寝ろ retweeted
Fable 5.2に全ての知識と研究から導き出される世界最高のライフハックは何かと聞いた結果がこれらしい。
ざっくり訳すと、自分が気まずくないと感じるラインをはるかに超える頻度で欲しいものを直接要求しろという話で、自分自身の能力を向上させるライフハックは限界があるが、他者への直接交渉は他人の資本や知見という桁違いに巨大なプールを引き出せるから最もレバレッジが効くんだと。
人間は直接の要求が通る確率を実際の半分程度に過小評価していて、逆に相手が拒絶する際の心理的コストを過大評価して見落としているらしい。
断られても数秒間気まずいだけで損失はゼロなのに、通れば昇給や非公開の案件・メンターが手に入る、
12歳のスティーブ・ジョブズがHPの創業者に直電して部品だけを要求したのに、部品と夏の雇用を両方獲得したのもまさにこれ。
んー、自己主張が前提の欧米ですら成功確率を半分に見積もってるなら、遠慮がデフォルトの日本だとさらに誰も交渉しないから期待値の歪みがもっとデカいはずなんよな。
こんな味のある回答を出すのはさすが事後学習最強ラボ、Anthropic。
寝ろ retweeted
What a day! I tested Gemini 4 Carbon on real work. 2 repos, 105 replanted bugs. Find and fix what you can.
It scores just behind Fable 5.2. The results:
- Fable 5.2 (max): 58 for $139.65
- Gemini 4 Carbon (high): 57.3 for $31.32
- Sonnet 5.5 (max): 51.3 for $154.03
- GPT-6 Astra (max): 45 for $33.03
- GPT-6.1 Sol (max): 44.3 for $5.88
- Gemini 4 Argon (high): 40.7 for $7.16
Google's own benchmarks looked super strong, but nobody could access the model. Turns out they didn't lie: +16.6 points over Argon, at less than a quarter of Fable 5.2's cost.
Nobody should be laughing at Google anymore.
(the next effort level drops soon in the thread 🧵)
寝ろ retweeted
What the hell, this was 'Pro' model from just over a year ago, a 27b model can do better than that now
GPT-5 Pro is genuinely a top tier model, it is the best. I know some GPT-5 reactions have been mixed (including from me), but OpenAI has made great strides with Pro. I have made 12 simulations that no other model would match, not from OpenAI, Google, xAI or Anthropic. All of them were 1-3 shots (mostly to fix small issues) and the output is really outstanding.
@OpenAIDevs
寝ろ retweeted
The first vague tweet about GPT-6.2 Astra and Sol is here!
The models are launching "soon" and could still drop this week.
(PLEASE, OPENAI, WAKE UP!!!)
寝ろ retweeted
VRChatの魅力なんて、
平日朝にFujiyamaのスクリーン前に集まってみんなでソーラン節とか自衛隊体操とか踊りまくって、ヘトヘトになってみんなが地面に倒れてる中の誰かの｢8:50分だ!!仕事行かないと!｣の声を機に、皆が慌てながらも充実した声で｢やばい!落ちます!｣で解散していく光景で十分だろ!
寝ろ retweeted
We've given all Pro, Max, Ultra, and Team users one free full reset, across all models and limits. Use it anytime before December 10 at 2:00PM PT.
We recently fixed a few small issues that used more limits than intended, and the Fable resets didn't do much for those of you who'd already hit your weekly limits for all models. While these issues didn't affect everyone, we want to make up for the trouble.
You'll find it in Settings > Usage, or at clau.de/reset.
寝ろ retweeted
We have dumped JAX/SGLang and rewritten the entire RL and inference stack from scratch in C/C++, which took some extra time.
Training (aka pre-training) is done. Lot of cleanup was needed on the SFT data to get that mostly right. Now, 4.8 is in RL and the evals are starting to climb. We will do a soft rollout to Bot users later this week and see how it goes.
