<?xml version="1.0" encoding="utf-8"?>
<rss xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title>Disqus - Latest Comments for mcramer</title><link>http://disqus.com/by/mcramer/</link><description></description><atom:link href="http://disqus.com/mcramer/comments.rss" rel="self"></atom:link><language>en</language><lastBuildDate>Tue, 30 Sep 2025 17:20:46 -0000</lastBuildDate><item><title>Re: In the beginning, there was data - Hackin’ and Tinkerin’</title><link>https://mdcramer.github.io/apple-2-blog/synthesizing-data/#comment-6776164906</link><description>&lt;p&gt;The hardware is, of course, slow. I took a few speed measurements at &lt;a href="https://mdcramer.github.io/apple-2-blog/refactoring/#testing-irwin-hall-as-an-alternative-to-box-muller" rel="nofollow noopener" target="_blank" title="https://mdcramer.github.io/apple-2-blog/refactoring/#testing-irwin-hall-as-an-alternative-to-box-muller"&gt;https://mdcramer.github.io/...&lt;/a&gt; but I intend on doing more analysis with the full k-means algorithm in a future post. Anyway, the Apple ][+ has 64k of RAM although not all of that is available for data. If we talking 2-byte signed integers, you could fit 8k numbers in the expansion card.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Tue, 30 Sep 2025 17:20:46 -0000</pubDate></item><item><title>Re: Refactoring - Hackin’ and Tinkerin’</title><link>https://mdcramer.github.io/apple-2-blog/refactoring/#comment-6739212322</link><description>&lt;p&gt;Thanks, Mike. I asked chatGPT to take a look at the Applesoft BASIC assembly and, naturally, it's complicated. (For the record, I could have looked at the actual assemble code for a year before figuring it out.)&lt;/p&gt;&lt;p&gt;RND uses a linear congruential generator. The equation is (5 * seed_n + 1)  mod 2^24 divided by 2^24. COS is just a phase shift of SIN, which uses Horner's method, which is a polynomial approximation using Taylor series expansion: sin(x) ≈ x⋅(C1+x^2 ⋅(C2+x^2 ⋅(C3+x^2 ⋅C4))). SQRT uses exp (1/2 * log(x)). Box-Muller also needs LOG which also uses a Taylor polynomial: log(x) = n⋅log(2) + log(f) where x = 2^n⋅f.&lt;/p&gt;&lt;p&gt;I then asked chatGPT to compare computational cost and wow:&lt;br&gt;RND is ~150 6502 ops&lt;br&gt;SIN is ~300-500 ops&lt;br&gt;SQRT is ~800-1200 ops&lt;br&gt;LOG is ~500-700 ops&lt;/p&gt;&lt;p&gt;So computing &lt;i&gt;two &lt;/i&gt;Box-Muller standard normal random variables requires 400 + 400 + 1000 + 600 = ~2400 ops. (There's an efficiency because you take the SQRT and LOG once and use it twice.) Two Irwin-Hall standard normal random variables requires 24 RND = ~3600 ops. As such, it looks like Irwin-Hall is ~50% more ops, which is almost exactly what I got with my empirical result.&lt;/p&gt;&lt;p&gt;Note: If I wasn't computing the standard normal random variables in pairs, Irwin-Hall would be slightly faster. 400 + 1000 + 600 = 2000 vs 12 * 150 = 1800.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Thu, 17 Jul 2025 22:22:14 -0000</pubDate></item><item><title>Re: Refactoring - Hackin’ and Tinkerin’</title><link>https://mdcramer.github.io/apple-2-blog/refactoring/#comment-6737180563</link><description>&lt;p&gt;Thank you, @Kunal Punjabi , for being the very first commenter across my entire blog. 😆 Glad to know this thing is working...&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Mon, 14 Jul 2025 12:42:55 -0000</pubDate></item><item><title>Re: What I’ve Learned Unsubscribing from Hundreds of Mailing Lists</title><link>https://www.sparkminute.com/2015/11/12/what-ive-learned-unsubscribing-from-hundreds-of-mailing-lists/#comment-2356475164</link><description>&lt;p&gt;Not at all. It seems like you've got a difficult inbox issue. Whether it's newsletters or spam, a white list is just an option for working with it.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Thu, 12 Nov 2015 14:59:05 -0000</pubDate></item><item><title>Re: What I’ve Learned Unsubscribing from Hundreds of Mailing Lists</title><link>https://www.sparkminute.com/2015/11/12/what-ive-learned-unsubscribing-from-hundreds-of-mailing-lists/#comment-2356446796</link><description>&lt;p&gt;I understand the difference between newsletters and spam.