Post

In 1980 an IBM 3033 mainframe cost $3,000,000 with 8 megabytes of RAM. Running assembly language it was the fastest machine in the Jan 1983 Byte magazine Sieve of Eratosthanese benchmark. A smartphone would blow it away.

Now that I’ve retired from the field, I’m getting around to reading books on the subject(s) I’ve collected and never read past Chap 1. A Small Matter of Programming by Bonnie Nardi (1993) And The Human Use of Human Beings: Cybernetics and Society by Norbert Weiner (1954)

It’s online here! archive.org/details/comp...

📌

📌

I haven’t consciously thought of that book over 20 years, but much like The Mythical Man Month it underlays a lot of my thoughts. I go back and reread it at some point.

There is actually a lot of material from the dawn of the age of personal/home computing that we need to revisit A lot of SV tech bros could use a history lesson This stuff makes me feel old, but it must be shared

Hell yes, this rocks. I’m getting it.

📌

Yes!

this really drives home how the goal for these guys is NOT to create a computer that is "sentient" or "thinking" or whatever adjective one wants to use for "person", but rather a Legal Entity incapable of legal liability.

Alfred Bester’s first novel The Demolished Man captures the first highlighted section beautifully in a surprisingly approachable way for Golden Age SF

humans can in fact possess wisdom unlike machines, however

The point is we don't make every human a psychiatrist or judge. Jobs that require wisdom, in theory, are supposed to only go to people who have demonstrated wisdom; whereas an LLM can never have wisdom, no matter how many individually wise training data items you feed it.

There is something very wrong with saying 'we have created a cheap efficient labor-saving device that can try to do some jobs' and leaping to 'we will entrust human psyches to it'. If tons of labor is being saved, freeing up lots of people, then maybe humans can be available to care for humans.

George Lucas’ first, “THX 1138” (1971) had Robert Duvall’s character doing a confession-style interaction with an AI face, in what looks like a bus bench and display of today.

bsky.app/profile/danh...

Joseph Weizenbaum’s “Computer Power and Human Reason” is available on Anna’s Archive…though I think we desperately need a reprint run. annas-archive.org/md5/532a20e6...

I've played with Eliza before.

The AI as Tarot reader. Same result.

Best I can do you right now is "visceral repulsion" bsky.app/profile/pars...

Funny, it still kind of is. A big server room has a LOT of fans going on.

> The question then arises, and it answers itself, do we wish to encourage people to lead their lives on the basis of patent fraud, charlatanism, and unreality? I fear in many ways we have our answer to this question...

I will not accept responsibility for this particular nonsense. This is all on those fucking billionaires.

And "affront", that needs to be used much more often too

📌

Flip coins? Computers are NOT good a generating random outcomes. Are they?

It really does seem like it should be obvious. A glorified chatbot does *not* have wisdom. It just has mathematically determined "vibes".

I read Computer Power and Human Reason many years ago and was impressed by his arguments. The above quote is the last two sentences of the final chapter.

What's funny is that LLMs are also no good at answering question of fact or logic. I've been trying to teach ChatGPT the asbestos regulations, and it reaches to the most absurdly wrong missinterpretations. The only thing they are good at is mimicking human language.

📌

📍

From the Wikipedia entry: “The book caused disagreement with, and separation from other members of the artificial intelligence research community, a status the author later said he'd come to take pride in.”

Thanks for the rec!

I recently read something kinda similar in theme- Frank Herbert, author of Dune, wrote a book in 1980 that is meant to demystify what computers are but has a lot of thoughts about AI and computer autonomy. You can feel the distrust he has for the machines. www.goodreads.com/book/show/13...

📌

Oh excellent! I would love to help promote that any way I can.

Sure, can you DM me your email?

Thanks, I've read some of his essays before, time for the whole book. It seems he was way ahead of the curve, prescient in that down to earth way of the truly wise.

OK, my tsundoku reading list is *already* causing a wobble in the earth's spin. Can we either just drop the AI thing already or just give everyone a month off to read?

I ask Joe to autograph my copy using my own Parker pen. Another time, I sat next to him during dinner at a conference in Portland. We both expressed dismay when Seymour Papert stated that teachers should STFU and let students explore to learn. He came to the dedication of Sci bldg at L&C college.

Taps the sign: bsky.app/profile/robi...

There! That’s where I’d heard of him. My prof in 1979 Intro To Programming ran us through the concept of the Turing Test, and then said a paper was still being discussed as to whether someone had made a machine to pass it.

Well, they would…

That’s some pretty strong street cred.

I think you are utterly, profoundly wrong.

The irony of a prescient critique of computational thinking being preserved through digital archiving is not lost. Weizenbaum would likely appreciate the meta-commentary.

Rad!

Between you and @niedermeyer.online posting about this, I grabbed the copy from my university library. Last checkout: 1997. I have until December to absorb it.

I have a copy in my backpack! It's on my list to read next!

I checked this out from the McGill library months ago and it's just been sitting on my desk. I should change that. But it also means that no one is has requested it in that same time. Yikes.

📌

And this is another excellent book. www.simonandschuster.com/books/Mind-O...

thanks for posting about this book!

Even in 1969, Hal in 2001 Space Odyssey was given an ominous role…

bsky.app/profile/chec...