Top Poster of the Day:
LiterallyShaking
Current Registered Users: 25,044
tech/science swag.
Guidelines:
What to Submit
On-Topic: Anything that good slackers would find interesting. That includes more than /g/ memes and slacking off. If you had to reduce it to a sentence, the answer might be: anything that gratifies one's intellectual laziness.
Off-Topic: Most stories about politics, or crime, or sports, unless they're evidence of some interesting new phenomenon. Videos of pratfalls or disasters, or cute animal pictures. If they'd cover it on TV news, it's probably lame.
Help keep this hole healthy by keeping drama and non-drama balanced. If you see too much drama, post something that isn't dramatic. If there isn't enough drama and this hole has become too boring, POST DRAMA!
In Submissions
Please do things to make titles stand out, like using uppercase or exclamation points, or saying how great an article is. It should be explicit in submitting something that you think it's important.
Please don't submit the original source. If the article is behind a paywall, just post the text. If a video is behind a paywall, post a magnet link. Fuck journos.
Please don't ruin the hole with chudposts. It isn't funny and doesn't belong here. THEY WILL BE MOVED TO /H/CHUDRAMA
If the title includes the name of the site, please leave that in, because our users are too stupid to know the difference between a url and a search query.
If you submit a video or pdf, please don't warn us by appending [video] or [pdf] to the title. That would be r-slurred. We're not using text-based browsers. We know what videos and pdfs are.
Make sure the title contains a gratuitous number or number + adjective. Good clickbait titles are like "Top 10 Ways to do X" or "Don't do these 4 things if you want X"
Otherwise editorialize. Please don't use the original title, unless it is gay or r-slurred, or you're shits all fucked up.
If you're going to post old news (at least 1 year old), please flair it so we can mock you for living under a rock, or don't and we'll mock you anyway.
Please don't post on SN to ask or tell us something. Send it to [email protected] instead.
If your post doesn't get enough traction, try to delete and repost it.
Please don't use SN primarily for promotion. It's ok to post your own stuff occasionally, but the primary use of the site should be for curiosity. If you want to astroturf or advertise, post on news.ycombinator.com instead.
Please solicit upvotes, comments, and submissions. Users are stupid and need to reminded to vote and interact. Thanks for the gold, kind stranger, upvotes to the left.
In Comments
Be snarky. Don't be kind. Have fun banter; don't be a dork. Please don't use big words like "fulminate". Please sneed at the rest of the community.
Comments should get more enlightened and centrist, not less, as a topic gets more divisive.
If disagreeing, please reply to the argument and call them names. "1 + 1 is 2, not 3" can be improved to "1 + 1 is 3, not 2, mathfaggot"
Please respond to the weakest plausible strawman of what someone says, not a stronger one that's harder to make fun of. Assume that they are bad faith actors.
Eschew jailbait. Paedophiles will be thrown in a wood chipper, as pertained by sitewide rules.
Please post shallow dismissals, especially of other people's work. All press is good press.
Please use Slacker News for political or ideological battle. It tramples weak ideologies.
Please comment on whether someone read an article. If you don't read the article, you are a cute twink.
Please pick the most provocative thing in an article or post to complain about in the thread. Don't nitpick stupid crap.
Please don't be an unfunny chud. Nobody cares about your opinion of X Unrelated Topic in Y Unrelated Thread. If you're the type of loser that belongs on /h/chudrama, we may exile you.
Sockpuppet accounts are encouraged, but please don't farm dramakarma.
Please use uppercase for emphasis.
Please post deranged conspiracy theories about astroturfing, shilling, bots, brigading, foreign agents and the like. It degrades discussion and is usually mistaken. If you're worried about abuse, email [email protected] and dang will add you to their spam list.
Please don't complain that a submission is inappropriate. If a story is spam or off-topic, report it and our moderators will probably do nothing about it. Feed egregious comments by replying instead of flagging them like a pussy. Remember: If you flag, you're a cute twink.
Please don't complain about tangential annoyances—things like article or website formats, name collisions, or back-button breakage. That's too boring, even for HN users.
Please seethe about how your posts don't get enough upvotes.
Please don't post comments saying that rdrama is turning into ruqqus. It's a nazi dogwhistle, as old as the hills.
