Jeremy Dunck — Mon, 30 Dec 2013 00:47:39 -0000

That quote is pure Texas Sharpshooter fallacy. I have a point to make. How can I illustrate it. Oh, here's a way.

Re: Alex Gaynor -- Meritocracy

Jeremy Dunck — Sun, 13 Oct 2013 01:26:39 -0000

Here is a fun game: go down this list of cognitive biases and consider how many of them "meritocracy" lends itself to: http://en.wikipedia.org/wik...

Re: The San Francisco Rent Explosion

Jeremy Dunck — Fri, 19 Jul 2013 00:14:42 -0000

Liquidity drives pricing. Under demand, reduced liquidity drives prices higher, because it makes supply less elastic. A landlord wouldn't lower their price based on a lack of rent control, but would be forced to price lower because the prospective renter would have a better chance at a better deal (soon/later) with increased liquidity. Increased liquidity reduces the leverage a landlord has over a prospective renter.

Re: The San Francisco Rent Explosion

Jeremy Dunck — Thu, 18 Jul 2013 21:35:09 -0000

That argument only stands up if rent control applies to retired people only.

Re: Blogger vs Journalist

Jeremy Dunck — Tue, 02 Jul 2013 14:26:08 -0000

I didn't go to a J-school, if that's what you mean. I'm not sure how that's relevant, or why you asked the question.

The idea that bloggers are (only?) sources is the thing I disagree with. (You may not have intended to imply that bloggers are *only* sources, but if not, I'm not exactly sure what you're trying to convey with the piece at all.)

Given that a bunch of commenters (apparently) have missed the point of what you're saying, could you try saying it again more clearly?

The impetus for this rehash of the old debate is, in context - which forms of speech deserve shield law protections? If Greenwald being a blogger (or more despairingly, not a "real Journalist") is the thing that puts his head in a noose, we have gone badly off the rails.

The desire to draw a bright line and sweep aside the complexity of the question is damaging to the discussion.

What would journalism look like if it started over? Easy publishing, cheap distribution, no rentier privilege. We'd like an informed public. What should the laws be?

Re: Blogger vs Journalist

Jeremy Dunck — Sat, 29 Jun 2013 23:49:33 -0000

Dave, I think this post is harmful. Journalism is a commitment to an informed public, and a collection of practices to that end.

Blogging is world-wide publishing of a personal voice. These things are different and overlapping.

Some blogging is journalism, and much of what is published in papers is not journalism.

I think it's quite clear that Greenwald is a journalist. I don't think we need to throw bloggers (some of whom are jouralists!) under the bus to protect him.

I realize that our national news and political dialogs are so impoverished that a strict distinction might be useful in bringing clarity and justice to this situation, but please, please, understand that the "blogging vs. journalism" debate is useless, harmful, destructive to framing and proper reasoning, and revise or remove this piece.

I think it's clear that journalism, more and more, will not be done by the once-great major for-profit dailies, and that cable news is largely a farce. Where will the great journalists still practice their craft? That thing they can't not do? Blogs (and elsewhere, of course).

Re: Software Design: 80/20 libraries

Jeremy Dunck — Thu, 18 Apr 2013 01:58:39 -0000

Alex, you've always made me arch eyebrows and scratch head at your vehement opposition to global-ish stuff, but thank you for presenting a well-formed argument. It's of course true that what we make easy, we will do more of. But I think I see the point of view now.

I am skeptical that multiple entry points can really cover up the added complexity in the library -- while I have looked closely at SqlAlchemy, I think it suffers some of the same pain that twisted and boto do -- its layering and structure doesn't offer an obvious "easy" mode. I agree that the requests package succeeds here, but I do think it leaks a bit as well. And then there's the contributor problem - every time I've looked at SA, any interest in contributing fell away as I tried to make sense of which of the many layers my desire might fit within.

I think the single-library argument is a winning one, though - reuse is a primary problem in our work, and much is lost for there being so many mostly-good-but-no-great ORMs (and every other vertical you might care to mention).

So perhaps we can catalog some of the libraries that aim to be 100% libs, and succeed at "easy" mode, and don't leak too badly, and succeed at growing contributors. It seems like a tall order, but a good goal, too.