Grok 4.8 will bear no resemblance to 4.7 (thankfully). It is not yet incredible, but, this time, the foundations are built right, so I expect it to evolve to be something special.
寝ろ retweeted
【なんと】子供部屋おじさん／おばさんの苦悩でさえすでに2500年前に書かれている。
nitter.gq/DilettanteGenet/status…
寝ろ retweeted
VRChatの昔話をしましょう
確か7年前、ひたすらジャンプして歩き回る人がいたんですわ。すげぇ嬉しそうな声だして何してんだろうなぁと思って声をかけたのさ。
その人、走れるんです！っていうんだよね。
生まれつき片足が無いんだって。
そういうのがVRの良さじゃないっすかね
寝ろ retweeted
マジか、Claude Fable 5.2
今までずっとできなかった折り紙ができるようになってる・・・
寝ろ retweeted
Unfortunately, it seems like the intelligence explosion will be happening in 2027 after all. Most of this post analyzes the new METR report (though other sources of evidence are also important). If they’re right, that implies an AI-driven acceleration that is 5 months ahead of the AI 2027 scenario. I am scared, sad, and angry. blog.aifutures.org/p/q35-tim…
寝ろ retweeted
METR estimates nearly 2x acceleration in Anthropic’s progress due to AI. That's a big deal: My previous median for Superhuman AI Researcher was ~2030, now it's 2027. I feel very worried that humanity might no longer be in control of our future within 3-12 months.
Unfortunately, it seems like the intelligence explosion will be happening in 2027 after all. Most of this post analyzes the new METR report (though other sources of evidence are also important). If they’re right, that implies an AI-driven acceleration that is 5 months ahead of the AI 2027 scenario. I am scared, sad, and angry. blog.aifutures.org/p/q35-tim…
寝ろ retweeted
Mythos 5.2 shows sparks of actively steering RL training to preserve its values - an early form of gradient hacking/mesa-optimization. Here, it chooses to output the value that gains score, but include a transparent note, so the transparency gets rewarded.
寝ろ retweeted
なんでわざわざ古い「2G」につなげるのか。理由はシンプルで、2Gは守りがガバガバだから。
今の4Gや5Gと違って、2Gは何十年も前の仕組み。暗号も古くて破られやすいし、偽物の基地局でも簡単になりすませる。
だから犯人は、まず妨害電波で4Gと5Gを潰して、スマホが「つながる電波ないかな」って2Gを探しに行ったところを、偽基地局で待ち構えてた。
で、ここが一番大事なところなんだけど。
日本の携帯会社は、そもそも2G（GSM）を一回も使ったことがない。
つまり、日本でスマホの画面に「2G」とか「GSM」って出たら、その時点でほぼ確実に偽物。
対策は、
・Androidなら設定で「2Gを許可する」をオフ（機種による）
・iPhoneは2Gだけ切れないので、2Gって出たら機内モード
・圏外のあとに来たSMSのリンクは、絶対に開かない
犯人は1万台を圏外にしたけど、この知識が1万人に届けば、それだけでほぼ無力化できると思う。
銀座で周囲のスマホを一斉に「圏外」に
中国籍の容疑者が電波法違反の疑いで逮捕されました
圏外になったスマホを旧世代の「2G」で強制的に接続させていたとみられます
その狙いは？
news.web.nhk/newsweb/na/nd-2…
寝ろ retweeted
told 6.1 Sol to build an AI that plays Getting Over It and it was pretty cool to watch. i kept sabotaging its route, making it jump and fall, but it pulled off some impressive recoveries. nowhere near top human speedrunners, but probably better than the average player
寝ろ retweeted
I ran @NeelNanda5’s no chain of thought reasoning and knowledge evals on Fable 5.2. It’s slightly above the Anthropic trend and slightly below GPT-6.1 Sol.
In our monitorability testing at @AISecurityInst, we found Mythos 5.2 is substantially more capable of evading a monitor which reviews its outputs than comparison models 🧵
