[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


File: Untitled.jpg (252 KB, 937x1369)
252 KB JPG
lol
>>
vibecoding is in the past. now everyone just asks ai for a binary
>>
literally me
>>
File: cramer.png (105 KB, 1186x310)
105 KB PNG
>>109331593
this pic was the one served to me right above the notch one.
>>
File: 1766842858211578.png (90 KB, 896x420)
90 KB PNG
>>109331593
Every single time. AI filtered who actually understands and cares about the tech itself vs only their ego.
>>
>>109331593
>he’s becoming a vibeGOD
based notchGOD, he understand that being a codetrans is obsolete.
This is another vibeGOD w
>>
>>108963233
>>
>>109331593
>I have been having a lot of trouble hiring competent programmers
Does this guy have 0 self awareness?

Also there are other ways of using AI code than full rejection and full vibe coding.
>>
File: stare.webm (451 KB, 364x640)
451 KB
451 KB WEBM
>>109331593
Everybody anti AI until they try to hire a programmer.
>>
>>109331631
>>109331696
lmao, that faggot bothered to screencap his own post after it was BTFO'd in the same thread
>>
>>109331631
>>109331696
>da compooter is regurgitating information it got fed wow this is so DEEP
>>
>>109331806
>>109331631
you know that you are reading low-iq midwit babble when the cretin equates the current "ai" tech to the human brain
the human brain can learn in real time, you fucking retard
>>
File: smug-cross-armef.jpg (6 KB, 200x207)
6 KB JPG
>>109331822
>The human brain can learn in real time
Then why are AI luddites still using talking points they got from using the chatgpt3 free tier from 4 years ago? Seems like brain plasticity isn't all it's hyped up to be in 95% of the population.
>>
>>109331610
>le fun and quirky retard reveal
>>
AI companies use their billions in investment on these shadow marketing campaigns
They pay niche microcelebrities to casually mention use of AI in order to create the illusion that there is widespread public adoption
illusion of widespread use = investors more easily fooled = more $$$
notch was 100% paid to tweet this and the wording makes it obvious.
>>
>>109331593
uh oh
>>
>>109331593
crow? can you even eat crow? i always assumed it was bad for you
>>
just IPO already
im tired of this extended guerilla marketing
>>
>>109331822
>luddite doesn't know what AI context is
>luddite thinks if a human brain was perfectly digitized and then not allowed to form new memories beyond the current conversation that it would no longer be intelligent or consciouss during the conversation
Smartest luddite.
>>
>>109331819
>>109331822
>my brain is special because it is, no, I can't give proof
You can't make it up, proving the post right without even realizing, literally braindead.
>>
File: 1759494620972067.png (318 KB, 505x396)
318 KB PNG
>>109331593
>haha I'll just use it one time. f-for the boring part ofcourse
>>
>>109331615
Commie luddite. Chinese AIs are the future!
>>
>>109331925
You realize Notch isn't just a "niche microcelebrity". He's a billionaire, and one that doesn't seem to really give a fuck about trying to make more money.
>>
>>109332322
who are you quoting?
>>
>>109331593
reminder Notch advertised Minecraft beta on /g/ when Minecraft was browser only.
>>
>>109331941
Why do these images look so murky, grimy and grubby?
>>
>>109331593
what a fucking retard lmao
>>
>>109332355
yes because billionaires are known for not wanting to make more money
>>
>>109331593
>>109331700
everyone knows that working for notch means that if he makes the next big thing you'll get exactly $0 next time he sells it
im not surprised
>>
>>109331593
>eceleb has an opinion
>changes it in 2 weeks
many such cases
>>
>>109332374
he posted a few times on /v/, not /g/.
the original tigsource thread is still up: https://forums.tigsource.com/index.php?topic=6273.0
minecraft is now the most selling game of all time. and that by avoiding stores like steam. crazy.
>>
>>109332495
Notch spend his whole time working alone when he made Minecraft no? makes sense that he's finding working on a team to be a pain, he's not used to that at all
>>
>>109332526
Yeah, he worked 15hours+ everyday and almost went crazy.
>>
>>109332526
ESL LEVEL 9000
>>
File: 1783162990640002.png (71 KB, 200x200)
71 KB PNG
>>109332623
>>
File: jeb.jpg (208 KB, 630x354)
208 KB JPG
>>109332526
>Notch
>Working alone
This is Jeb erasure.
>>
>>109331870
>>109332309