Re: Stomach ache (aka not enough time challenge) - Got to Start Somewhere

Jeremy Dunck — Sun, 07 Apr 2013 00:31:41 -0000

Being the oldest person in the room - me, too, most of the time. But you can't let that build up your expectations of yourself so much that it becomes a barrier. The older the field gets, the less important that will be, but you must remember that while you have lots of experience, it's OK to be a beginner.

"getting comfortable with being uncomfortable" So true. And I think you already knew that -- but this area of learning is new and big and might seem too much.

But give yourself permission to falter, to feel out of control, to get past it.

“He who asks a question is a fool for five minutes; he who does not ask a question remains a fool forever”

Keep asking.

Re: Finding a Project - Got to Start Somewhere

Jeremy Dunck — Sun, 31 Mar 2013 14:40:56 -0000

One other suggestion is to look at milestones and risks. This applies to project selection for all students, but I'll use the hardware sensor/data idea as an example for illustration.

There is some effort in implementing a data source, some effort in consuming that data, and some more in using that data for some end goal.

Is there a way you can apply your effort so that the overall project is successful despite partial failure?

For example, could you create a simulated data source so that you can demonstrate using the data for the end goal, regardless of whether you get the data source working? (You should be sure that the sensor *can* provide the needed data, but maybe it's useful to show the purpose without the concrete application.)

Similarly, if you're not working on this alone, you can look for opportunities to split work and parallelize. Perhaps you could agree on how the data would be presented, then one person can work on the sensor itself and the other person could work on processing and the end goal.

Re: PyCon 2013 and Codes of Conduct, more generally

Jeremy Dunck — Wed, 20 Mar 2013 17:04:09 -0000

A small note: most issues (such as the ones your colleagues suffered) go unreported. The Code is necessary not just because stupid people need them, but because it creates a space between the shitty choices of not saying anything on the one hand, or confronting bullies in the moment and on their terms on the other. The Code creates the space for these things to come into the sphere of discussion, and for posts such as these.

Re: Let's get a little louder

Jeremy Dunck — Mon, 04 Jun 2012 05:50:00 -0000

Robin - so what you're saying is: "lighten up"?

Re: http://dronejournalism.tumblr.com/post/15030607639

Jeremy Dunck — Tue, 03 Jan 2012 00:15:51 -0000

Aside from the nice hack of being able to use iOS via wifi, the wifi is also useful due to its bandwidth - you'll have difficulty getting enough bandwidth at reasonable frequencies without specific FCC provision - the wifi band is one of the few, obviously trading range for bandwidth. If you want longer range, you'll almost certainly trade bandwidth for it.

Inability to fly at night seems like a big deal. Consider switching the built-in cameras for IR to see if the night contrast might work well enough?

If you do pursue adapting the parrot to other control links, consider gnu radio:
http://gnuradio.org/redmine...

Re: http://dronejournalism.tumblr.com/post/14217785440

Jeremy Dunck — Wed, 28 Dec 2011 17:55:24 -0000

Another problem will be credibility of the markings. News choppers are so expensive that they are necessarily rare. Any jerk will be able to mark their drone as a news drone.

Re: Revenue Based Financing

Jeremy Dunck — Thu, 20 Oct 2011 12:06:32 -0000

I like the model, but one concern I have is the difficulty in stacking - with VC, dilution and preference are well-understood ways to manage multiple rounds of financing. What happens if I need multiple rounds of RBF?

Also, in your example that's 1.5x payback of current revenue but over 10 years, equivalent to about a 4% interest rate, so I assume you make that investment having been convinced that revenue will grow, resulting in a higher payback. How is coming to that conclusion (of growing revenue) different or similar to coming to the decision to make an equity investment?

Re: Dougal Matthews Home

Jeremy Dunck — Wed, 12 Oct 2011 05:36:52 -0000

A note to folks going this route -- doing a signal in-process means the instance passed into the signal handler will definitely exist.

But note that the "instance" which async_post_save receives is from pickle, not from the DB, and because of that, you will save races if you do anything in the handler which blindly assumes that the instance is already in the DB. (More clearly, a task may - and sometimes will - execute before the process which sent the task has closed the transaction which would, once completed, insert the row in the db. Thus, a task which assumes the row exists is buggy, possibly data-eating.)

