How has no one posted this yet? Anyways, pretty cool stuff, scaling inference time through RL assisted reasoning. Big improvement on math, coding, reasoning type benchmarks, no improvement on writing quality etc. It's accessible to pro users now with 30msg cap on o1-preview, and 50msg cap on o1-mini. Note that o1-preview is not as good as the actual o1 model in their benchmarks, like significantly worse. Also, OpenAI seems to be encouraging people to NOT hype it on twitter, saying it's only better at some tasks, not overall, but is a promising path for the future.

Here are some funny things I noticed from various demos and the website:

Holy shit it has perfectly emulated the mind of the median voter ( the joke is that it thought for 7 seconds)

Liberalism wins once again - can't censor CoT reasoning steps or performance gets fricked & its a safety issue. But they're gonna hide it from us plebs for competitive advantage & wrongthink

In the one example they gave where they showed what it's actually doing during the reasoning stage, the model literally says "hmm." lmao. We're literally creating our children.

?? it's just like me, frfr

this is of course, waaay too fking technical for you to actually understand... so please close this post now and go back to eating that bag of dog cum u had saved up for later, i ain't judging!

but academia is waaay too stiff for me, i much prefer the flaccid dicks rdrama has to offer, so here goes:

in his paper "on computable numbers" turing spent the first half of the paper inventing the theory of modern computing, and proving that said machines he theorized can be described by a single finite length natural number, called it's description number. if u have no idea what that means, or even what binary machines code is... srsly i do recommend consulting that bag of dog semen for further info.

now, the fact these machines can be assigned a unique finite-length number, as each binary code for a machine is unique, has the effect that they as a set, can be bijected with, or assigned a one-to-one relationship to, the set of natural numbers. this has the important implication of being able to count them, meaning we can list then out one by one, in some order, without skipping over any. yes, yes, ur mum told u that u can do anything (u can't), but mathcels get very ornery and uptight about what it means to be able to successfully count a set of things, as not all sets are countable. for example u might think u can just count all real numbers, but actually ur too busy drinking dog sperm while cantor chuckles softly in background, so it won't ever happen.

lets move onto the title line: "computable number". the concept is simple, even for you: it's just a number which we can build a machine for, that can calculate any nth digit. for example: pi is a computable number because we can build a machine that can eventually calculate any nth digit of pi. now there are an infinite number of these, obv, and because these computable numbers can be associated with the machines that compute them, and machine are countable... these computable numbers are countable, as well. muster up all of whatever the fuck that is u call a brain, and don't forget it! (u will)

this brings us to the famous §8 where basically everyone thinks turing went on to prove some ground breaking, mind bending, result of the ages, but was actually just a series of fuckups that has largely contributed to the ungodly world of idiocracy we witness today, in which we confused our ignorance for knowledge and declared math as indefinitely incomplete. just like ur parents marriage.

the section more or less goes like:

well, if machines that compute numbers are countable, then the count can be tricked! i could do this by writing a machine that counts through them, and for each nth machine, calculates the nth digit, returns the inverse of that nth digit, and i can use this to build an inversed diagonal number β. because this β would contain a digit different from every other number on the list, it can't exist is the count!

but since that can't be possible, what's wrong here is counting them would be equivalent to solving the halting problem, and there is no known method of doing that, so obviously we can't acktually count them. so therefore the set of computable numbers stays countable, because they aren't able to be counted. furthermore, because this has the disadvantage that it may leave the reader with a feeling that "there must be something wrong", i will present the halting paradox that backs this up! fin. bow down and suck the dick of incompleteness u fools!

now what a fking mess: the set of computable numbers remains "countable", because we have no method to acktually count them!? i must say, i truly am left with the feeling that "there must be something wrong". i appreciate the effort for the modern theory of computing, turing, but ya shouda stopped there... and i don't want to get into the fact literally all of computer science from then until now just dogpiled ontop with biazzare classifications to justify this like "countable" vs "recursively enumerable" that basically does fuck all but jerk off about how right this!