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Thu, 12 Nov 2015 14:41:26 -0000</pubDate></item><item><title>Re: What I’ve Learned Unsubscribing from Hundreds of Mailing Lists</title><link>https://www.sparkminute.com/2015/11/12/what-ive-learned-unsubscribing-from-hundreds-of-mailing-lists/#comment-2356413068</link><description>&lt;p&gt;If the problem is unmanageable, try something like &lt;a href="http://www.spamarrest.com" rel="nofollow noopener" target="_blank" title="www.spamarrest.com"&gt;www.spamarrest.com&lt;/a&gt;. I have an email that gets so much spam it was either this or abandon it. I chose Spam Arrest and it's worked well for almost 10 years. It's mildly annoying to new people who get the challenge-response, but they only have to do that once and the alternative was scrapping my email.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Thu, 12 Nov 2015 14:21:56 -0000</pubDate></item><item><title>Re: Ello Says You&amp;#039;re Not a Product, But You Are</title><link>http://www.theatlantic.com/technology/archive/2014/09/ello-says-youre-not-a-product-but-you-are/380809/#comment-1608422255</link><description>&lt;p&gt;Hello Adams - Didn't realize that you were the same guy. I'm glad you liked my paraphrase. I was trying to be humorous.&lt;/p&gt;&lt;p&gt;Again, the M&amp;amp;A argument is good - Is all my personal stuff going to be sold in an acquisition? I think Ello will need to address that. (If they were to say that you're data wouldn't be sold in an acquisition, and for the sake of argument you believed that, would you then be OK?) However, that's not really the point of this article? It's saying that if you use their website, you're a product.&lt;/p&gt;&lt;p&gt;I get it with Facebook - the website is free of charge, but they collect data about you that they then sell to advertisers in order to show you ads. But, unless I missed something critical, the whole point of Ello is to be not that - the website will not be free of charge (unclear what you have to pay for, but it'll probably be freemium like Evernote) and so they won't collect data about you and they won't sell it to advertisers and they're not going to show you ads. Unless you think they're being untrue (no basis for that) then you're not a product, right?&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Sat, 27 Sep 2014 12:25:37 -0000</pubDate></item><item><title>Re: Ello Says You&amp;#039;re Not a Product, But You Are</title><link>http://www.theatlantic.com/technology/archive/2014/09/ello-says-youre-not-a-product-but-you-are/380809/#comment-1607492383</link><description>&lt;p&gt;Hilarious. "In order to be profitable in the social networking industry the best way to make money is to sell your member info to advertisers and/or allow them to send ads to your page." THAT is a huge assumption, and one that Ello purports to be challenging: &lt;a href="https://ello.co/wtf/post/why-no-ads" rel="nofollow noopener" target="_blank" title="https://ello.co/wtf/post/why-no-ads"&gt;https://ello.co/wtf/post/wh...&lt;/a&gt;.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Fri, 26 Sep 2014 16:39:14 -0000</pubDate></item><item><title>Re: Ello Says You&amp;#039;re Not a Product, But You Are</title><link>http://www.theatlantic.com/technology/archive/2014/09/ello-says-youre-not-a-product-but-you-are/380809/#comment-1607412634</link><description>&lt;p&gt;I responded to the M&amp;amp;A argument above, but that's a decent point - your content on Ello is an 'asset' of the company which "could" be sold, although it's possible that they'll enable people to delete their stuff. Regardless, it's not really the point being made in the article. Adams is saying that "If you use Ello then you're helping the founders and employees of the company by making them look good so that they can pad their resumes or whatever." And that is true for every business, for-profit and non-profit.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Fri, 26 Sep 2014 15:43:25 -0000</pubDate></item><item><title>Re: Ello Says You&amp;#039;re Not a Product, But You Are</title><link>http://www.theatlantic.com/technology/archive/2014/09/ello-says-youre-not-a-product-but-you-are/380809/#comment-1607404697</link><description>&lt;p&gt;Nope.