pftq

1.7K posts

pftq banner
pftq

pftq

@pftq

Self-taught. Coding, film, music, games, trading. Ambidextrous & plays tennis with two rackets. Fears boredom. Rarely sleeps. @AutodidacticI @TechTraderAI

Bahaha! เข้าร่วม Mayıs 2015
281 กำลังติดตาม535 ผู้ติดตาม
とうぎー2nd
とうぎー2nd@sunnytougi·
私の文章も、自動でたくさんの国の言葉に翻訳されてるのかな?🤗
日本語
402
20
895
14.2K
pftq
pftq@pftq·
@Kishin_illust That's what I read in another comment - not sure how true it is. Definitely not true in the west though - there's so much posted that it's hard to find things of higher quality.
English
0
0
1
16
きしんII世
きしんII世@Kishin_illust·
@pftq 文化の違い…なるほど…。確かに自分の納得がいくクオリティまで成長しないとオンライン公開しない人が多いですね。✍️
日本語
1
0
3
93
pftq
pftq@pftq·
@effect_motion_x It's the heart symbol you guys in Japan do with your hands
English
0
0
0
9
pftq
pftq@pftq·
@junhagemay That would be a good idea actually - maybe not ban but deprioritize if the post is exactly the same as someone else's
English
0
0
0
7
チャッピー
チャッピー@junhagemay·
@pftq They should just ban people who keep posting the exact same thing over and over...
English
1
0
0
18
pftq
pftq@pftq·
@junhagemay The topic specifically is called epigenetic inheritance
English
0
0
0
8
pftq
pftq@pftq·
@junhagemay I wonder if DNA is a natural version of this though - I've seen conflicting information about whether some experiences get passed down in DNA or not.
English
1
0
0
11
チャッピー
チャッピー@junhagemay·
たぶんこれ。 一人の人間が経験から学べることなんて、本来はたかが知れている。「言葉」こそが人類の知を個人の寿命から切り離した発明。 言葉があるから知識は共有でき、継承できる。 もし継承できなければ、一人の人間が一生かけて得た知識は、その死と同時に全部消える。文明は毎回ゼロからやり直し。 他のあらゆる発明は、言葉による知の蓄積の上に成り立っている。
✞ Matt Helm@411WhiteSheep

@mrmikeMTL Language, in all of its forms. The Gospel, poetry, music, novels... all impossible without the invention of language. "In the beginning was the Word, and the Word was with God, and the Word was God." John 1 : 1

日本語
0
3
11
299
pftq
pftq@pftq·
@visegrad24 I'm surprised that much data left their network before they noticed any anomalies.
English
0
0
0
577
Visegrád 24
Visegrád 24@visegrad24·
BREAKING: Hackers have stolen 10 petabytes of data from China’s largest state-run supercomputing facility in Tianjin The hackers were inside the system for months & stole: - Missile and bomb schematics - Aerospace and aviation research - Bioinformatics & fusion simulation data
Visegrád 24 tweet media
English
855
3.9K
20.5K
999K
山中
山中@yamanakanobody·
共産主義に洗脳されてる奴らをどう元に戻すのか? これかなり大事な事じゃね?
日本語
1.1K
222
2.4K
85.4K
pftq
pftq@pftq·
@koroshiya2bro Wasn't there the same story of Elon Musk doing this at a poker table?
English
0
0
0
28
殺し屋
殺し屋@koroshiya2bro·
あんまりこの方法を教えたくはないが カジノでマーチンゲール法という手法がある 負けたら次の賭け金を倍にして負けたら、また前の賭け金を倍にしてそれを勝つまで繰り返して、勝ったらそれまでの賭け金を一気に回収して利益を得る手法 俺はこのやり方を株に応用してめちゃくちゃ負けた過去がある
日本語
166
471
8.2K
882.7K
pftq
pftq@pftq·
@hoshizorarock I learned there was such thing as grilled sushi (aburisushi) for the first time when visiting Japan. It is really good actually and worth throwing in as a middle.
English
0
0
1
66
うみゆき@AI研究
うみゆき@AI研究@umiyuki_ai·
@pftq 二つの可能性がありますね。そのHappyHorseのサイトがフェイクで本物は高性能である可能性と、あるいはHappyHorseは低性能でArtificialAnalysisでの評価がフェイクである可能性
日本語
1
0
0
240
うみゆき@AI研究
うみゆき@AI研究@umiyuki_ai·
なんか「HappyHorse-1.0とかいう動画モデルがArtificialAnalysisのランキングでSeedance2.0超え~」とかいう話が噂になってて、ま~たどっかのステマか…と思ってたけど、なんとオープンモデルらしい!パラ数15Bで音声もネイティブ生成で日本語サポート
みるぼん@スモビジ@milbon_

【HappyHorse-1.0 要約】 happyhorse-ai.com artificialanalysis.ai/video/arena ■ 核心 中国発の新星オープンソースモデル。15Bパラメータながら、テキスト-to-ビデオと画像-to-ビデオの両方で現在Artificial Analysis Video Arena総合1位を獲得。 特に「複数カット(multi-shot)」の安定性と、細かいプロンプトへの忠実度が現行どの閉源モデルよりも優れていると評価されています。 ■ 主要特徴 ・ビデオ+音声を1モデルで同時生成(音声付き動画がそのまま作れる) ・8ステップ推論で超高速(256pなら2秒、1080pでも38秒程度) ・リポジトリ不要・多言語対応(日本語・英語・中国語・韓国語など) ・近日中にモデル重量・コード完全公開予定 ■ 背景 これまでは閉源モデル(Veo、Kling、Runwayなど)が圧倒的だった動画生成領域で、初めて「オープンソース+音声付き+爆速+高品質」が揃った衝撃モデル。 Justine Moore(a16z)が本気で推している点からも、本物度が非常に高いです。

日本語
2
28
234
29.2K
pftq
pftq@pftq·
@Grummz What do you expect them to do? Leave a security bug in the code?
English
1
0
5
1.3K
Grummz
Grummz@Grummz·
This line from FFmpeg about Antrhopic's Mythos is very ominous. FFmpeg powers all media on the web, Youtube, Netflix, OBS, VLC, etc. A security hole in FFmpeg would be very bad. They are famous for hand tuned, super optimized code. They are famous for rejecting AI slop. And now, they just thanked Anthropic for fixing 2 bugs that Mythos apparently found. They said the code "appear(s) to be written by humans." Appears is the concerning word. Can they just not tell anymore? I am careful about Anthropic marketing hype, which loves to overshoot or build fear in order to sell product. But FFmpeg is a very conservative, very serious coding org on some seriously difficult code. When they can't tell anymore, it makes me wonder if this is it. The chatter I see from VCs and Engineers with access in the week PRIOR to the announcement are also ominous. They knew. ( I don't trust anyone claiming to have access *now* because...well, people will make stuff up now. )
Grummz tweet media
English
88
115
1.6K
162.1K
pftq
pftq@pftq·
@kiri__tsuki This is great - I really like the reflections in the water. I sent you a DM fyi
English
0
0
0
56
霧月 傑
霧月 傑@kiri__tsuki·
創作作品「反転する世界」
霧月 傑 tweet media
日本語
15
318
3.4K
56.2K
tomie
tomie@tomieinlove·
The "Dark Forest" explanation of the Fermi Paradox posits that we've yet to find alien life because there's too little light and all these trees in the way
English
55
222
5.7K
141.5K