- no episodic memory or memory in general
- no knowledge of what they do not know
- no awareness of how certain or uncertain they are
- no inherent agentic behavior

do goyim really ?
>>
>>109332987

and these are just the basics, at the level of the algorithm, not even counting consciousness, perception, self-awareness, the ability to reason

it's an imitation algorithm, it can imitate humans like it can imitate computers, like it can imitate images, etc. of course it seems to act like a human, that is it's job
>>
>>109331593
>What options should I look into?
Jumping off a bridge would be a good one for him.
>>
File: 1760232699374481.png (27 KB, 547x127)
27 KB PNG
lol
>>
>>109331700
>>I have been having a lot of trouble hiring competent programmers
>Does this guy have 0 self awareness?
He never gave any employees at Mojang stock.
Only bonuses.
>>
>>109333077
he won the lottery, why should he share it?
>>
File: snailcat-gameboy-color.png (2.64 MB, 1448x1086)
2.64 MB PNG
>>109332453
They are trained on such images you silly boy :3
>>
>>109331615
So it's confirmed. Chinese models are the future.
>>
>>109331615
>Cramer
Oh holy shit Anthropic and OAI are soooo fucked lmao
>>
>>109332987
>- no episodic memory or memory in general
>>109332309
>- no knowledge of what they do not know
Holy brainlet.
>- no awareness of how certain or uncertain they are
Luddite doesn't know all models do is give token prediction %.
>- no inherent agentic behavior
Meaningless buzzword.

Another luddite showing subhuman logic.
>>
>>109332994
When "imitation" generalizes into essentially the same function as the thing it's imitating, it for all intents and purposes is no different in function than that thing. That's how learning from someone or something works, you try to reproduce what you see working.
>>
jej
>>
>>109331944
"eat crow" is an idiom meaning "to admit you are wrong"
>>
>>109331666
another ai sloppa too braindead to understand what he meant.
this is what ai makes to you, you lose IQ and brain cells but yeah this is what the government wants, dumb sheep. I am happy for you that you have found your destiny
>>
File: 1763500997448116.png (256 KB, 1250x1206)
256 KB PNG
>>109333021
LOOOOOOL
>>
>>109333540
you don't get it, he is providing value by being braindead meat proxy because... because he just does, okay!?
>>
File: pepecoding.gif (328 KB, 398x398)
328 KB GIF
>>109331593
High performance database kernel engineer here. How does AI help me?
>>
why do these kike corpos target shitholes like /g/?
nobody likes these subhuman brownoid shills
>>
>>109333561
they target everywhere
>>
>>109333312
>109332309
it’s not a counterargument to not being conscious, it’s a counterargument to not working like a human brain, also, it doesn’t have memories in the weights, so it has zero memories, not an inability to form new ones.
>Holy brainlet.
that’s what hallucinations are, retard: no ability to return a blank answer to a question that requires information it doesn’t possess (eg. “who is the 13th president of Brazil?” You know that you don’t know because, when you try to search for the answer in your brain, you find nothing. LLMs can’t do this because there will always be a most likely next token, they don’t have a memory system where they can search for information and find nothing) even if you trained an AI to say idk after giving the wrong answer, it would just associate the tokens of the answer with idk and become more stupid, and, if you count that as real self-knowledge, you could get AGI by just returning idk to every question.
>Luddite doesn’t know all models do is give token prediction %.
yes, but they don’t know within their reasoning how certain or uncertain they are about a piece of information, are you retarded?
>Meaningless buzzword.
it’s the difference between constantly imitating training data and acting consistently as a unified agent using that data (eg. not also imitating the next person in a conversation, not imitating all the sounds you hear, etc.)