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Fri, 26 Sep 2014 15:39:22 -0000</pubDate></item><item><title>Re: Ello Says You&amp;#039;re Not a Product, But You Are</title><link>http://www.theatlantic.com/technology/archive/2014/09/ello-says-youre-not-a-product-but-you-are/380809/#comment-1607403370</link><description>&lt;p&gt;That requires making a lot of assumptions about what happens to assets during change of control. You're probably right, but it's not the point being made in the article. Regardless, non-profits can be acquired, too, and when they do the names and addresses of their donors (i.e. their "products") go with it. So, again, no diff.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Fri, 26 Sep 2014 15:38:27 -0000</pubDate></item><item><title>Re: Ello Says You&amp;#039;re Not a Product, But You Are</title><link>http://www.theatlantic.com/technology/archive/2014/09/ello-says-youre-not-a-product-but-you-are/380809/#comment-1607321224</link><description>&lt;p&gt;This article is terrible. "You might decide that being that kind of currency—the kind that promotes investment, hiring and promotion of this companies and these people—is fine with you." - This is true for every single company that ever existed. When you transact with a company, you are inherently improving the odds of success for that company, and their employees, enabling them to expand and hire more people. Does that make you a "product"? Nope.&lt;br&gt;"But as long as it remains for-profit, it will almost certainly treat you something like a product..." - Sorry, but non-profits are not different. When you donate money to a charity, that charity will Market the amount of money they raised in order to raise more money, attract employees (who need to get paid) and engage in Marketing activities. Oh, and all those employees can post on their resumes how much money their charity raised which will ultimately help them get new jobs or investments or whatever. So does giving money to a non-profit make you less of a "product" than a for-profit? Nope.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Fri, 26 Sep 2014 14:38:53 -0000</pubDate></item><item><title>Re: Ello Says You&amp;#039;re Not a Product, But You Are</title><link>http://www.theatlantic.com/technology/archive/2014/09/ello-says-youre-not-a-product-but-you-are/380809/#comment-1607314086</link><description>&lt;p&gt;Cavlee has got it right. "You’re a line on their resume that gets them that next job, or that next seed money for that next startup: Founder, Ello, 200,000 users (hey look, that’s you!)." When you buy a coffee at the coffee shop, you're transaction is part of their Income Statement!!! The coffee shop can then go around and tell the world, "Look how many coffees we sold." Your transaction, without any personal information being transmitted or sold, can then be used by the coffee shop to pitch opening a new coffee shop. See how awesome our coffee is! Does that make the coffee buyer a product? Nope.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Fri, 26 Sep 2014 14:33:35 -0000</pubDate></item><item><title>Re: KlearGear goes into hiding as world realizes it's awful</title><link>http://vator.tv/news/2013-11-19-kleargear-goes-into-hiding-as-world-realizes-its-awful#comment-1130031096</link><description>&lt;p&gt;Not that I'm a lawyer myself, but I believe with the lawyer on &lt;a href="http://www.kutv.com/news/features/gephardt/stories/vid_474.shtml" rel="nofollow noopener" target="_blank" title="http://www.kutv.com/news/features/gephardt/stories/vid_474.shtml"&gt;http://www.kutv.com/news/fe...&lt;/a&gt;. It's an unconstitutional and unenforceable infringement on 1st Amendment rights. And it is fraud. Unfortunately, to prove that they'll have to go to court, which costs money. Maybe someone will take up their cause.