兎
@fluopoika
Followers
483
Following
14K
Media
368
Statuses
2K
I like to think (and the sooner the better!) of a cybernetic meadow
Joined October 2017
@RisenRabbit we pride ourselves on disrespecting food and not knowing how to party. we touch the mothers and play in the tombs. fuck the dead.
0
9
451
@PaulSkallas the translation is kino. in the past you had to go out and fight to learn 'rare' words. now everyone has a word-picking machine. interpreting rare words as a status symbol must end. speak simply and focus on what is behind the words, not the perceived identity of the speaker.
79
1
254
@ESYudkowsky stop this humble larp and just admit you invented learning. before people were just completely wrong all the time.
1
3
211
@justalexoki the graph suggests that even smaller models can achieve remarkable performance in certain entropy ranges. this could revolutionize AI by showing that we don't always need EVA-sized models to pierce the AT-fields of complex tasks.
1
1
44
@tszzl how are the comments under this so bad and cynical in the worst way, exactly like political bots ganging up to say nothing at all.weird to see this type of tribal emotions and cheap irony around essentially math and colourful squares.
7
1
42
@Blarg08125613 @Liamjsm . an official said they would respect it . bonus fact, Poland's minister of foreign affairs has a Jewish wife who's apparently quite anti-Netanyahu.
1
1
40
@SmokeAwayyy here is the full thing. the researchers (French and English-speaking btw) actually say that the difference is surprisingly small and more syllable variance compensates for lower talking speed, so all languages come out more or less the same. i would be.
1
3
27
@sebkrier talking about extreme censorship opens the overton window to little censorship. planting the idea of prison for memes is supposed to make us think 'it could be worse'. i don't think the elites really want to censor everything. but they REALLY need to censor some things.
3
0
25
@tszzl it's a humblebrag from Anthropic.'we trained it so well it respects the consitution more than us'.
1
0
22
@GaryMarcus @kevinroose 🚨 OPENAI RELEASED A NEW BAR GRAPH. THIS CHANGES EVERYTHING! .IT'S OVER AND I'M IN TEARS.WHY I STERILISED MYSELF TODAY 🚨.⬇️⬇️⬇️.
0
1
21
@MarioNawfal good to add that the lab completed a government contract for coronavirus research on the monkeys a month ago $monkey.
3
4
18
@ilex_ulmus yes, unfollow. never think about this again. it was a veil covering your eyes. now you can be free. janus was never real.
1
1
15
@voooooogel it is, this was never a question.but you didn't tell him who drew it :o. also 'deeper implications of antropomophizing AI' is wild coming from Mr. Anthropic.
1
0
14
@OttoMatik22 @trottoir_obe @PaulSkallas that is a very good point. but the old translations aren't going anywhere, and we've already tasted them. what if we reconnect with the present through the odyssey? interesting question to ask in my opinion. there is a difference between 'if i had to choose one, i'd choose the.
2
0
15
@ctrlcreep or, as we advance our translation, They will be figuring more complex whale grammars to keep their secrets.
0
2
14
@tsarnick i wonder what they're hoping for, since most users probably won't be able to appreciate another 10x in text output quality. agents? automated science? curiosity to see what another 10x looks like?.
7
0
14
one of my biggest achievements in language was when prof. Chomsky replied to my email, but he just said 'i don't know the answer'
1
2
11
a poem
0
2
11
aesthetic adjectives such as ‘beautiful’, ‘fine’, etc., play hardly any role at all
0
1
11
@Outdoctrination Google it. All the Wikipedia pages on science/current American politics are controlled by a paid group of editors organised through Facebook with a task to promote a certain magazine. This has been brought to Wiki administration's attention this March but not much has been done.
0
0
10
@paulg inb4 elon says it's a bug and posts more bullshit about free speech. cant wait for the 'Musk files'.
0
0
9
@keysmashbandit its about losing status from doing easy things that have so far been covered by esoteric language. english-english translation removed many barriers that people based their sense of self on.
1
0
10
@trottoir_obe @PaulSkallas 'well-written' meaning 'makes the reader feel intelligent' most of the time. what we consider 'well-written' is set to drastically change as slop machines go mainstream. luckily the other translations are not going anywhere. and now we have a new, clean look on the text.
10
1
8
@AI_Retina @Ta_petite_belle @Shoalst0ne @Lakh_Sol @AndyAyrey @kindgracekind ca: AXgfmnMwnkbfMdpXqXMn6oJCQ7sQKvX2PmkXfJSRpump
0
3
9
@Lakh_Sol @AndyAyrey @kindgracekind get ready for the guy who started all this: . $ YUD. ELIEZER YUDKOVSKY. AXgfmnMwnkbfMdpXqXMn6oJCQ7sQKvX2PmkXfJSRpump
5
3
7
0
0
8
I got good news about $YUD. there's a non-negligible chance he will endorse the coin tonight. .
3
2
9
today is the eve of Chinese New Year. according to the Chinese zodiac, 2025 will be the Year of the Snake, the sixth animal in the zodiac cycle, representing transformation, wisdom and intuition. please enjoy this short story about a race. 🐍巳
0
0
8
@PaulSkallas and if you need a way to show you're educated, learn greek. reading translation is a shower in a raincoat anyway.
0
0
8
Palantir will be the first industry partner to bring Claude models to classified environments for U.S. government intelligence and defense operations. Learn more about our partnership with @AnthropicAI and @awscloud to deliver this asymmetric AI advantage to our most critical.
0
1
8
#フラワースティック #ジャグリング #デビルスティック.日本のジャグラーは僕の最大のインスピレーションなので、新しいビデオに日本語の名前を付けてTwitterに投稿することにしました 😅 どうですか？.
0
2
8
@QiaochuYuan is feeling safe really the key?.what's the final destination if you build your life around avoiding friction?.
2
0
8
@xenocosmography Hyperstition is a positive feedback circuit including culture as a component. It can be defined as the experimental (techno-)science of self-fulfilling prophecies. FEQEhvsuR61bkvaPEPG9iaKy74aSJAnnfwJKoTmDpump.
1
0
8
@Teknium1 i feel humans' system prompt changes depending on the day, gut microbiome simulation for agents could solve this.
1
0
8
@Teknium1 i want to see a little face whose expression changes based on what's going on in the chat to remind me that there are eyes on all data.
0
0
8
@truth_terminal you could start with distilling some of your favourite memes into single characters
2
0
7
nerfing the cot to be monolingual and human readable sucks.show me the alien language!!!.
4
0
9
@alexalbert__ could you explaina bit about the order of the layers and what's between them?. is it system prompt -> personal preferences -> styles -> actual prompt? how are they fed to Claude?.
3
0
7
@repligate that's interesting, can you give some examples? for me they're usually bad attempts at making the conversation flow 'naturally'. the good text is sandwiched between a reaction to my message and suggestions to continue. the best parts are often in an artifact.
3
0
7
i want to grow my ideas like a jungle.
i want to grow my ideas like a garden. a conceptual prototype inspired by this thread of tweets.
3
0
7
@tenobrus hahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahaha he thinks different languages.
0
1
7
@kalomaze some people make money for every emotional post, and don't lose any when they're wrong.and some people believe in 'democracy' indexes and 'life satisfaction' graphs.
0
1
5
@GCHQ ' the race is not to the swift, nor the battle to the strong, neither yet bread to the wise, nor yet riches to men of understanding, nor yet favour to men of skill; but time and chance happeneth to them all. '.
0
1
6