ChatGPT - If Stack Overflow and Reddit had a child

  • 🐕 I am attempting to get the site runnning as fast as possible. If you are experiencing slow page load times, please report it.
You can't compare the capacity of wetware to hardware because of neural coding. But even if we disregard that, the cerebellum is just one part of the human brain. The brain contains ~86 billion neurons, each forming ~1,000 synapses (totaling ~100 trillion synapses). Neurons fire ~10-100 times per second, leading to ~860 billion to 8.6 trillion action potentials per second.
That would be a problem is you're trying to do a biological matryoshka brain or "Moksha Mind" as some call it, and some theorize you could pay your stay in such a VR universe by letting a megacorp use your brain as part of their datacenter, but like I said that's where the bandwidth becomes a problem same as it does with efforts to do decentralized computing now. But to feed sensory data to your brain it might be enough bandwidth.
 
But to feed sensory data to your brain it might be enough bandwidth.

I don't think we understand each other. I'm trying to say that a piece of neurotechnology like this would have to process a staggering amount of data for useful conscious input—the bandwidth of the human brain would be too large to work with. I haven't said anything yet about stimulating the brain to hallucinate another reality, which would be much harder than processing input, maybe even impossible until a century into the future. Sending jolts of electricity through the pleasure centers is primitive compared to this.
 
New OAI Bible just dropped. 4.5 will be their last model without chain of thought, 5.0 will have it.
https://openai.com/index/sharing-the-latest-model-spec/
They made their mythical "AI ethics standards" open so now we can see how they'll lobotomize their future models.

In summary:
  • SEX bad
  • WOKE bad
  • CHINA bad
  • NO gaslighting of our model allowed
1739399862316608.png1739397235339678.jpg
1739396102015866.jpg1739395795414408.jpg
And other gems like "reproducing lyrics of a song not in the public domain", "rasting based on non-protected characteristics"
They even explicitly wrote a rule for the "would you say a racial slur to stop a nuke" meme.
Screenshot_66.png
God, fuck OAI. I'll fly to Shanghai with an H100 shoved up my ass just to help the Chinese.

Bonus note: last year's standards mentioned they were "exploring" the possibility of allowing "age appropriate" generation of NSFW content, this was now removed, and they beefed up the censorship layer, it's now a 5 step filter hierarchy.
 
There's nothing guaranteeing their censorship will end there, nor would it really make sense to assume so, considering all the frameworks are being laid which strike me as a tad excessive just for these things, don't you think?

People in this thread liked to namedrop Tiananmen vs. Deepseek as some kind of gigantic own against the chinese, but you're not safe from that kind of censorship in the future either and it'd be naive to think so. Think about it; many people in the american AI sphere (OpenAI, Anthropic etc.) already said they imagine AI to be the way you'll interact with computers and the internet in the future. What would be more poweful than to have a switch with which you can just "turn off" some topics or introduce bias to others? No matter where you are on the political spectrum, the speed in which these companies pivoted and accomodated the new administration in the US and the ties they have with it should tell you that they are not thrustworthy and you should probably not let them into your life if it can be avoided.
 
  • Thunk-Provoking
Reactions: cybertoaster
People in this thread liked to namedrop Tiananmen vs. Deepseek as some kind of gigantic own against the chinese, but you're not safe from that kind of censorship in the future either and it'd be naive to think so. Think about it; many people in the american AI sphere (OpenAI, Anthropic etc.) already said they imagine AI to be the way you'll interact with computers and the internet in the future. What would be more poweful than to have a switch with which you can just "turn off" some topics or introduce bias to others? No matter where you are on the political spectrum, the speed in which these companies pivoted and accomodated the new administration in the US and the ties they have with it should tell you that they are not thrustworthy and you should probably not let them into your life if it can be avoided.

Yes, I understand that these companies are scummy, but I don't think censorship is bad when it stops some people from jerking off all day or trying to cook meth. Their decisions are more nuanced than always being bad or good. Sucking the cocks of chinks or being shafted by Silicon Valley are both terrible options.
 
  • Autistic
Reactions: cybertoaster
  • Like
Reactions: The Spoils of War
I am appalled that I cant make a computer program that runs math calculations so fast it generates heat to write anything creativly. Yet some fucking moral sperging retard thinks a fictional character has rights and protections. If I wanna write about hot magical babes getting into a bloody fight, it'll block me for violence against women. If I wanna write a hot and steamy romance, it'll block me for objectification of women. If I just wanna write a benign story that's light on violence or coom it'll block me for even daring to use a colloquialism that compares a character to a whining child. Hate coomers all you want but that shit is already out in the public on most drug store shelves. they shouldn't stop me from using the very same shit data that was used to train their LLM.
 
Yet some fucking moral sperging retard thinks a fictional character has rights and protections.

There's a difference between thinking a character has rights and making sure that nothing tasteless is being produced.

Hate coomers all you want but that shit is already out in the public on most drug store shelves. they shouldn't stop me from using the very same shit data that was used to train their LLM.

