jplewicke

There's also HireFire, which has support for scaling all your Heroku process types, not just your web processes: http://hirefire.io/ .

I've used them for a few projects which honestly have never actually needed auto-scaling, but they're very easy to use overall and at $10/month are way underpriced. They do have a slower polling frequency (1 minute instead of 15 seconds) though.

I would actually guess that the timing is mostly driven by the vesting schedule from the Salesforce acquisition. They announced the acquisition in December 2010, and it closed fairly quickly thereafter. I think 2 years is a standard vesting period after an acquisition, so now is a good time to leave unless he wants to make a career of it at Salesforce.

You see stories all the time like "Founder X leaves ____ 2 years after acquisition by ____." There's one just from today about a Tweetdeck founder: http://techcrunch.com/2013/05/31/iain-dodsworth-leaves-twitt... .

It's been around for a few years, and I'd definitely recommend it. It's very comforting when you find yourself saying "Wait, what if I need to do refunds or dunning emails? Oh, that's already handled." I've saved many hours of development time by starting with it as a base, and would recommend it to anyone writing Rails-based SAAS.

It has a few wrinkles though, mainly from being actively developed in 2008, with mainly bugfixes and Rails 3 support coming since then. The main ones I'd highlight that tripped me up:

- Expects payment gateway authentication credentials to be stored in a YAML file in source control instead of loaded from the environment.

- A number of the controller actions trigger immediate email sending, which can cause customer-visible 500 errors if the email sending fails and is a less reliable way of making sure the email gets sent.

- Use Stripe API v1 instead of the current v2.

- A lot of the customer email templates are kind of boilerplate-y and would not win the Patrick McKenzie stamp of copywriting approval. I believe they're loaded in from the gem rather than from the app.

My main piece of advice would be to vendor the entire gem right off the bat -- you're likely going to be tweaking several different parts of it, and there's almost no active development that you'd have to incorporate.

Most importantly, Bitcoin was the world's first distributed cryptocurrency without a centralized currency issuer. The blockchain was and is a very important advance over the previously existing Chaumian cash cryptocurrencies.

Thanks!

It sounds like one of the Bitcoin alternatives is scrypt based, and they are putatively getting a roughly 10x speedup from GPUs: https://en.bitcoin.it/wiki/Litecoin#Scrypt_Proof_of_Work . Is that a new development, or is it in the neighborhood of how GPU-resistant it was supposed to be?

I wasn't inconvenienced by it, but I remember myself and everyone at Dartmouth receiving a single email inviting us to join Facebook at the same time somewhere in the early winter of 2004. It wasn't a viral action like "your friend has added you", but was instead a direct invitation to sign up. I remember my roommates and I all immediately signed up and wasted the first of many hours on there.

I think the email was from another Dartmouth student, so I assume they just paid someone to send the entire campus an email.

They order from Applebee's secret menu, which is centered around obscure cuts of locally raised meats that are prepared slowly using sous-vide and finished with pan-searing.

Going to tack on:

5) Quantitative portfolio management. This usually involves managing a portfolio of stocks, either long-only or long-short. People usually come into this from either a heavy math/tech background or from doing academic finance research. Skills include portfolio optimization using linear/quadratic/etc. programming, machine learning, simulation and backtesting. Some companies that do this include:

- Acadian Asset Management - AQR Capital - Goldman Sachs Asset Management - Janus InTech - MDT Advisers (discl: I work there) - Renaissance Technologies Institutional Equities Fund

Thanks! I'll look forward to it.

These are a few things off the top of my head that I'd love to read a blog post on:

- SAAS Pricing when marginal costs are fairly high. For example, how would you price a service that offered human translation of a company's social media posts into N different languages?

- Accounting / corporate structure for bootstrapped SAAS companies: do you use an LLC, and if so things to take into account for cash vs accrual accounting, pass-through vs. corporate taxation, guaranteed payments vs. distributions, finding a good accountant, etc.

- The 8,000 word "stake through Bitcoin's heart", or at least the part that talks about how and why banks are going to stop taking Bitcoin exchanges as clients. There is a surprising amount of doublethink going on to allow people simultaneously believe "Bitcoin and traditional banking are natural enemies", "at least one Bitcoin exchange will always be able to access the banking system", and "the nominal price of Bitcoins can only increase in the long run."