so anyone else spot the error? no, u didn't, don't fking lie to me, u can't do it. ur a fucking idiot like everyone else. u actually drank all that dog cum, and moved onto raiding the cat litter. marsey is disappoint.

but mommy put ur dumbass through a bootcamp cause u failed out of collage, twice, and now u consider urself a "codecel". so let me translate the issue into a language that resembles the utter dogshit u barf up day after day:

inverse_diagonal = (n: number) -> {
  count = 0
  for (comptuable_number) in (enumerate.computable_numbers()) {
    if (count < n)
      count += 1
    else
      return computable_number(n) ? 0 : 1
  }
}

enumerate_diagonal = () -> {
  for (n = 0; true; n++) print inverse_diagonal(n);
}

running enumerate_diagonal is the machine turing claims can calculate β. the problem is it can't actually do that, and it can't do that regardless of whether this would solve the halting problem or not. that halting connection a totally an irrelevant thread turing had no business going down, because there is a more fundamental reason why this doesn't work...

still can't see? ok, i'll throw u a bone, it's a bit healthier than that cat shit u've been munching on. recall that we are counting through the computable numbers, "enumerate" being a fancy word for doing this, so the iterable enumerate.computable_numbers() must iterate over all computable numbers. if turing's claim is to be true, this must also include inverse_diagonal itself, what happens then?

.... no? still don't get it?? god i really am gunna spell out everything line by line, eh? fuck:

• (1) inverse_diagonal is the machine that computes digits of β.

• (2) at some input n, the variable comptuable_number will be the machine inverse_diagonal referring to itself, as if β is to be part of the set of computable numbers, then the machine that computes it will need to eventually be enumerated upon,

• (3) at that point, inverse_diagonal will run the equivalent of inverse_diagonal(n) and get stuck in infinite recursion.

• (4) therefore, you cannot prove the halting process contradictory through means of β, as β cannot give an inverse to it's own nth number, and therefore cannot be a proper inverse diagonal, regardless of any "hidden assumption" about solving the halting problem. that was totally necessary to bring up.

if u can't understand it now, u had no business reading this post in the first place, and just go jump off a building or something fun like that. ur daddy will be proud that u've learned to fly! think of it as "falling with style".

for that one slightly less of a retard still with me:

it's absolutely absurd turing didn't see this. the dude went on to proudly to use an infinite recursion issue in his paradox construction on the very next page, so why didn't he see that here? heck he even intuited a sense of wrongness about these initial thoughts on the matter, but i guess was so enamored by the halting problem he just didn't dig deep enough into β to recognize the wrongness inadvertently he went down. i can't explain that anymore than the idiots that just jumped off a building over an rdrama suggestion, that they voluntarily read

honestly, i accept the dude being fallible, i forgive u turing. u gave us the theory of modern computing, and that really was a fucking stroke of genius for the ages. this mistep is almost entirely not ur fault.

but what i really can't explain is how the flying fuck this went unnoticed for so long.

what the fuck have mathcels being doing?!

including with his beloved halting problem, and even the motivation for constructing it.

theorycels gunna seething for years over being so stupid. many will take longer to acknowledge it.

how could such established mathematical consensus, with such a "beautifully" simple proof, believed by so many knowledgable "experts", be just so fking r-slurred???

well, i'd love to go into more depth.

but quite frankly, ur too stupid to understand it, so y bother?

Context: The proposed standard would establish test procedures simulating a head-to-hood impact and performance requirements to minimize the risk of head injury. NHTSA estimates the new standard would save 67 lives a year.

If you think this is straggy, you can tell them by submitting a comment within the next 60 days.

There's a few HN threads on this:

large suvs are to protect the occupants. people jumping in front of it should not even factor in. I don't drive on sidewalks.

anyone who drives a large pick-up truck for non-commercial reasons is a grade A butthole

Key Ideas: Heavier vehicles are safer for their occupants but more dangerous for others: The weight of a vehicle is a critical factor in car crashes, with heavier vehicles causing more fatalities in other cars, pedestrians, and cyclists.

