Disqus - Latest Comments for Jay_Levitt

Re: User Interface Obsession.

Jay Levitt — Sun, 18 Oct 2009 16:05:08 -0000

Funny.. I was just looking at a Yamaha mixing board I've had in storage. It's very fancy, all the bells and whistles - 96 channels, LED channel names, motorized faders, programmable knobs, all sorts of automation and effects.

And then I turn it on and I am reminded: I last touched this board three years ago. I got an iPhone two years ago. Suddenly the Tandy-100-era blocky monochrome not-touch-sensitive display looks very, very dated. If punched-hole printer paper started streaming out the top, nobody would be surprised.

Re: On Competition

Jay Levitt — Mon, 25 May 2009 20:38:18 -0000

Agreed. I don't know why people are so paranoid about competition. It's GOOD. Here's why:

1. Competition breeds innovation
2. Innovation breeds success
3. I want to be successful.

Therefore, I want competition. (Likewise, I don't want intellectual property, which is a temporary monopoly.)

One of the key factors in AOL's early success was Prodigy. IBM and Sears spent millions of dollars on store displays and advertising, teaching the public what an online service was and why you'd want one. Suddenly, selling AOL went from "Well, do you know what a modem is?" to "We're like Prodigy, only better." That's much, much easier.

Investors,of all people, should understand this. Your business model is to invest in startups and wait for a liquidity event. Aren't you worried that someone else could copy that idea?

Nah, me neither.

Re: Dreams

Jay Levitt — Fri, 24 Apr 2009 13:04:53 -0000

Double post.. nice cache-expiration bug, Disqus..

Re: Dreams

Jay Levitt — Fri, 24 Apr 2009 13:04:12 -0000

It's not bad for its day, but you should really read _The Halo Effect_ - which points out the flaws of this and other "company X is successful due to practice Y" books (and confirms that flaw with a little "Where are they now?" update).

Sad reality: Management style only contributes about 10% to company performance. The other 90%? Yet to be discovered, and possibly random.

Re: Confessions Of A Pack Rat (aka My Document Retention Policy)

Jay Levitt — Mon, 26 Jan 2009 23:24:05 -0000

I can guarantee they don't have it :) In fact, the AOL mail system was, quite intentionally, never backed up. Because the cost of doing discovery on THAT, for every lawsuit that might have involved an AOL account, was huge. Instead, we relied on multi-site app-level replication and fault-tolerant hardware to prevent data loss. It never lost a single message, and the last full outage was in 1998.

I'm with Fred. Archive everything; with full-text search, you don't even have to organize it. If you're worried about your messages being taken out of context, what better solution than producing the context?

Re: What If Your Model Is Wrong?

Jay Levitt — Mon, 12 Jan 2009 11:04:19 -0000

What if models don't work - period? In the past few years, we've seen increasing evidence that:

1. Statistics predict better than rules engines (e.g. Google vs. old-style "linguistic" search engines)
2. We can't escape our biases, and our models and experiments reflect it
3. Even "validated" models have base assumptions, and we never know what they are (e.g. counterparty credit models that presume the global housing market never fails catastrophically)
4. See also behavioral economics, cognitive neurology, decision-making theory, etc.

In software, we've largely moved from a "prove correctness" mindset to an "acceptance test" mindset, because the latter is more pragmatic AND more useful. Maybe economics needs to do that? Forget the models. Figure out how we measure "bad", figure out what knobs we can turn, watch the numbers, and tweak the knobs.

Re: Always Treat Money Like It Is Your Own

Jay Levitt — Mon, 05 Jan 2009 13:17:12 -0000

"Skin in the game" is important, but I disagree that treating it like your *own* money is the right mental model. When I spend my own money, I'm making an emotional decision. Am I buying something that will give me enjoyment? Will I be happier tomorrow, or next year? Will the other kids start liking me now?

When I'm spending my company's money, it's a financial judgement. Will we get a good ROI? Is this the best use of our resources? Is it a long-term or short-term investment? It's a dollars-to-dollars comparison. (Though I've often said that "You can't put a dollar value on money.")

Spending investors' money as if it were my own would be very bad. We'd have secret tunnels in the office, laser tag, some really nice skis, and quite possibly some really nice ski resorts. But no revenues.

Re: GUIfy your Ruby apps with Shoes

Jay Levitt — Thu, 25 Dec 2008 12:14:23 -0000

The "buzzwords" example doesn't work on OS X 10.5.6, Shoes r1134; I get

Error in /Applications/Utilities/Development/Shoes.app/Contents/macOS/li... line 394:
(!) no such file to load -- json/ext/parser

I do have the latest json gem (1.1.3), but I think Shoes uses its own gem repository anyway..?

Re: What does a liquidation preference spreadsheet look like?

Jay Levitt — Sat, 06 Dec 2008 12:39:57 -0000

Sure wish I could download that into Excel...

Re: Why The iTouch Is Inevitable

Jay Levitt — Sun, 09 Nov 2008 18:59:33 -0000

I think you may have just invented the Palm Pilot.

Re: The Sorry State of Blogging Software

Jay Levitt — Sun, 02 Nov 2008 06:22:23 -0000

If you're going to write some code anyway, have you thought about pitching in on Radiant? It's got a lot of neat CMS-like features, a terrific template-tag system, a strong extensibility framework. While blogging is doable today, it's not yet frictionless. The next major release (0.7) is expected to be the "blogging release", and I'm sure they have a good to-do list. Personally, I think the big things it's missing are:

* Simple "new blog entry" screen; right now, you have to add a new page just like any other CMS
* Better admin-side navigation of categories/tags/etc; the current admin system isn't really aware of the semantics of that metadata
* There's a metaweblog extension, but it may need updating

Re: Adam @ Heroku. Punting

Jay Levitt — Wed, 10 Sep 2008 07:41:13 -0000

Weird.. I had the glitch too in FeedDemon, but I am NOT seeing any comments past the big dump at 3:25:08 yesterday... I wonder what's different?