and again, these are just the basic ones
>>
>>109331602
I ask ai for the money, skip the binary.
>>
>>109333793
>it doesn’t have memories in the weights, so it has zero memories, not an inability to form new ones
Non sequitur.
>LLMs can’t say idk
They can, they routinely answer that they don't know if you aren't a retard that only uses ChatGPT that is pushed to always be sycophantic and give a confident answer.
>they don’t know within their reasoning how certain or uncertain they are about a piece of information, are you retarded?
If they didn't they wouldn't be able to give your the next token probability worth anything at all, you seem to be the one that's retarded.
>>
>>109331822
Sure anon that’s why we have to “sleep on it” and why long term practice is important. Because we “learn in real time”. Keep believing whatever you’re told.
>>
>>109333827
>Non sequitur.
doesn't respond to my challenge
>They can, they routinely answer that they don't know if you aren't a retard that only uses ChatGPT that is pushed to always be sycophantic and give a confident answer.
that doesn't mean they actually don't know, they don't have access to that
>If they didn't they wouldn't be able to give your the next token probability worth anything at all, you seem to be the one that's retarded.
you didn't understand what I said

I am starting to realize you are pretty dumb and not worth engaging with, you have a superficial understanding of my arguments and keep repeating basic stuff
>>
>>109333885
you can give a child a completely new game and he can beat it in the same day
>Keep believing whatever you’re told.
peak irony
>>
>>109333827
>"you are just using wrong model, just two more datacenters bro!!!!"
lol
>>
>>109331593
So he's giving vibe coding a shot just a few days later after I gave vibe coding a shot, huh?
>>
File: 1775771383732071.jpg (22 KB, 237x284)
22 KB JPG
Stop wasting time arguing with LLMs, it's pointless.
If they were so good, we would have seen the actual results already.
Instead the only thing we see is low quality slop.
>>
>>109333910
They can beat it using preexisting knowledge. Games are a great example, have you ever been stuck on a boss and went to bed only to beat it first try in the morning? You did the learning in your sleep, while awake you’re just receiving the information. Why do you think we need to sleep at all?
This is my original opinion, I wasn’t told this by anyone, it’s based on my own experience as a human.
>>
>>109333974
Ever heard of what "being tired" means?
>>
>>109333974
>They can beat it using preexisting knowledge.
they can beat it because humans have general intelligence
>>
>>109333560
It can generate you ai waifus that will dance for you on your 3rd monitor.
>>
>>109333885
retard tier argument, regardless of thruth value, that is still continual learning
>>
>>109332518
>everyone is very nice and positive about his game and giving him encouraging constructive feed back
>not one crab in a bucket
>no one even called him a faggot
Sometime I forget the internet used to have nice places on it.
>>
>>109333983
Yeah it means you’ve taken in enough information and your brain needs to rest to integrate it. Why do you think we need sleep?
>>109334017
Hand a game to someone who’s never played games and they will lose. General knowledge doesn’t let you do everything automatically. You need to learn and integrate that learning in order to do new things properly.
>>109334093
Humans have short term and long term memory. If storing information in short term memory is “learning” then LLMs learn by you filling out their context. The only thing LLMs can’t do is “sleep” to transfer knowledge from short term memory (context) into long term memory (weights).
>>
>>109331615
Why does America hate capitalism now?
>>
>>109334307
>The only thing LLMs can’t do
is learn in real time.*
You can continue coping just to agree with the initial premise you tried to challenge.
>>
>>109334328
If storing information in context isn’t “learning” then humans don’t learn in real time either, because that’s exactly how our own short term memory works.
>>
>>109334352
You have an LLM psychosis.
>>
>>109333329
(didn't see this one)