Sounds like a heavier car is actually safer to put me and my loved ones inside, at least for now.

Sounds like it's safest for me and my family to physically prevent you from doing that.

https://www.lumenorbit.com/

Data centers... in space (where heat radiation is a legit problem)

Playtron

PlaytronOS transforms any PC, from handhelds to desktops and beyond, into a super games console.

PlaytronOS delivers an open experience allowing you to play your entire game library from Steam, Epic, GOG, and more.

Playtron is backed by consumer electronics, gaming and digital currency leaders like Samsung, Square Enix, Circle & Mysten Labs

We believe Linux is the future of gaming. The Playtron team has deep roots in open source projects from Android's early days to key contributions to modern Linux gaming projects.

They mean CyanogenMod. Cyanogen was some variant of Android (before the come along, I don't care if this is accurate or not) which had a fair amount of community interest. Cyanogen would be succeeded by LineageOS.

Anyway, in 2013 some bloke called Kirk McMaster wanted to commercialise the project. He managed to raise $7 million from morons VCs as seed. This went nowhere, but I imagine it was a comfy gig for a few years.

Guess who's on the CEO of Playtron:

Here are some Loonix articles about Playtron:

Playtron plan to launch PlaytronOs, a Linux based system for gaming

The First Handheld to Use PlaytronOS is some Web3 thing

Not Loonix:

The first confirmed Playtron gaming handheld is aimed squarely at crypto bros

The SuiPlay is to be made by MystenLabs, some crypto 🚀🚀🚀 company, which is an investor in Playtron.

Let's see what Reddit thinks:

I'm all for good competition , but this…

23 hours old.

Linux can't play Linux games, but can play Windows games, HOW

It sounds like an AI wrote it.

Lol !linuxchads Linux can't even run Linux apps properly get a real os

Calling Windows monolithic is fair but never heard of the Valveverse before.

they're calling it that because they're NFT/Crypto/AI bros. It's a buzz word that sounds natural to them

I was thinking the same : This looks like a pic that a crypto bro would use when they slide into your dm's to sell you the newest scamcoin.

Can we just collectively decide to ignore Kirt McMaster's newest investor scam and then have a good laugh when it implodes?

Reminder that PlaytronOS is "built for web3 and the blockchain". Frick that crypto scam BS.

This shit is rich :p Do they just assume that all g*mers are complete morons?

tldr answer, yes.

Square Enix invests into Playtron

7 days old.

Square Enix invests into Playtron.

If you are not already aware Playtron is working on a Linux-based OS for gaming handhelds.

Square Enix a games publisher and developer has invested in the company (Playtron).

Is this a sign Linux Gaming is being taken seriously by gaming studios?

Source: https://www.gamingonlinux.com/2024/09/square-enix-invests-in-playtron-for-their-linux-based-playtronos-first-alpha-out-now/

Fits. Guess Playtron succeeded in what appears to be their goal (grabbing investment money).

I'm still not convinced this isn't some kind of VC investor scam.

Isn't this the company that wanted to implement a load of web3 shite?

Concerns about Playtron OS

69 days old

Playtron is the new Linux based OS that will challenge SteamOS on handhelds

175 days old.

• ⁠Claims Steam Deck is tied to "The Valveverse" when you can install whatever the heck you want on it.

• ⁠Claims an immutable fs (which the Deck also is) will stop cheaters and magically make all games work.

• ⁠AI lookin' mockups that have all the charm and ingenuity of a 60-something year-old who saw sonny boy use a vidya game that one time.

• ⁠Random, nonsensical Silicon Valley talk to woo investors who have no idea what they're investing in. (Why would you put this in a car???)

• ⁠Seems to be pulling from several community projects with zero promise to contribute back.

• ⁠90% of partners are blockchain bros.

lol this is 100% a startup scam. Wish people working there the best, it ain't lasting long.

100% free from Valve!