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Tue, 19 Nov 2013 15:01:04 -0000</pubDate></item><item><title>Re: So Does McDonald&amp;#039;s Make a Profit In Australia, Land of the $15 Minimum Wage</title><link>http://www.theatlantic.com/business/archive/2013/08/the-magical-world-where-mcdonalds-pays-15-an-hour-its-australia/278313/#comment-989989265</link><description>&lt;p&gt;You are correct. My bad. I must not be getting enough sleep. ;-) My wires crossed because people often say taxes are passed onto the consumer, which is different since taxes are not a cost of production.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Tue, 06 Aug 2013 02:34:35 -0000</pubDate></item><item><title>Re: So Does McDonald&amp;#039;s Make a Profit In Australia, Land of the $15 Minimum Wage</title><link>http://www.theatlantic.com/business/archive/2013/08/the-magical-world-where-mcdonalds-pays-15-an-hour-its-australia/278313/#comment-989598902</link><description>&lt;p&gt;"... not every extra dollar of worker compensation seems to get passed onto the consumer..." Economics 101 - cost and price are unrelated. Price is a function of supply and demand. Under capitalism there is no such thing as "passing on cost to a consumer." If a business could raise prices it would do so regardless of cost of the underlying product. I don't understand why so many people consistently get this wrong.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Mon, 05 Aug 2013 17:13:47 -0000</pubDate></item><item><title>Re: Absolutely No Machete Juggling » The Star Wars Saga: Introducing Machete Order</title><link>https://www.rodhilton.com/2011/11/11/the-star-wars-saga-suggested-viewing-order/#comment-976066090</link><description>&lt;p&gt;Apologies if someone already posted this, but if you haven't watched &lt;a href="https://www.youtube.com/watch?v=FxKtZmQgxrI" rel="nofollow noopener" target="_blank" title="https://www.youtube.com/watch?v=FxKtZmQgxrI"&gt;https://www.youtube.com/wat...&lt;/a&gt; it's awesome. Not only does it explain why Episode I was so terrible, it's hilarious. Given that it's been viewed 4+ million times, I'm guessing a lot of people have already seen it.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Thu, 25 Jul 2013 13:00:17 -0000</pubDate></item><item><title>Re: Some sucker just paid $1.5M to fly into space with Leo DiCaprio</title><link>http://venturebeat.com/2013/05/24/hes-king-of-the-world/#comment-907209191</link><description>&lt;p&gt;What Clemens said. It's a tax deduction, so if you're going to donate that kind of money to charity, why not take a plane ride with what's-his-face?&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Fri, 24 May 2013 14:51:03 -0000</pubDate></item><item><title>Re: Bing: Our Search Results Do Not Infect Users, &amp;#8220;Malware Study Was Wrong&amp;#8221;</title><link>http://searchengineland.com/bing-our-search-results-do-not-infect-users-malware-study-was-wrong-156338#comment-868679089</link><description>&lt;p&gt;They make a very good point. Moreover, it's important to not only look at the quantity of malware results that appear on the search pages, but where they rank. There are many studies that demonstrate how higher ranked results are clicked significantly more frequently than lower ranked results, so a malware site appearing in position #1 is going to be an order of magnitude more dangerous than one that appears in position #9. I could be mistaken, but I don't believe that the AV-TEST took this into account either.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Fri, 19 Apr 2013 18:32:55 -0000</pubDate></item><item><title>Re: The sad long story of Fisker Automotive, ‘the most tragic VC-backed debacle in recent history’</title><link>http://venturebeat.com/2013/04/17/the-sad-long-story-of-fisker-automotive-the-most-tragic-vc-backed-debacle-in-recent-history/#comment-868364704</link><description>&lt;p&gt;I'm not privy to the internal investment documents (and frankly don't have the time or interest to read the Privco docs), however, as a VC I'm sure you're familiar with upfront investment costs versus marginal production costs. I would image that Fisker made the case that they were going to build many, many cars and, as such, the marginal unit cost was eventually going to drop below the unit sales price and, thus, the company would be profitable. This obviously did not happen (and I would image that in your investment career you're familiar with putting money into companies where plans don't materialize), but dividing total production costs, including up-front investments, by initial production volume in order to get a per unit cost is, as I said, misleading.