P.S. A Column On Things

By PAUL E. SCHINDLER JR. I am from Portland, Oregon, Beaumont ’66, Benson High ’70, MIT ’74. Some things are impossible to know, but it is impossible to know these things.

I should have known that the readers of this column would be fascinated by Douglas Adams, who weighed in on AI last week.

I met Douglas Adams at the news conference for the Starship Titanic game in 1998. I’m surprised he didn’t appear to plug it on the Computer Chronicles. There was a large attendance at the news conference, mostly people from media that couldn’t do the game any good but who wanted to see Adams.

 This is where I discovered that his locution “This is obviously some strange usage of the word ‘safe’ that I wasn’t previously aware of”[1] was a catchphrase that he dropped into casual conversation. He used it twice that day.

I remember being impressed by his wry British wit. I own all the H2G2 books and all the British radio shows. I’ve seen the movie.

As a journalist, my personal favorite of his witticisms was, “I love deadlines. I love the whooshing noise they make when they pass by.” He was notorious for missing deadlines. 


[1] I have been polishing this quotation for decades. My version: “This is some interesting new definition of [blank] I wasn’t previously aware of.” Mine is three words shorter and pithier

Posted at 9:10 pm Permalink No Comments

A Tip o’ the PSACOT Hat to Daniel Dern, for this free NYT link:

How Accurate Are Google’s A.I. Overviews?

The company’s A.I.-generated answers look authoritative, but they draw on an array of sources, from trustworthy sites to Facebook posts.

By Tripp Mickle, Cade Metz, Dylan Freedman, Teresa Mondría Terol, and Keith Collins
April 7, 2026

The reporters spoke with firms that study A.I. hallucinations before selecting Oumi and its A.I.-verification model, HallOumi, to evaluate Google’s accuracy via a widely used benchmark test, known as SimpleQA.

Posted at 9:10 pm Permalink No Comments

It is not only idiotbots that get confused and conflate results. Google can too.

In answer to my recent column on the subject, S.M. Oliva checked in:

“A favorite example of this in my own research is Tom Snyder. No, not the former late night talk show host, but an early programmer of educational software. Tom Snyder Productions, Inc., was one of the first “edutainment” software developers. He never hosted a talk show, although he later got into television production. Yet I’ve seen more than a few sources conflate the two Tom Snyders over the years.”

Plus, I forgot to mention my friend Larry King. Not former talk show host Larry King, but rather the journalist.

Posted at 9:05 pm Permalink No Comments

I had it, I lost it. I had it, I lost it. Just another day with the single worst user interface on the Internet: LinkedIn. If you don’t respond immediately, the thread disappears, never to be found again. Yes, I looked in my activity. Yes, I checked all of my connections. Nada. So, I don’t get to react to the original posts. 

With the exception of UIs produced by amateurs, this is the worst. The god-awful LinkedIn interface has done it again. They sent me an email that indicated I had a new message. Clicking on the link in the email sent me into deep oblivion. A few minutes of attempting to find it in the bewildering and useless message interface was fruitless.

I think this just proves we had AI slop before we had AI. I think a bot could have designed a better interface. It appears that, like so many other lousy UIs, this one results from the accretion of a bunch of random decisions yielding an impenetrable morass, which no human being ever sits down to rethink. But of course, if they rethink it, the familiar but crappy interface morphs into one that is no better, and is unfamiliar to boot. 

Posted at 9:27 pm Permalink No Comments

You may or may not recall that in the early days of the Internet, when HTML was simple, if you liked a page you went to the source version and borrowed big chunks of clever HTML. Now you can do the same thing with AI prompts, if the original prompter is careless–as in the sense of doesn’t care.

I was looking for a cover for my song How Many Times, which asks the question “How many times have we been together,” with the stinger “Would I do it again? You Bet.” Here is a discussion of the use of AI for that cover.

Posted at 9:20 pm Permalink No Comments

I can prove it is the 21st century. The power went out. We were switched automatically to our backup generator.

PG&E texted and emailed the equivalent of “We think your power is out, but we’re not sure. What do you think?” Instead of asking me to send a pointless email or text, it took me to a new form. “What’s your address?”

Then “it may take us a minute or two, but we’re going to have a look at your meter.” Less than a minute later, “The power is out at your meter. We’ll get on it as soon as it is convenient for us.” (OK, that last part is a little snarky).

I was pleased to be able to provide useful information. The only thing that would be better, but why didn’t they look at my meter without asking me?

Posted at 8:47 pm Permalink No Comments
Author's image

Paul E. Schindler Jr.

SEARCH PSACOT 2015-Present

Sign Up For Weekly Email Notification

* indicates required

Email Me

Email Me