Post

Not thinking, and not even telling you ... it's just billion parameter autocomplete with prompts, checks and guardrails, and software to display or text-to-speech.

Mine just said it was “icing the cake” … very excited for when a cake spits out of the usb port! 🍰

Trump has created an "Enemies List" including David Pakman & Brian Tyler Cohen: www.youtube.com/watch?v=1WHn... www.youtube.com/watch?v=T8mY... BlueSky will move into his sights very, very soon.

It’s not a wait spinner—it’s cognition.

I will have Gemini find a link for me and if I don't check it, sometimes it's Rick Astley and Gemini is trying to troll me.

I strongly doubt Claude or any LLM has any inner world or internal perspective required for consciousness, and as many others have pointed out, Richard Dawkins was simply dazzled by an effective mimic. Claude has no sense of itself; the algorithms just give the illusion that it does.

Famously, travel sites like Expedia had to add false loading screens saying they were checking various sites because people didn’t trust instantaneous results and saw more value in what they perceived as results that took time and effort.

YOOOOO Apple Human Interface Guidelines!! 🙌🏼

www.youtube.com/watch?v=klGW...

Bring back the spinning pizza wheel. Teach users patience rather than quasi-satisfied expectation.

I rely on the alien analogy. Super intelligent aliens come to earth. We don’t know how they think or what they know but they seem to demonstrate abilities far beyond ours. If they’re not human-intelligent aren’t they still intelligent?

This is why Section 230 no longer applies.

You mean that my game isn't really reticulating splines or aligning planets while it's loading? Damn.

No maybe about it! A I will make all of it so much worse I hate A I

"Running your input through a giant nueral net"

My engineering and ethics class in university was the #1 class people skipped

It needs to go backwards. Outside of the window frame. Outside of the screen. Let the progress bar be free!

Off on a tangent, I’m thinking of the Windows file copy dialog and what Microsoft did to it in … was it windows vista? 7? Where it had this graph and all this stuff Mostly to hide the fact that copying lots of / big files on windows has some tragically slow backend that’s not fixable easily.

No, we just need mountain dew and poptarts. Much cheaper than RAM 😁

That'd be designers

You think?

Until recently, this wasn't "companies" doing this so much as machine learning researchers. Probably the first notable historical offense here was the coining of "perceptron" and "neural network." But companies like Anthropic are offenders now. They recently called some feature "dreaming" 🫠 FFS.

Yes, the automatic completion point is the crack in the whole debate. Most of our own thinking is exactly that, retrieval that feels like thought. So if we call ours thinking and theirs mere completion on identical behavior, is the line about the process, or about who we decided gets the word?

Just because an AI program refers to itself in the singular does not mean it’s a person. It’s another programmer lie.

BS article, they do not "believe" anything... They are fancy autocompletes based on human (written) stories ... So after feeding the Wikipedia Article of 2001: A Space Odyssey and other examples it's completely predictable what the autocomplete of that input would be

I wish that were the case when I took the GMAT for the first time: Computer: do you wish to submit? Me: yes Computer: fail Instead instantly returning the results, a spinner for a couple seconds would have been nice.

I think it has one benefit... it's gotten a bunch of developers to learn synonyms for thinking, postulating, considering, etc...ruminating is one of my favorites...do they even know what that means?

I don't think it's the programmers coming up with the marketing horse shit

Just look at X

The emergent behavior was definitely not programmed.

this behavior cannot be programmed only trained. Models aren’t programmed as much as they are grown.

The actual plumbing doesn’t tell you much about what’s going on in the guts of a model with trillions of weights once it’s been trained. There a huge amount of complexity there that one really understands.

I mean neutral nets were meant to mimic the brain.

Fundamental it boils down to what's "reasoning"? At its core, LLM has a long, long list of "if A, then B", "if B, then C", and so on. The difference from traditional programming is that instead of explicitly mapping B to each other, the "B" can be fuzzy matched.

I’ll bet you don’t have a definition of understanding or comprehension that allows for testing whether it’s happening.

No Animal brain runs on code. A computer chip has a decent number of connections nowhere near as many as a brain, but what’s happening in a computer chip is driven by software which is nothing like an organic brain. A brain takes in data directly and reacts to it without any set of instructions.

In other words, an LLM can't reason for the same reason as a car can't fly. It lacks wings. And you can't just add internal feedback connections to an LLM because it relies on back propagation which only works with feedforward networks.

But can it clutch my pearls for me? It's becoming tiresomely repetitive... I'd like to outsource that please

My smart TV was being careless last night.. it was doing this little dot going around in a circle while changing colours, instead of loading this BKFC thing I found on ze net... I had a glass mug and beer in my hands and stopped pouring, pondering the purpose of the pursuing pulses.. then decided no

To be fair, I make random noise when I'm being creative too

"maybe"?

they need to take an "ethics of programming" class

As a 40 year developer, I resemble those remarks

Also, thinking = calculating. It's not capable of "thinking."

Yeah, but if a webpage is loading slowly, then a bunch of users all hitting refresh and flooding a struggling server with requests doesn’t fix things. A little animated gif that makes people think they’re being attended to helps stop them from mashing things.

And honestly we really have to define the stuff our brains do if applying the same words to tech...

It comes off as pretentious to want to discuss what someone puts behind the term "thinking". Who am I ask? I have no background neither in philosophy nor cognitive sciences. The thing is, most of the people who claim LLMs can think don’t have such a background either.

What’s it doing if it’s buffering and not really reasoning or thinking? 🤔

I'm sure AI will fix this.