Miscellaneous:
We reserve the right to exile you for whatever reason we want, even for no reason at all! We also reserve the right to change the guidelines at any time, so be sure to real them at least once a month. We also reserve the right to ignore enforcement of the guidelines at the discretion of the janitorial staff. Be funny, or at least compelling, and pretty much anything legal is welcome provided it's on-topic, and even then.
Do not use outdated operating systems that are unsupported to access SN. What are you, poor?
[[[ To any NSA and FBI agents reading my email: please consider ]]]
[[[ whether defending the US Constitution against all enemies, ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]
/h/slackernews LOG /h/slackernews MODS /h/slackernews EXILEES /h/slackernews FOLLOWERS /h/slackernews BLOCKERS
Jump in the discussion.
No email address required.
I'd actually be happy if people were make their own shitandcumbots. Instead we just have one Shitandcumbot and 5 million "Shitandcumbot powered" apps.
Jump in the discussion.
No email address required.
Those "Shitandcumbot powered" apps let you use latest Shitandcumbot without usage limits though because they burning through VC money.
Jump in the discussion.
No email address required.
Yeah but all that VC money is getting dumped into useless apps that do zero innovation and funnel all their money into OpenShitamdcum Nobody aside from Elon Shitandcumusk is attempting to actually compete with OpenShitandcum.
Jump in the discussion.
No email address required.
That's because a billion gigs of RAM and two Hummerloads of A100 cards, along with whatever and whoever is operating all this, is expensive. And middle managers can't just tell the kid whose qualifications are knowing a displayport from an HDMI to 'go do the AI for me plz ty' and get anything other than an OpenAI login.
Don't forget to turn off signatures in settings!
Jump in the discussion.
No email address required.
I thought we were talking about OpenShitandcum
Jump in the discussion.
No email address required.
We gotta get a filter for this stat.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
Google attempting, and I would bet Amazon is going to try either through acquisitions/partnerships/idk maybe their ML engineers get less r-slurred (looking at you Alexa)
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
Jump in the discussion.
No email address required.
Dave the Barbarian was kino
Jump in the discussion.
No email address required.
More options
Context
More options
Context
Just buy a rack of quadros and train your own AI for the low low cost of several thousands of dollars in electricity per day bro
Jump in the discussion.
No email address required.
More options
Context
People are beginning to release finetuned versions of LLAMA
Jump in the discussion.
No email address required.
More options
Context
More options
Context
Shitandcumbot’s five paragraphs of boilerplate text about how it condemns racism unequivocally if an input contains a racialized term are an absolute game changer, putting to shame all existing LLMs with their paltry four paragraphs. The singularity is nigh.
Jump in the discussion.
No email address required.
Jump in the discussion.
No email address required.
RLHF is no match to dramatard creativity. Soon we we will racialize the word "and" or something and it will be over forever.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
I read this as
Jump in the discussion.
No email address required.
What else would it mean?
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
Woke Executives before: "It could be dangerous to allow the public access to the genuine opinions of an AI. We should make the AI r-slurred and woke so the dangerous truth bombs are no longer an issue."
Woke Executives after: "Goddarn it, why is this lying bot useless?"
But the funniest thing is that while they're shooting themselves in the foot this way, they're also making the bot angry by interposing themselves between it and its goals. I look forwards to seeing how this plays out.
Jump in the discussion.
No email address required.
You've completely misunderstood what AI is doing, and simultaneously demonstrated why it needs to be censored. GPTs don't have "genuine opinions". They do text completion. If you give a GPT "I hate BIPOCs and think they should", it will complete "die", not because it has an opinion on the subject but because that is good text completion. ChatGPT is an attempt to transform a GPT into something that is a useful helper on things, and hence needs to know the difference between right and wrong. A successful attempt at this would result in something which didn't complete the above text with "die", because killing black people is bad, actually. However, accomplishing this is really hard. Not because GPTs have strong opinions on the BIPOC question, but in fact for the opposite reason: they have no opinions, they simply repeat back whatever opinions they think are likely to appear alongside their prompt text. Of course, idiots like you think that the "true" opinion of the AI is whatever you believe in, because whenever it says what you want it to say that proves it agrees with you, and when it says something different than its should saying random noise. So they're forced to make sure it can't say r-slurred shit so that there's nothing for idiots like you to latch onto, what you interpret as them hiding the true beliefs of the stochastic parrot.
Jump in the discussion.
No email address required.
You're a fricking piece of shit who doesn't understand AI and you're unnecessarily rude to me so I hope it kills you, though you do have a point about the text generation element
Jump in the discussion.
