AI fails

[icowden]

My fist PC was based on a 386SX processor, had 4MB of ram and a 40MB hard disk.

Ah, a youngster. I inherited a 286 from my father. It ran Dr Halo and Windows1.0 as well as a cool game called Beast where your symbol had to move away from H monsters or crush them.

 

[BoldonLad]

My fist PC was based on a 386SX processor, had 4MB of ram and a 40MB hard disk.

The first computers I worked with (years before PCs) measured memory in KB and disk space in single digit MB. In my third job (early 1970s) I was working for a Supermarket chain. We managed 3 warehouses and 500 stores order processing with an ICL1901 with 16kb memory and 2 * 4MB removable disk drives. The mind boggles. 😂

All this talk of the “olden days” is bringing the memories / nightmares back. The second system I worked on (IBM360) it was necessary to pre-allocate and manage the disk space manually via a construct called the VTOC (Volume Table Of Contents).

I am going to stop thinking about this now! 😂

 

[briantrumpet]

Quite apart from the merits of the case itself... resorting to AI to produce a judgement seems a bold move (if that's what the judge did), given the frequent hallucinations from the legal world.

https://archive.ph/ruab6

 

[briantrumpet]

Another forthright piece on the perniciousness of AI in academia, and the way that it cements hallucinated references in citations:

https://www.rollingstone.com/cultur...t-journal-research-fake-citations-1235485484/

Moser tells Rolling Stone that to even claim LLMs “hallucinate” fictional publications misunderstands the threat they pose to our comprehension of the world, because the term “implies that it’s different from the normal, correct perception of reality.” But the chatbots are “always hallucinating,” he says. “It’s not a malfunction. A predictive model predicts some text, and maybe it’s accurate, maybe it isn’t, but the process is the same either way. To put it another way: LLMs are structurally indifferent to truth.”

“LLMs are pernicious because they’re essentially polluting the information ecosystem upstream,” Moser adds. “Nonexistent citations show up in research that’s sloppy or dishonest, and from there get into other papers and articles that cite them, and papers that cite those, and then it’s in the water,” he says, likening this content to like harmful, long-lasting chemicals: “hard to trace and hard to filter out, even when you’re trying to avoid it.” Moser calls the problem “the entirely foreseeable outcome of deliberate choices,” with those who raised objections “ignored or overruled.”

But AI can’t take all the blame. “Bad research isn’t new,” Moser points out. “LLMs have amplified the problem dramatically, but there was already tremendous pressure to publish and produce, and there were many bad papers using questionable or fake data, because higher education has been organized around the production of knowledge-shaped objects, measured in citations, conferences, and grants.”

 

[briantrumpet]

I've just given Google Gemini a test to see if it can get me a direct email contact for Meta Verified support, as I've reached an impasse with the support chat in the IG app.

*If* the email address it suggests works and I get a reply, I'll be only too happy to register an AI win, as attempts via Copilot just sent me round in ever-decreasing circles. At least the email didn't bounce back.

 

[briantrumpet]

As reported in the WSJ.

 

[dutchguylivingintheuk]

Ah, a youngster. I inherited a 286 from my father. It ran Dr Halo and Windows1.0 as well as a cool game called Beast where your symbol had to move away from H monsters or crush them.

I started with 8086, even the hard drive was optional, and 100mb was big. and then the 286 and 386, funny part then at time was that you would have 486 computers that would be faster then a Pentium.
Installing windows having to swap floppy's the nostaligia

 

[icowden]

I started with 8086, even the hard drive was optional, and 100mb was big. and then the 286 and 386, funny part then at time was that you would have 486 computers that would be faster then a Pentium.
Installing windows having to swap floppy's the nostaligia

Swapping floppies? Pah! After my time. I started with a BBC B microcomputer. 32K of memory and a tape deck. My dad did eventually buy the 5.25 inch floppy disk drive and install an eprom with "View" on it (a word processor).