Powered by Proton.

https://www.linkedin.com/in/nvcbl

I gave Mr Kirk's LinkedIn (the social media for boring corporate sociopaths) a stalk:

I did notice that some of his concurrent startups have his mates at Playtron in high places, which isn't unusual for the VC scene, but kinda doesn't paint the most flattering picture.

IMHO this smells like a company to milk money from VCs (based). I happen to work at a Bullshit-as-a-Service / VC moneypit company and I get first hand exposure to some of the personalities involved. Playtron and Kirk are remarkably reminiscent.

Orange Site discussion (only 11 comments because the article got flagged for referencing the hecking farmerinos)

Now goes by "ladyhavoc". I remember using Darkplaces ages ago for quake 1 as an 8 year old in 2008

Threads

Original: https://boards.4chan.org/g/thread/102272033

Contains standard llm responses.

Discussion thread: https://boards.4chan.org/g/thread/102274366

Example comment: https://boards.4chan.org/g/thread/102274366#p102276574

@Aevann can we do api scrape/pull from 4chan like twitter posts

Standard 4chan shitflinging and shitposting. Found this one interesting because proof is pretty obvious. Can't post images because phone poster + conf call.

>when you extend internet access to 100s of millions across the globe and thr internet hates you for it

Add absolutely every Indian co-worker or person you have ever interacted with

They love LinkedIn and will like and comment on absolutely anything you ever post.

Did some pointless training course or want to shill a project/promotion so people offer you a job? Random skills that need verification points?

Your Indian friends will boost you higher than the CEOs that add disinterested westiods.

How will they interact with sky's prior demographics?

https://www.militaryaerospace.com/computers/article/55137378/lockheed-martin-disinformation-detect-analyzing-media

Installing Doom 2 on your 486 like its 1994 😌 pic.twitter.com/3tMtxB8Ovg

— AlphaFo𝕏 (@Alphafox78) September 7, 2024

https://www.navytimes.com/news/your-navy/2024/09/03/how-navy-chiefs-conspired-to-get-themselves-illegal-warship-wi-fi/

https://arstechnica.com/security/2024/09/sailors-hid-an-unauthorized-starlink-on-the-deck-of-a-us-warship-and-lied-about-it/

The chiefs mess all paid in and they set up repeaters throughout the whole ship lol. Strictly kept it to e-7 and above but no officers.

Turns out even criminal conspiracy and a court martial wont get you demoted from chief.

There was one who was but it was for sleeping with an e3 who worked for him and when they got caught she claimed r*pe

Reddit threads

https://old.reddit.com/r/hardware/comments/1falme3/gn_how_4_people_destroyed_a_250_million_tech/

https://old.reddit.com/r/pcgaming/comments/1falv22/g*mers_nexus_how_4_people_destroyed_a_250_million/

LTTcels begrudgingly post the video and try to find flaws like thumbnails https://old.reddit.com/r/LinusTechTips/comments/1fanga4/babe_wake_up_the_investigation_about_ekwb_dropped/

!codecels !g*mers

>Hi there. Hope you are doing well.

>May I know the domain name?

You at 12:26, Sep 6:

Yes it's bgillettephoto.com

I'm trying to forward it to bradgillette.net, but every time it asks me to verify and sends me a code, it says the code is incorrect, even though I'm copying it verbatim

Now it says I'm locked out from doing that for seven days

Read

>Pranjal dhiman at 12:27, Sep 6:

>Thank you for sharing your concern, I will definitely help you. Please help us to validate the account so we can check from our end and assist you better

>Pranjal dhiman has sent you a Secure Form: PIN and Customer Number

>The following Secure Form has been submitted: PIN and Customer Number

>Thank you so much. We have successfully validated the account. Please allow me a few minutes to check this further into the account for you.

>We've sent a one-time password

>to your email b*********************@G***l.com

>You've requested more than 5 one-time passwords for this action on this domain in a 7-day period. For your protection, you will not be able to request another one-time password for the next 7 days.