With the right answer to this question, the original comment would need to be revised to "I have a friend who worked for a now-bankrupt marketing company until the Google anti-spam team found out about said company."

On the other hand, he's unilaterally refused to disclose actual audited returns for the funds that he's been involved with. This is very convenient for marketing his books and views, but creates an inaccurate impression of how profitable his approach really is: http://falkenblog.blogspot.com/2010/08/nassims-selective-ret... .

If you don't mind me asking, I'd be curious to hear what decisions you've made on cash vs accrual accounting, pass-through vs corporate taxation, and the use of guaranteed payments vs distributions.

I know a lot depends on the specifics of each situation, but haven't heard a lot of details on this from people running single-member bootstrapped SAAS LLCs before :) .

And if you're selling something which can be resold for most of what you've been paid, then either you or your payment processor is giving away free puts on Bitcoin. And if you offer refunds at par value in USD, you're also giving away free calls on Bitcoin.

I'd love to see the converse, where the goal is to write code that looks really malicious but is actually a useful and helpful program -- like an apparent rootkit that actually implements a full webserver.

It's definitely safe to adjust pricing after launch, even for subscription products. The key is to grandfather users in at their current rate if you are raising prices, and to charge them less if you are lowering prices.

You can probably even do a direct A/B test on price if you want, although you have to be careful and considerate if you do so.

Markets are only hard-to-predict in terms of future prices and returns. If all you want to do is to compare the relative value of two assets, the current price is a great place to start out. For example, the market is pretty sure that Exxon Mobil is worth vastly more than Groupon. Likewise, a prediction market does just fine at telling you stuff like "Romney is much more likely than Ron Paul to be the Republican candidate."

There's also two in Cambridge (one by Fresh Pond and one on Mem Drive near the BU bridge) and one in Coolidge Corner in Brookline.

I believe Patrick is referring to this Bitcoin exchange being hacked: http://news.ycombinator.com/item?id=5043122 .

However, they have a legal amount of minimum parking spaces per residence/store, which mandates auto-centric development. If you search for "Houston" in http://www.streetsblog.org/2010/09/01/shoup-to-otoole-the-ma... , there's some good discussion of this issue.

Calling her a monkey is unbearably and unquestionably racist, and I think you might want to take a minute or two and think about why you said that.

From http://www.aaronsw.com/weblog/semmelweis :

"Looking at ourselves objectively isn’t easy. But it’s essential if we ever want to get better. And if we don’t do it, we leave ourselves open to con artists and ethical compromisers who prey on our desire to believe we’re perfect."

The perverse incentive wouldn't be to seize more assets -- it would be to generate more guilty convictions for defendants whose assets have been seized.

In case anyone is wondering about Noah Kagan's past sexist behavior at conferences, http://sherprog.com/2011/07/10/noah-kagan-and-the-faceless-b... and http://news.ycombinator.com/item?id=2749858 is another example.

I'm really not sure why he's being invited to speak if he still hasn't learned to be so misogynistic.

It may be infeasible at massive scale, but many systems do it right now with approximation algorithms, heuristics, and integer programming. Fedex, Walmart, Amazon, and most other industrial or logistics corporations have teams of people with a background in discrete optimization or operations research building and improving similar systems.

I'm not aware of any public APIs for this at present.

Best wishes to him and the rest of your family.

What about vibrating motors placed in different locations? I've always been intrigued by this kind of vibrating compass-belt: http://www.gradman.com/hapticcompass .

I've never been more excited to give a website my email. Half the reason I check HN these days is to see whether Patrick has posted anything interesting and actionable. If giving him my email and some money will incentivize him to take some time from his products and consulting to share what he's learned, that's a win in my book.

I would have been glad to sign up even if the landing page had just been a text field to enter my email and a disclaimer saying "Every second email I sent you will be spam intended for Latvian philatelists.", so I think offering a 45 minute video and a transcript is a gracious touch beyond what Patrick needed for his email list MVP.

On the other hand, if you can eventually measure long-term value and you can keep track of users over a long enough cycle, then you can eventually run an A/B test on long term value. This depends on accurately tracking users, having a quantifiable way to calculate long-term value, and eliminating other sources of error, but it should be absolutely doable.

You can also measure how well your proxy stat does correspond to long-term value and adjust your expectations accordingly.

Previously discussed at http://news.ycombinator.com/item?id=1277004 , with some good commentary on simulations of various stopping rules.