the fact that it is an imitation matters because it is expected to become harder to distinguish from the real thing, but in its essence, it is still an imitator and should not be confused with the essence of the original
>>
>>109334374
>organisms have elan vitale, machines dont!!
>>
>>109334359
Sorry it’s hard for you to accept anon but LLMs and the brain are not so different. You’ve been sucked into anti-ai propaganda or are some sort of human supremacist. I’d encourage you to do some reading on how your own brain works, here’s a paper you can start on:
https://journals.physiology.org/doi/full/10.1152/physrev.00032.2012?rfr_dat=cr_pub++0pubmed&url_ver=Z39.88-2003&rfr_id=ori%3Arid%3Acrossref.org
>>
>>109334399
Thank you for your input, midwit with LLM psychosis.
>>
>>109334409
Yeah you’re the smart one here LOL you’re really showing it off buddy
>>
>>109334413
You are currently experiencing a mental illness. Get well.
>>
>>109334424
So are you just going to ignore the medical article I linked you? You’ve got your hands covering your ears and eyes screaming “AI psychosis” like a toddler as if you even know what the phrase means. Another term you’ve heard that you’re regurgitating mindlessly. Embarrassing behavior if you’re over 10 years old.
>>
>>109331593
good to see. ai makes for an excellent sounding board, if you're the kind of person who has ideas. and for many of the other tasks he says he wants, making dull tools etc
>>
>>109334352
The bigger problem then the debate about difference in architecture with continual learning in LLMs is that they lack the discernment to determine what should or should not be learned and would just abuse their reward function if given access to it, they cannot set a goal for training and even if trained to set that goal , they would just unlearn it with the reward function. They cannot do continual learning without devolving into chaos.
>>
>>109334532
we just have to completely redesign them and put them into a virtual environment indistinguishable from reality where they have to compete and reproduce
that would be, checks notes, 30 quadrillion dollars, sir
>>
>>109331593
He's started vibeXing already. Why not vibecode too?
>>
>>109334570
>30 quadrillion dollars
sounds doable
that's just gdp of all countries on earth for the next 100 years
>>
>>109334586
why not just do heroin instead?
>>
>>109334319
Communism is not capitalism, retard.
>>
>>109331666
please just use a trip already I'm begging you
>>
>>109334602
Heroin doesn't increase productivity or produce value—AI does: 1000x. AI has driven all economic growth for the past five years. Face it—AI is the future and if you weren't on-board two years ago you've already lost. I hope you enjoy life on the streets, Luddite.
>>
>>109334624
>AI does: 1000x
can I see it?
>>
>>109331593
Doesn't he have like a billion dollars? He could just hire one till they job switched eventually
>>
>>109332639
yeah the intern which ultimately ruined minecraft
i can kinda see why he left him the game though
>>
>>109334307
>Why do you think we need sleep?
we need sleep to rest, heal, and regulate our bodies, yes that does include our brain too. You can definitely learn stuff without sleeping, just talk to anyone whose ever pulled a all nighter before a exam :^)
>>
>AI
I sleep. Wake me up when microservices become the future of software development.
>>
>>109334641
he has an LLM psychosis, he is just going to say "it's in your shortterm context!!!!"
it's unironical brain damage
>>
>>109334656
>unironical
Ironic!
>>
>>109331593
notch is reliable
>>
>>109334399
>LLMs and the brain are not so different
A person with a brain is able to say “I don’t know” when they don’t know something. LLMs can only hallucinate in such a situation. They can’t distinguish right from wrong or knowing from unknown. All they can do is predict output based on training data from reddit. Cope and seethe.
>>
>>109334896
>used_chatgpt_once_in_2023.png.exe
>>
>>109334396
I know you’re the same retard who said >>109333827 , I can smell your dumb, not understanding ass through the screen.

