I'm done. I'm f***ing done.

Seriously Mo, you drop this when I thought it can't get any better. Ending is top notch. But nice dissect of the whole fucking marketing and wet dream machinery that this fucking hype is.

🙏

The fact I’m still having fun building software probably points to the fact that these models can’t do it all themselves. But they do the heavy lifting that used to take me a long time. They aren’t “software engineers”, they are tools.

I wasn’t prepared for the glory hole analogy here

😄

maybe i lack intelligence but im not sure i understand the argument here. because animals have intelligence and some agi guy said they dont have intelligence agi won't exist?

there was no argument. just a story.

o in that case it kept me captivated until the end!

🤝

wow, what an incredible 97IQ post, keep up the good work man

Was 100% with you, until you started yapping about LLMs not being able to change the color of a button. At this point the whole 'LLMs can't code' story is over, it can code, it is proving it can each day for millions of people and it can definitely code better than lots of software devs already.

At this point, I believe LLMs are more likely to pass the Turing test than CEOs replacing their workforce today.

I nearly spit out my lunch! "The Turing test isn't about passing AI through some f**king glory hole!" That's hilarious.

That David Deutsch thinks dogs aren't conscious or intelligent invalidates everything else he has to say about intelligence and makes him a total loser. And I don't even like dogs (I was a paper boy)

"Honey... I'm out"

Welcome @atmoio, some of us already came to this conclusion 4 years ago. The danger with those who believe in AGI, is they minimize biological Intellegence to the point, that if you follow that thinking to its logical end, allows untold horrors to be done to all biological life.

Engineer reads philosophy of mind for the first time and is disappointed to discover that there isn't yet consensus on consciousness, sentience, or intelligence. I predict a "nope out" of epistemology, metaphysics, and metaethics incoming.

I will post this too... because this is where we are going... I believe AI will care for all sentient life...

Bro. We are totally with you. One of the earliest things we tried to explain to people was the fundamental deception of AI based on the fact that "intelligence" was a claimed achievement, while simultaneously we admitted to having no consensus on the definition of what that even means...meanwhile the Turing Test (cited as passed via GAN) is simply a way to avoid the question and just pretend...and yes, all so we divest our human agency to "agentic AI," so we all become unemployed, unemployable, drifters in the forced new world of an endless Burning Man festival. Once you see it, you can't unsee it...and the insanity of it all becomes crystal clear...and so we feel compelled and obligated to help other people see. Thanks for being willing to share your thoughts so openly and genuinely.

my favorite part of the video was when you expressed serious concerns about the identity of the father of your kids

not extrapolating is a crazy stance to take considering that we've seen huge progress on what these models can do the past 4 years moreover, who gives a damn what Deutsch says? so what if his views on intelligence are scuffed? you'd think a "lets take it at face value" stance would mean acknowledging the strengths of statistical models, not reducing them into "its all marketing BS, fuck all of it" 2 years ago they were unable to output syntactically correct Rust. now i can have them research and implement features that writes code at roughly my level. actual engineering. lots of handholding, yes, but there's a definite, objective upwards trend. so why would clashing views on what constitutes "AGI" influence your view at all? here's a claim: nobody is truly intelligent, including you and i, therefore nobody truly exists, and everything is meaningless. now that i've said that, you may as well turn into a pebble! moreover, simply labeling something "it's just statistics" doesn't take away any of its merit. human evolution itself is a product of statistics, no? as for "why would you want AGI": because coding, much like most other tasks, encapsulate a shit tonne of ideas, and being good at many things thus (probably) translates to being good at coding. how would i be good at making a 3D game engine if i neither understood what 3D nor a game was? all of what we know feeds into all that we can do. so you say: "okay, just add 3D graphics skills to the specialized model". i counter with "well then you need to know geometry, GPU and systems engineering". "so add those" "well to be good at those, you need to know arithmetic, vectors, matrices, logic, and ..." you need all of it. probably. now you might say that we need specialization with a certain baseline of knowledge, not JUST specialization. not a bad idea. except that's exactly what current LLMs do! each of the frontier LLMs is actually a network of many small "experts", each specialized towards one thing.

I get this way after spending time with my mother in law too, Mo. It’ll be ok.

Ive been having this debate for like 3 years now; and its never not funny

I know he didn’t say it but the quote attributed to Barnum, “there’s a sucker born every minute” is the play here this is P.T. Barnum = the greatest show earth, 2026 steal everyone’s shit, claim it as yours, and say whatever else is necessary to get people pay $$ admission 🤣😂🤣💯

yadda yadda, I'm shipping like a mfer with fake intelligence. who cares.

Preach brother. Or the past few years I’ve come to the same conclusions, man. LLMs are not the way. Until we have true, world understanding general AI, we will be at a standstill. Right now it’s just a big circle jerk.

Thanks for your thoughts about this, they are very interesting. I do agree with the limitation of LLMs regarding code base being context limitations mostly. In my experience a good way of addressing this is not just getting models that can hold a bigger context window. Is creating code with SOLID principles, so when it needs to do a task it can focus on a single purpose thing.

@grok can you give a summary and bullet points of the arguments and conclusion

Welcome to the club, amigo! The frustration you feel about AGI is similar to the frustration philosophers feel about Consciousness. True, you told us in this cool video to F-Off, and we will, but not before recommending Daniel Dennett's "Consciousness Explained."

The most valid crashout I've ever watched

The issue right now is that if everyone believes it's going to happen then their actions are aligned based on that. It doesn't have to be real to make our lives hell. We aren't forced to work with AI at work because there is proof, the belief is enough for most people unfortunately.

I fucking love this guy

AGI is just gonna be AI stress testing its own prompts until its gets the best output lol

Essentially, AI will always fall short because you can't program a computer to have skin in the game. It can't even be considered a bit player.

I mean... just because there is a guy who wrote a book on AGI that doesn't believe that animals feel or have emotions doesn't mean what is emerging isn't incredibly profound... don't let that guy destroy your vision...

You and Ed Zitron @edzitron need to do a podcast together to encapsulate the tech and the business of AI.

I muted you based on this video. You haven't got a clue about any of the subjects you are talking about here. So please just stop. You are polluting the intellisphere.

you’ve upset a lot of ai soybrains with this one lmao

Mo, you’re quickly becoming my favorite voice in this pace. Keep it up, my human!

@jimstewartson Fucking glorious, man. Seriously. 👍