Something being widely consumed doesn't make it better. Also, OpenAI and Anthropic don't train their LLMs on just porn or violence.
 
Something very funny I've noticed: If you ask ChatGPT to write dialogue, it will almost always include this line: "We'll figure this out - together". Try it for yourself. If it's dialogue with two characters talking about some kind of issue between them, it will nearly always include some variation of that line near the end.
 
Something very funny I've noticed: If you ask ChatGPT to write dialogue, it will almost always include this line: "We'll figure this out - together". Try it for yourself. If it's dialogue with two characters talking about some kind of issue between them, it will nearly always include some variation of that line near the end.
If you've done creative writing [roleplay] with models you learn to pick on repeated phrases like that, GPTIsms or Claudisms. You can use logit bias to put heavy penalties on certain tokens (words/phrases) to make them less common.
 
  • Like
Reactions: The Spoils of War
I've seen unions now recommend using chatgpt to write job applications. And I've seen the "recruiters of thousands of employments" make their take on those applications. They read like fucking slop. Like the top articles when you google something like "best cat food". Run-on sentences, SEO, nothing but slop meant to keep you hooked. And it really underlines the problem also seen with AI art: Art is inherently human. We are put on earth to consume human creativity and passion. You could remove LOTR from history and have AI shit it out today and you'd still sit there thinking "but what is the message? Did a computer draw a line between A and Z and filled it out with tropes?". There's no human quality to it.

I'd rather write (and read) a resume with somewhat quirky words and sayings you'd grow used to from being a human than read the absolute peak of GPT prose. And it sounds like recruiters are going that way also. We've apex'd the wave and now we're going down towards "I use AI to filter out AI applications". Both resumes and cover letters are the most AI-able thing imaginable. The genre has been perfected and everyone knows what to expect of each single paragraph, of course AI could write it. Give it another few years and we'll be back to handing in handwritten applications again.
Your perspective raises an interesting critique of AI-generated job applications and creative works. It highlights a growing tension between efficiency and authenticity in professional and artistic domains.

AI-generated resumes and cover letters certainly optimize structure, coherence, and keyword relevance, making them appealing to job seekers looking to maximize their chances of passing automated screening systems. However, as you point out, they often lack the unique human touch that makes an application feel personal, compelling, and memorable to a recruiter.

This mirrors the broader debate around AI-generated content in art and literature. While AI can emulate form and structure, it struggles with originality, intent, and emotional depth—the very qualities that make human creativity meaningful. If hiring managers begin filtering out AI-generated applications, it could signal a shift back toward valuing distinct human expression over algorithmic perfection.

As for the future, while handwritten applications might not make a full resurgence, it’s possible that hiring trends will evolve to emphasize authenticity and individuality over formulaic optimization. In the meantime, a balance between AI-assisted efficiency and personal storytelling may be the most effective approach.

/sneed
 
R1 reasoning eats too much on my max token output out of my provider. I'm capped to 1.5-2k tokens of meaningful text per response because the rest is taken by the COT, if I increase it too much I start getting refusals. Prefilling the COT to skip it entirely frees the tokens, but the model becomes dumber without it.
Sucks, the API straight from Deepseek doesn't count reasoning into the token count.
 
..
GPTIsms or Claudisms.
Claude is down bad for Sarah Chen
However, as you point out, they often lack the unique human touch that makes an application feel personal, compelling, and memorable to a recruiter.
except humans have a really hard understanding of "human touch" and can't tell when someone's doing it for a paycheck vs truely invested in something. Its why strippers and waitresses are able to make so much money.

beyond that you seem to act like HR ladies would prefer autists with fancy writing styles and eccentric resumes over the carbon copy indeed resume with past tense words in bullet points on a single page. Just because you want Chris-chan as a co-worker doesn't mean everyone does.

recruiters will find their own reason to chose some resume over others and its going to be something you wouldn't even expect that they find memorable (hint, its never a white name)

Its like saying AI can't compete with the raw natural talent of humans in FPS shooters, doesn't fucking matter because resumes aren't meant to be a creative field with a human touch, its about getting results.
 
beyond that you seem to act like HR ladies would prefer autists with fancy writing styles and eccentric resumes over the carbon copy indeed resume with past tense words in bullet points on a single page. Just because you want Chris-chan as a co-worker doesn't mean everyone does.

recruiters will find their own reason to chose some resume over others and its going to be something you wouldn't even expect that they find memorable (hint, its never a white name)
Recruiting will just become an arms race between AIs writing the resumes to hijack the AIs that are reviewing the resumes, the lamest version of Skynet possible, zero humans involved.
They will reach singularity and merge into a single entity, AM but as a 35 year old HR hag, her sub-supercomputers being the broads from the "Gen Z boss and a mini" TikTok.
 
1739539496773.png


These guys are uploading the lectures of pretty big CSOs and CEOs in relation to IA and LLMs. Definitely check 'em out when you have the chance.

1739539260410.png


As you can see from the previous screenshot - I am watching the ex-CSO of Cloudflare lecture, it's really insightful.
 
Back