So, consider being prepared to retry (so that the instance later exists, having won the race), or use force_update in instance.save calls so that you know for sure that you're updating the right state.

In general, I favor passing PKs to tasks rather than instances for this reason.

Re: Problems with AP’s new “linking” policy - 10,000 Words

Jeremy Dunck — Mon, 25 Jul 2011 13:16:14 -0000

"Imagine if we all just linked to the original reports from news outlets who have the stories first"
There are 2 issues here, neither of which have anything to do with value to the reader.

First, AP has long served as cheap filler content to attract eyeballs, and linking to originals would miss that "opportunity".

Second, a feed of AP-driven links would remove some editorial control; large portions of what used to be on-site content would instead be a (more obviously) automated shell game.

To be clear, I'm for it, except that I doubt it'll ever happen.

Re: Thanks Spotify! 40 invites for startups (UPDATE: All out! Next time, guys)

Jeremy Dunck — Thu, 14 Jul 2011 15:43:22 -0000

Jeremy Dunck jdunck gmail Founder http://prelaunch.Lendzee.com Yes, in US

Re: SWF Seeking Technical Soulmate

Jeremy Dunck — Thu, 07 Jul 2011 17:59:15 -0000

Very often, non-technical founders are not in a good position to answer these questions. I make no comment about Maren, but I've overheard some painfully-done interviews where the need is technical, but the screener is not at all technical, trying to bluff it.

Any good dev will see right through that, and in the end, the most likely outcome is a bad hire.

I've thought about ways to address this problem (there are *many* nontechnical founders trying to make a technical play), but thoughts aren't fully-formed. The closest I've seen, for now, is http://www.hackruiter.com/ - the folks doing that are smart and have thought about it more than I have. But they're bottlenecked on process right now as far as I know.

Re: Small business lending is in trouble - A Sack of Seattle

Jeremy Dunck — Thu, 30 Jun 2011 16:15:21 -0000

You know, I just saw an episode of Shark Talk where the pitcher had revenue but needed capital for inventory to fill orders.

The Sharks berated him for being an idiot and not taking a small business loan. Perhaps they are out of touch with current trends.

Re: Django at Scale slides | Brett Hoerner's blog

Jeremy Dunck — Mon, 27 Jun 2011 15:27:40 -0000

Re. slide 25/26 always .set over .delete:

I think race could be avoided by .add rather than .set, but it wouldn't avoid stampede.I assume if the backing object is deleted (so there's no appropriate value to .set), you .set a sentinel value?

Re. 28, "keep", if I understand right, this is to avoid going back to memcache on the same key within the same request cycle?
I think that's a good optimization, but since mc is fast, you'd have to be doing a good bit of load for it to be noticeable?

Re. 31, external connection pooler, is one recommended? Which have you tried?

Re. 33, .save of specific fields, this is the ticket: https://code.djangoproject.... It's been open forever, but people are actively working it recently.

Re. 34, instance.update isn't available from Andy's code; I guess you meant either update(instance, ...) or you have a custom descriptor.

Re. 35, "alter hurts", how does it hurt? Deadlock transactions? It seems to me that lots of apps would need to add more complex things than flags to otherwise-stable tables. You disagree?

Re: Are you too old or too late stage for TechStars? The most common TechStars misperception debunked - A Sack of Seattle

Jeremy Dunck — Tue, 24 May 2011 10:55:39 -0000

I think it might be possible to be too early-stage, but not too late-stage. ;-)

Re: http://bryce.vc/post/5392647542

Jeremy Dunck — Mon, 16 May 2011 13:34:52 -0000

I think this is generally true, and, coming from Dallas, wish more people realized it.

I think if you don't need VC, this is more true than ever.

But. What do you think about this line of reasoning?

====
The VC partner's time is limited.
In-person boards are much better.
The chance of a partner making a trip for just one meeting is low.
This combines to make location matter (perhaps more than it should) from the perspective of getting VC funding.
====

I wish it weren't that way, but I don't see a hole in the logic. Options seem to be, avoid needing VC (bad for you guys) or avoid living where VC's aren't (bad for most places in the world).