Why Machine Learning Scares Us

So it turns out that the flavor of AI that is finally reaching the mainstream is the large language model. And we’re uncomfortable with it not only for the obvious reasons that there’s a machine acting eerily like a human, but also because we can’t explain precisely how it’s doing what it’s doing. And in that sense it’s like our own brains, and by extension so much about our universe. We’ve developed staggeringly complex science and simultaneously can’t explain how we think or where the universe came from or even make general relativity and quantum mechanics play nice together. This general situation has been unsettling for a long time, so we invented gods to explain the universe. But the gods didn’t make AI.

Dall-e created image
An image created for me by DALL-E

Passwords Are Now Outdated

There’s a lot to be said about how to secure our online information while providing a good experience; hasn’t anyone written this book yet? In any case, I think a good, uncommon, guideline would be:

Do not use human memory

To remember and recall takes work, and I’d rather not expend my precious attention and cognition on computers. And, human memory is fallible. Most passwords break this guideline, but we now have biometrics and password managers to provide security without relying on memory and it’s time our security practices catch up to the technology.

Two-factor authentication, similarly, feels like a lazy, flawed approach which requires too much human work.

The cost of war

From now until November I’ll probably be blogging more on American politics, both to refine my views and because this here is my little soapbox. You’ve been warned.

The Bush campaign’s assertion that “The world is a better place without Saddam” is absolutely true. Saddam is an evil man. But this statement looks only at the benefit of the Iraqi war and not the costs. We started a war to remove Saddam and find his weapons (the latter based on circumstantial evidence). We did remove Saddam and there were collateral benefits like scaring Syria into giving up it’s arms program. Both good things.

What is this worth to us? It’s hard to say, but the line has to be drawn somewhere. When I look at how much the war cost I think we crossed way over any reasonable person’s line.

First we should look at the cost in human life. The coalition fatalities number over 1,000. The estimates of Iraqi civilian fatalities range from 7,000 to between 21,700 and 55,000. Around 5000 Iraqi soldiers died. The “lucrative” contracts resulted in at least 111 contractors missing or dead. And at least 30 journalists are dead.

We might remember how utterly stunned we were at the nearly 3,000 deaths on September 11, 2001. If we multiply that number by four it still isn’t the at least 13,000 people killed in Iraq.

The Cost of War site has a great financial analysis (summary: $122 billion and counting). The figure of $1,700 cost per U.S. household is particulularly interesting when you compare it to the few hundred Bush gave each household in tax savings.

The Institute for Policy Studies’ costs of the war is a more complete list, including indirect costs on the economy, health, and international relations.

Update: Thinking more about this, I have to remember how evil Saddam was, and quantify it likewise. It’s thought that he has killed a million Iraqis, both through war and in terror, in a country of 22 million people. Comparing the numbers — as cold as that is — it seems worth it.

John Z. points out that we spent four times the gross national income of Iraq during the war, raising the idea of how we could’ve given that money to Iraqis to fix the problem instead. Technically it’s against international law to put a price on a national leader (not that international law has stopped the Bush administration), but some economic incentives to the Iraqi army might have done wonders to avoid any war at all. After all, we’re the capitalists, we should know how to spend this money better than anyone.

Internet Appliance

On Matt’s blog I found a place to raise my freak Internet appliance flag, ‘If we look at what people actually use PCs for, there’s a bell curve that includes email, web, digital photos, and a couple other things. Why not spend $299 for that? ‘. He points me to the Emailer plus: emailing and surfing for $60. At that price point, it’s less about replacing the PC than complimenting the PC and replacing the landline phone.

The cost of software

Interesting that a price policy in something as relatively low cost as blog software has spurred such good thinking about price policies. Mark Pilgram gives us the long-term view: ‘In the long run, the utility of all non-Free software approaches zero. All non-Free software is a dead end… It’s not about money; it’s about freedom.

And Mark Bernstein gives us the short-term view: ‘The only thing I expect, if I’m paying $700 for a software package, is that it be worth $700. If I buy a $700 program, and it promptly saves me $750, I’m as happy as a clam.


Irony was cited as a trait of my generation, but it seems even stronger in the upcoming generation, increasingly reality-show self-aware. Is this a trend, growing irony over time?

Douglas Coupland, author of Gen X, later said, ‘When you’re younger, you think a little irony is all you need. You think it’ll get you to the grave, but it won’t. Loss always seeps through. You do need to deal with it.‘ Can the jolt of reality still counter irony when reality itself is mocked?

Blog software mayhem

My MT Alternatives post back on May 6 turns out to be prescient, with Movable Type upping their prices and competitors finally coming out of the wood work. Regarding MT’s prices, mostly I’m disappointed in the we-want-everything-for-free attitude of the complainers (heard before when Blogger had load and security issues), I thought that went away with the dot com crash. The MT folks worked long and hard to do something better, gave a lot of it away for nothing, and now the competitors are standing on their shoulders. If folks want to switch, fine, but spare us the whining.

The issue is summarized nicely by Liz Lawley.

Play sound!

So I’ve got this big, expensive computing machine under my desk that’s capable of generating some serious music, and yet out of the box it’s rather difficult to get it to play anything at all. Sometimes all I need is the A below middle C to tune my guitar, ya know? I found Audacity which helped me generate 30 seconds of such a note. iTunes converted it into a 400k AAC file, and now you can download 220Hz.m4a for your tuning pleasure (you’ll probably need to right click/control click to save it).


I’m trying to forumulate my rationale to myself of why I don’t want a piece of the Google IPO. It has something to do with 1) IPOs are bad for companies in general, 2) Google has explicitly warned it won’t try to make returns in the short term. Given that Google took the search crown in the short term, someone else is just as likely to take it away, and 3) I can’t perceive much difference between their showcase product and Yahoo’s.

While I’m getting those thoughts straight, see their soul-sucking new blog, and their counter-philosophy new banner ads. Those Morgan Stanley folks really know how to whip a company into shape!