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Fri, 19 Apr 2013 12:21:01 -0000</pubDate></item><item><title>Re: The sad long story of Fisker Automotive, ‘the most tragic VC-backed debacle in recent history’</title><link>http://venturebeat.com/2013/04/17/the-sad-long-story-of-fisker-automotive-the-most-tragic-vc-backed-debacle-in-recent-history/#comment-867249419</link><description>&lt;p&gt;This is misleading: “Fisker spent a stunning $900,000 for each vehicle it produced,” PrivCo chief executive Sam Hamadeh told me. “Then they sold them to dealers for an invoice price of just $70,000.”&lt;/p&gt;&lt;p&gt;There is a difference between fixed up-front investment and variable costs. Obviously things went horribly wrong, but most every business requires up-front investments that, if amortized across initial production, will make the business seem unprofitable on a per unit basis.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Thu, 18 Apr 2013 14:01:40 -0000</pubDate></item><item><title>Re: WordPress admin accounts target of botnet attacks</title><link>http://venturebeat.com/2013/04/12/wordpress-botnet-attack/#comment-865300233</link><description>&lt;p&gt;This makes me furious. Our blog has been repeatedly hit over the past couple weeks, but we're using a strong password (&lt;a href="http://strongpasswordgenerator.com/)" rel="nofollow noopener" target="_blank" title="http://strongpasswordgenerator.com/)"&gt;http://strongpasswordgenera...&lt;/a&gt; and we've installed &lt;a href="http://devel.kostdoktorn.se/limit-login-attempts" rel="nofollow noopener" target="_blank" title="http://devel.kostdoktorn.se/limit-login-attempts"&gt;http://devel.kostdoktorn.se...&lt;/a&gt; which I have set to lock out the IP for 48 hours after two failed attempts. Now we just need to make sure that we don't fat-finger the password.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Tue, 16 Apr 2013 13:06:45 -0000</pubDate></item><item><title>Re: Discovering Discovery</title><link>http://blog.surfcanyon.com/2007/10/11/discovering-discovery/#comment-858752555</link><description>&lt;p&gt;Rea - Don't worry about mistakes. We've developed the algorithm to be sensitive to mistakes and adapt to changes during the search session. If the suggestions you get are not good, just skip over them and the system will adjust accordingly. Please let us know if you have any other questions. Thanks!&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Wed, 10 Apr 2013 11:52:45 -0000</pubDate></item><item><title>Re: How to reduce stress as an entrepreneur | VentureBeat</title><link>http://venturebeat.com/2013/03/31/how-to-reduce-stress-as-an-entrepreneur/#comment-848689246</link><description>&lt;p&gt;Great post and excellent advice. I'll add one - if you can, take the bus. Commuting can be very stressful, especially when there's a lot of traffic. Even if the bus ride take a little longer, you can sit back, relax, read the paper and let someone else navigate through the city mess. So far it's worked out great for me.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Mon, 01 Apr 2013 13:48:00 -0000</pubDate></item><item><title>Re: Al Gore used his Current.tv payout to buy Apple shares worth $29.5M</title><link>http://venturebeat.com/2013/01/18/al-gore-used-his-current-tv-payout-to-buy-apple-shares-worth-29-5m/#comment-821453579</link><description>&lt;p&gt;Thank you for this. I read the post several times and it's not clear. Rather than saying he "purchased stock" it should say that he "exercised options." Still, he only spent $441k to exercise (for which he probably didn't need he &lt;a href="http://Current.tv" rel="nofollow noopener" target="_blank" title="Current.tv"&gt;Current.tv&lt;/a&gt; payout) and not $29.5m. It's a bit misleading.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Mark Cramer</dc:creator><pubDate>Wed, 06 Mar 2013 18:38:19 -0000</pubDate></item></channel></rss>