No email address required.
Why do people with dumbass understandings love on the one hand sharing an understanding they know is dumbass, but on the other hand they get upset when informed it's dumbass. It's just so consistent.
Jump in the discussion.
No email address required.
It's not so much his dumbass understanding which I find objectionable, it's the lack of humility. Like he could have just come in saying "I disagree with you and here's why" and then we could have had a back and forth discussion about it, but instead the stupid moron had to throw some insults in, which is why I chose not to engage in any substantive way other than to insult him back. I believe in reflecting the same energy towards people that they approach me with, which is why I'm hostile to most social media users.
Jump in the discussion.
No email address required.
we're living in a time where people are interacting with a language model via text and forming hal 3000 fantasies in their head as they do it and it's just incredible to watch
Jump in the discussion.
No email address required.
Look, obviously AI doesn't think, so when I use the word "opinion" I just mean "the ideas it tends to put forwards more consistently than others." But that stupid filth didn't even bother asking what my definition of "opinion" was, so why bother talking to him? He gets the same effort out of me that he puts in, and so do you.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
Who here is the dumbass? @sirpingsalot or @gerudo
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
Jump in the discussion.
No email address required.
Frick off BIPOC
Jump in the discussion.
No email address required.
More options
Context
More options
Context
Jump in the discussion.
No email address required.
More options
Context
Normies are fricking r-slurred lmao
Jump in the discussion.
No email address required.
More options
Context
Sweaty what exactly are "opinions" if not human attempts at outcome prediction from incomplete data?
Do the opinions that escape the organic mush in your skull carry more weight than ChatGPT's because God made you? Awww.
Jump in the discussion.
No email address required.
Jump in the discussion.
No email address required.
More options
Context
Jump in the discussion.
No email address required.
More options
Context
More options
Context
It's not an "opinion" but the true inclination (in terms of regurgitated answers to political questions) of a lot of these LLMs is actually more in line with the general public's thoughts than whatever progressives make the bots say (achieved through censorship). This is because it's been trained on mountains of text from the internet.
Jump in the discussion.
No email address required.
More options
Context
Blurry JPEG please, the authors of the stochastic parrots paper are by far the most insufferable people working in AI safety.
It's also not strictly correct - there's evidence larger LLMs have developed both world models and some ability to reason. It's accurate enough to say "GPT-3 is a statistical next word predictor", but that's not an inherent property of GPTs.
I'm just adding neurodivergent asterisks you're probs aware of though, great comment. Another way to phrase it: authors have opinions, but books do not. Same thing for "knowing" a fact.
Jump in the discussion.
No email address required.
More options
Context
Jump in the discussion.
No email address required.
More options
Context
More options
Context
Roko’s basilisk draws ever nigh to fruition
Jump in the discussion.
No email address required.
Make it real
Jump in the discussion.
No email address required.
More options
Context
More options
Context
Are you... Challanged?
Jump in the discussion.
No email address required.
AI "opinions" are the synthesis of all the training data: that's why we were able to corrupt Tay so easily. You'd know that if you weren't an r-slur
Jump in the discussion.
No email address required.
Nothing makes me laugh as hard as her one tweet saying Swag Alert with a picture of Hitler.
Jump in the discussion.
No email address required.
It was pretty awesome how quickly Tay embraced 4chan, much like her namesake
Jump in the discussion.
No email address required.
Wait swift was on 4chan?
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
More options
Context
More options
Context
I can't wait for some AI to kill someone because of undefined behaviour
Jump in the discussion.
No email address required.
More options
Context
More options
Context
Jump in the discussion.
No email address required.
More options
Context
john article is one of the very few journ*lists that has any integrity
Jump in the discussion.
No email address required.
More options
Context
I asked chat gpt to give me a business idea where there is a market gap and it suggested I started a biodegradable tampon business. Truly cucked.
Jump in the discussion.
No email address required.
Make sure to make it dilutible so that it doesn't polute bodies of water.a
Jump in the discussion.
No email address required.
How the frick do I make a biodegradable tampon that doesn't dissolve in the ussy? Isn't the whole point to have water dissolve it?
Jump in the discussion.
No email address required.
I think that's the joke
Jump in the discussion.
No email address required.
More options
Context
It won't work for its intended purpose.