Yeah, when I call and hit an automated assistant and it wastes my time by thinking, only irritates me.

When I was in IT, I had a desk motto from IBM: "People should think. Machines should work." Very true!

either that or the human brain is an llm

That is a specific type of reasoning. There are others. Perhaps ‘abstract reasoning’ was the intended meaning of the op? Even then, current llm models appear to at least mimic abstract reasoning.

the real tell is that the big ai companies even have employees at all. if it does what they say it does, then why all the people?

THIMK

The media giants that sell your attention don't like that you use ad blockers and clear your cookies, they don't like that you left facebook either, so AI is the gateway to those media centers to continue to misinform you.

IF Humanity = Seek THEN Elon Musk Had Rich Parents

The biggest truth it's not the developer or the programmer especially if they are the one who builds , it's mostly the companies that give them so much of money they just can't refuse but to do what they say .

Definitely.

LLMs don’t learn, so you can’t taint them with prompts.

The sims was never really reticulating splines.

Folks need to read their output more, or use a vector space visualizer. Diffusion text models have it. It's interesting. It makes it quite clear this thing acts very differently to how we understand baseline human thought I blame the corpo models because they gate that away from nontechnical users

You are telling me that Sim City 2000 was never actually reticulating my splines? My splines were unreticulated? I'm so embarrassed and ashamed.

Real. Instead of debating sci-fi "AI consciousness," we urgently need real regulation on the actual developers and tech oligarchs who face zero ethical constraints. Technocapitalism without guardrails is just a fast track to algorithmic exploitation.

Just because a person says it is reasoning or thinking doesn't mean he is either.

Maybe the entire AI industry needs more guardrails. They are pushing way too fast to make everything AI and that usually leads to poor oversight and mistakes. And we already know AI can make mistakes.

I shall configure mine to use the synonyms: "Generating slop..." "Resolving premature conclusions..."

I don't think many programmers were involved in this case. Claude Code in particular feels like it was vibe-coded.

some times we do it because users are impatient sods who need reassuring it's working and not to mash F5 over and over

sure thing! I couldn't agree with you more on this

“What if artificial beings maximized their reward functions without moral constraint?” we all said, as the dow shot up, and millions died. Corporations have been misaligned artificial beings for centuries. Paperclip manufacturers have been doing the Paperclip Problem forever. Iron mines are a thing.

GIGO - Garbage In Garbage Out

Or ethics.

This post hit home as I recently had to replace my TV w/a Smart TV. Basically that's all you can buy now, so I thought Samsung would be a sound choice but wrong. The TV's stupid; its lousy software freezes up regularly, it won't recognize valid passwords -- it's truly a piece of 💩 & intelligent? NO!

👍

“The question of whether a computer can think is no more interesting than the question of whether a submarine can swim.” ― Edsger W. Dijkstra

LLMs don't have to be instructed to lie. They take every input as the truth and regurgitate it as if it were true. As ever, garbage in, garbage out.

👍🏼 maybe user need more technology knowledge and understanding technology literacy is today as important like reading/writing was in medieval age

The Sims lied to me when it said it was reticulating splines?

I don't understand this when we repeatedly see AI having "aha" moments of incredible reasoning, like AlphaGo's Move 32 or OpenAI's recent disproving of Erdos' geometry conjecture. These are beneficial intellectual capabilities that we haven't had before. Maybe we need a new word to describe them.

We have used "thinking" as a progress bar for decades because that is something the end user can more easily understand. The LLM companies lying about it actually thinking in marketing is an ethics issue. Don't blame the devs.

Agreed!

Like a magical fantasy guiding crowd source space of professionals. We can call it IEEE, or something... 😅

You would be shocked how many talented engineers who should know better are still saying things like how much they get out of the"deep reasoning" sessions they have with their little LLMs

EXACTLY what I said today.

Yeah, I believe it. But I worry a bit more about software being able to lie to programmers about what it is actually doing...

Is anyone actually stupid enough to believe the service is really *thinking*? Would it make any difference if the message is "you are number 5 in the queue for the processor"?

Brendan Dawes mentions this in one of his books (AIDO I think). In a video editing web app, user testing showed disbelief of the immediacy of rendered output - users thought it must’ve made a mistake and started the render again. So, they added a fake loading bar to give the illusion of processing.

I kinda like the approach of most of Claude's "spinner verbs"—perhaps excepting a few too-humanish cases such as "musing" and "thinking"—to convey the please-wait-while-I-work-on-that concept with some amusement. 😆 bsky.app/profile/info...

It has no mouth?

It thinks a scream

I lost all respect for programmers when my Amiga software that I bought and paid for just stopped working even though the computer works fine. Same problem with my old Iphone apps, they just stop working one day. So now, I never buy any software, F you all

It's a matter of semantics, how you define what thinking is. Do airplanes really fly?

It was never ambiguous for intelligent people. Unfortunately, the world is full of retards.

Old saying in programming, "Garbage in, Garbage out". 🗑

Why ai can be manipulated Just ask elon!

Yeah but they want to get paid. Morality has no place in the workplace more now than ever.

No! The Dreamcast IS thinking! You can't take it away from me!

"Reasoning models" is so telling of how completely divorced the AI industry is from truth. So, because LLMs make up unverifiable BS output, we just made an LLM making up equally fake reasoning for why it did.

I fully agree with the 1st point... ...but if you believe that the programmers are the problem here, you may need to widen your approach. 🙂

If you code a program giving it false information as input, you don't get to be surprised if you get prejudiced and false outputs. It's really just as simple as that.