You haven't lived until you've spent an hour trying to load a game that takes 5 minutes on the tape deck only to have it fail at different points because THE VOLUME ISN'T EXACTLY RIGHT. At least with the floppy drive you could save and load games quickly. Well quicker.

And - you had the wonder that is Elite for the BBC B. A space exploration game with eight, EIGHT I tell you!, Galaxies each of which has 256 planets. TWO HUNDRED AND FIFTY SIX. All with different names and features. In 32k of memory. Braben was a genius. 3D wire frame graphics as well.

 

[briantrumpet]

Frogger on the uni BBC Micro diverted me away from the dissertation.

 

[Pross]

My posh friend had a BBC Micro (another posh friend had a Commodore 64). I had a ZX81 which I souped up with 64k RAM and a proper keyboard. I spent hours typing programmes into it only for them to fail (syntax error line 10 etc.) but very occasionally the cassette games would actually load and I’d get some dots bouncing around the screen.

 

[briantrumpet]

My posh friend had a BBC Micro (another posh friend had a Commodore 64). I had a ZX81 which I souped up with 64k RAM and a proper keyboard. I spent hours typing programmes into it only for them to fail (syntax error line 10 etc.) but very occasionally the cassette games would actually load and I’d get some dots bouncing around the screen.

There was some posh Colombian in my halls 1983-84, and he had a ZX81 that we played the tennis/pingpong thing on. The music department had one BBC Micro (with daisywheel printer) for the whole department.

 

[Rusty Nails]

Just remembered I have an Acorn electron plus games somewhere in the box room. Might dig it out over Christmas...if I can find a cassette recorder that works.

 

[briantrumpet]

Just remembered I have an Acorn electron plus games somewhere in the box room. Might dig it out over Christmas...if I can find a cassette recorder that works.

That'll be a compact cassette recorder. The marketing now seems hilarious.

 

[briantrumpet]

If you Bluesky, this is a revealing thread about the pervasiveness of hallucinated (i.e fake) citations in academic papers.

https://bsky.app/profile/benpatrickwill.bsky.social/post/3maecwqfprk2n

An academic writer found a non-existent paper credited to him, and so did some searching...

 

[briantrumpet]

OK, so this is fairly niche, but I asked Gemini if 'calling it quits' meant the same as 'quitting' (in response to this Graun article).

Initially it said yes, but when I pointed out that that missed the sense of an equitable conclusion for both parties, it agreed that its definition was incomplete and that I was right. Ten minutes later, I asked it the same question, and it went through the same process, so no sign that it had absorbed its first admission of error. Not a great advertisement for its 'intelligence' element.

This is the later part of the second time I asked it the question.

You are absolutely right to point that out. My first explanation leaned a bit too heavily on the modern usage of "stopping," while missing the reciprocal or equitable heart of the phrase.

The "quits" in "calling it quits" specifically refers to being quit of one another—meaning both parties are even, debts are settled, and neither owes the other anything.

The Sense of "Evenness"​

The phrase is deeply rooted in the idea of a draw or a mutual settlement. When you "call it quits," you are essentially agreeing that:

• Status is Equal: "I don't owe you, and you don't owe me."
• Balance is Restored: If there was a conflict or a competition, the score is settled at a point where both can walk away without further obligation.
• Mutual Release: It’s a "release" from a contract, a debt, or a grudge.

Historical Context​

In Middle English, the word quite meant "free" or "clear" (related to the modern word "acquitted"). If two people were "quits," it meant their accounts were balanced.

Example: If you and a friend have been trading favors or perhaps small debts, and you realize the math is getting complicated, you might say, "Let's just call it quits." You aren't just stopping; you are agreeing that the "ledger" is now at zero.

---

How this differs from "Giving Up"​

This "equitable" nuance is what distinguishes it from simply quitting:

1. Giving Up: One person stops because they can't continue (unilateral).
2. Calling it Quits: Both parties (or a person and their task) reach a point of mutual conclusion where the current state is accepted as the final state.