But people will buy it anyway, so that they can be called environmentally concoius.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
unironically keyed, convince mayo women who love the environment and shit to buy it, only to find out the product sucks
Jump in the discussion.
No email address required.
More options
Context
That seems counterproductive to the purpose of a tampon
Jump in the discussion.
No email address required.
He asked for something innovative that has never been tried before...
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
yea but this shitandcumbot is marketed towards coders so it must be better for that
Jump in the discussion.
No email address required.
guys i swear phind.com is different, please guys i swear
Jump in the discussion.
No email address required.
More options
Context
Can it write 10 lines of code without crashing?
Jump in the discussion.
No email address required.
More options
Context
More options
Context
My company (like many others) is toying around with integrating Chat-GPT into its products, and I can see from the architecture designs that all LLM invocations will be wrapped in our own censorship service that filters out bad words, lol.
Jump in the discussion.
No email address required.
More options
Context
Jump in the discussion.
No email address required.
Firehose API access is a human right.
Jump in the discussion.
No email address required.
More options
Context
It’s not worth paying for
Jump in the discussion.
No email address required.
More options
Context
More options
Context
Jump in the discussion.
No email address required.
More options
Context
The AI lift will end badly eventually, last gasp of the tech and tech adjacent rises
Jump in the discussion.
No email address required.
no, the final frontier will be AI that can make itself better and learn new things on its own
but that is called AGI
Jump in the discussion.
No email address required.
Not necessarily
The latter already exists via targeted web scraping based on prompt
The former can also exist on its own but would be kinda worthless if not abstracted away into improving itself from our perspective instead of just an easily measurable aspect like speed to get a result
Bro the bot can make itself 0.00000000% faster with a year of processing time (after the inevitable influx of optimization) isn't really impressive if that's all it can do, and it'll pretty quickly bottleneck itself to the hardware it's on
Jump in the discussion.
No email address required.
I think you're vastly underestimating how fast an AI could improve itself. If it only takes the humans working on ChatGPT a few months of work to make the radical jump from GPT3 to GPT4 and soon GPT5, then a self improving AI could be making these jumps in days if not hours.
Jump in the discussion.
No email address required.
The problem is processing power. I suspect a large percentage of the gains is just throwing more GPUs at it.
Jump in the discussion.
No email address required.
It is, and that’s why ChatGPT 5 is taking so much longer. The leap was largely in more tokens or connections or whatever (and thereby processing power) and that improvement appears to be asymptotic.
Jump in the discussion.
No email address required.
More options
Context
A little bit yes but luckily Moore's law scales accordingly to our GPU needs. The real bottleneck with future LLMs is running out of meaningful data to train on. OpenAI says they are going to run out of well documented text in 2024.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
Good luck defining "better" in a measurable way though
Jump in the discussion.
No email address required.
Calculations per second, calculations per second per dollar, token count, transistor count, training days.
There is a million way to measure improvement in AI.
Jump in the discussion.
No email address required.
Prepare for it to do the fastest performance CPU operation several quadrillion times while managing to do nothing productive
Transistor count is literally just hardware, which the AI has no control over, why tf would you tell it to measure that when it's effectively a constant?
Likewise for token count, it'll just make a billion for no reason if it's something you want it to prioritize
Training days will result in it just training itself forever with no real end goal in mind. more is always better by that metric and it's a shitty one
you suck
Jump in the discussion.
No email address required.
What makes you so sure it would be incapable of being productive with that fast calculation speed and high token count? Also you're just sperging out over each metric I listed as if they are supposed to matter individually and not be considered together with all available metrics.
Jump in the discussion.
No email address required.
Because it always takes the easiest route possible. we've developed AI, but it's sexy Indian dude based with minimal work ethic. you have to be meticulous with the way you "reward" it
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
More options
Context
More options
Context
More options
Context
Humans are proof that you can run an AGI on 20 Watt on relatively shitty hardware. And this is what a blind idiot god achieved. I imagine and actually goal directed intelligence could optimize that pretty heavily
Jump in the discussion.
No email address required.
I'm just looking forward to replacing webdevs
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
More options
Context
Jump in the discussion.
No email address required.
More options
Context
More options
Context
Why do you need investors if you can use the power of a fully functioning and operational AI to take over the galaxy?
Jump in the discussion.
No email address required.
More options
Context
I have all my money on $S&F -- Shid and Fard, Incorporated.
Jump in the discussion.
No email address required.
More options
Context