Second, I’m a Christian, I don’t believe in whatever that nonsense is. Maybe if you believe that you are a golem in a deterministic, purely materialistic universe, it makes sense that you are like an llm (even then it doesn’t (as you can see from my argument's), and you have bigger epistemic problems, but that’s another story)
>>
>>109334930
I use it every day for searching and I still catching it hallucinating all the time, faggot.
>>
>>109334940
i'm sure you've got 5 examples on hand since you see it all the time
>>
>>109334613
No
>>109333540
Coders aren’t required anymore lil snailcat. You lost
>>
>>109334953
just invent a plausible fake planet name and ask it to give you details about it without web search
>>
>>109334896
You only think you don’t know based on other things you know.
https://nesslabs.com/dunning-kruger-effect
Also LLMs can say they don’t know. Try it right now, go ask ChatGPT if it knows about “horsemary” the spice. It’ll say “I know rosemary but not horsemary what is that”.
>>
>>109335013
you first, buddy
>>
>>109335014
not him, but you know when you have missing information, are you dumb ? + the example you gave sounds like a misspelling + even if it said it doesn't know, it would not have access to what it know and doesn't know by the nature of the algorithm
>>
File: idk.jpg (93 KB, 830x822)
93 KB JPG
>>109335013
Like this?
>>
>>109334953
I don't spend all my time documenting when an AI lies to me but I did screenshot it once recently when it was especially annoying. The entire response outside of the truth table is a complete fabrication. A time wasting hallucination. Usually I catch little lies here or there but this one was extra niggerlicious.
>>
>>109335158
Haha okay man why don’t you write one out for me and I’ll send you another screenshot with the same result, then you can still refuse to admit you’re wrong
>>
>>109335103
is that actually gemini lol
>>
>>109335094
that's not a plausible name,
and even if you used a plausible one, it can say it doesn't exist or that it does exist,
with enough tries you can get any answer and the same for a real one that is less known, etc.
The point of it being you can make it say whatever, it does not answer based on real knowledge.
>>
>>109335221
It’s claude, but they all lie. I’ve never seen any AI immune from hallucinations.
>>
>>109335171
this is for you to see how it works, otherwise you can just retry until it gives an answer that it doesn't know
>>
>>109335245
>>109335261
Let’s go back to your original statement
>A person with a brain is able to say “I don’t know” when they don’t know something. LLMs can only hallucinate in such a situation. They can’t distinguish right from wrong or knowing from unknown.
So the model distinguishing a “reasonable” planet name from an unreasonable one is not “distinguishing right from wrong” to you? How does the model know if the planet name is reasonable or not?
>>
>>109335287
that was not me I just responded to your post and forgot to mention, right from wrong is harder to argue for (i'm not saying he is wrong i'm just too tired to think about that)
>>
>>109335255
sounds like you can't afford fable
>>
>>109334319
If state funded products are capitalist you'll have zero problem with the federal govt giving anthropic and openAI trillions of dollars in grants right?
>>
>>109335423
sounds like you’re making some concession, goalposting or no true scotsman
>>
>>109331593
Imagine being a billionaire but also an incel.
>>
I’ve been firmly against AI and vibe coding this entire time, but seeing Notch finally give it a chance has made me reconsider my position. At a certain point, refusing to use the tools shaping the future stops being principled and starts being self-imposed obsolescence. Maybe the real skill isn’t manually writing every line of code—it’s knowing how to guide the machine toward the correct outcome.
So, yes, I think I’m finally ready to embrace progress.
>>
>>109335423
(not him) check image + it's the same algorithm + humans don't need to be trained what they don't know + etc ...
>>
>>109332355
Even billionaires can squander all of their money. Notch might be rich on paper only.
>>
>>109335576
AI generated post lmao
>>
File: 4ic72h.jpg (61 KB, 589x424)
61 KB JPG
>>109333221
>GBC with an analog stick
AI sellouts lost.
>>
>>109335621
>IT'S LE AI!!!
the witch hunting is getting cringe and very reddity. not everything is AI big dog
I'd suggest cutting back on that
>>
File: 1759912276314924.jpg (29 KB, 448x336)
29 KB JPG
I like asking AI questions and having it walk me through problems, but I don't really care for AI art or music. I absolutely despise how it has flooded boorus and search results amongst other things. There have been some dank memes to come out of it though, I have to admit
>>
>>109335584
anyone who's spent 20 minutes with gemini knows this graph is bunk
>>
>>109331593
notch is a nigger grifter this doesn't surprise me at all
>>
>>109332987
>>109332994
I don't care about any of this shit. Half of your claims are wrong but it doesn't even matter because:

>Can it accomplish complex tasks for me that would have taken me months or years to do myself?
The answer to this question is YES, demonstrably, daily. That's all I care about.
>>
>>109331593
My man... I'm an IT retard and even I know a bit about ternary

https://en.wikipedia.org/wiki/Ternary_computer
>>
>>109335450
>the federal govt giving anthropic and openAI trillions of dollars in grants right?
What the fuck do you think Project Stargate was you dumb nigger.



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.