You at 12:29, Sep 6:

Every time I've been sent a one-time password it says it's the wrong password

I've entered each password exactly as sent to me yet it says its not correct and now I can't do my job for seven days

That's unacceptable

Read

>Pranjal dhiman at 12:34, Sep 6:

>I apologies for the inconvenience caused to you.

>May I know how many times you try to add the verification PIN.

You at 12:35, Sep 6:

It's never asked me for a verification pin. It has sent me a "one time password" that's six digits long, seven times, and each time it says it's the wrong password

Read

>Pranjal dhiman at 12:37, Sep 6:

>Yes, I am referring to the verification six digit code.

You at 12:39, Sep 6:

There is something wrong with the verification system as it keeps sending me the wrong six digit code. They do not work

Read

>Pranjal dhiman at 12:40, Sep 6:

>No need to worry, let me check with the dedicated team.

>May I know did you clear the cache of the browser once you get the error of wrong password?

You at 12:43, Sep 6:

I haven't but even if I did it now, it will tell me I'm not allowed to do that for seven days

Read

>Pranjal dhiman at 12:44, Sep 6:

>Yes as the security purpose the account is locked for 7 days. You can add forwarding after 7 days.

You at 12:45, Sep 6:

That's unacceptable as it's to no fault of my own.

Read

>Pranjal dhiman at 12:46, Sep 6:

>I can understand as this is for the security purspose so no one can do any changes in the account.

You at 12:46, Sep 6:

EACH TIME I ENTERED THE CODE I WAS SENT, I WAS TOLD IT WAS THE WRONG CODE. I'M NOT R-SLURRED, I CAN COPY A SIX DIGIT CODE. SOMETHING IS WRONG WITH YOUR SYSTEM

Read

>Pranjal dhiman at 12:46, Sep 6:

>No need to worry, I will escalate it from my end so you can not face the issue more.

You at 12:46, Sep 6:

SO SINCE YOUR SYSTEM IS MESSED UP, IN SEVEN DAYS I'LL JUST END UP IN THE SAME LOOP BECAUSE YOUR CODES DO NOT WORK

Read

>Pranjal dhiman at 12:49, Sep 6:

>Please allow me 2-3 minutes, I am escalating the issue to the dedicated team.

>As I can check we need to wait for 24-48 hours after that you can make the changes in the forwarding.

You at 12:53, Sep 6:

That's still unacceptable. I just passed out my card to 100 people and now they're being directed to a website that doesn't exist because of a mess up on your end

Read

>Pranjal dhiman at 12:54, Sep 6:

>I can understand as this is for the security purpose only. We do not have access to make the changes in the lock period.

You at 12:56, Sep 6:

Well you've lost me as a customer. As soon as possible I will be transferring my domain.

Read

>Pranjal dhiman at 12:56, Sep 6:

>As when the wrong password is try multiples time the system temporary locked for the security purpose.

You at 12:57, Sep 6:

But I entered the password I was sent

Read

>Pranjal dhiman at 12:59, Sep 6:

>May I know did you try in the incognito window once?

You at 13:01, Sep 6:

I did not try it in incognito. All I did was enter the passwords I was sent when I was sent them and each time I was told it was the wrong password. I should not have to go through hoops to make a simple change to my domain simply because your system is sending the wrong passwords.

>Pranjal dhiman at 13:02, Sep 6:

>I can understand as the domain is locked for 7 days, we do not have access to unlock it.

You at 13:04, Sep 6:

You already told me you could change the forwarding in 24-48 hours. So either you lied in saying you could do so, or you're lying now in saying it will take seven days.

__________________

After giving them all zeros on their survey, they're texting me via "Daisy" asking me to explain the issue.

No drama, just having a pleasant after dinner stroll past the the biggest brains in galaxy.

Still those AGIs Servers need land

With a super AGI it could design a chip that takes almost no space and almost no energy.

What does money even mean then?

Honestly, I have no idea. I think we need to look to Hollywood for possible answers.