Log inSign up
Log inSign up
songkeys 🐿️🦋@song.work
1,347 posts
songkeys 🐿️🦋@song.work profile banner
@songkeys

songkeys 🐿️🦋@song.work

@songkeys
i love (formerly hate) 𝕏 (formerly twitter), an ad platform (formerly social platform) to spam (formerly tweet). 🌍 普通话/中文/华语/国语/汉语/broken engslih
songkeys.eth
bento.me/song
Joined July 2014
1,519 Following
2,359 Followers
RepliesRepliesRepostsRepostsMediaMedia

Log in or sign up for X

See what’s happening and join the conversation

Continue with phone
or
Log in with username or email
Terms·Privacy·Cookies·Accessibility·US TIDA·Ads Info·© 2026 X Corp.
  • @songkeys
    songkeys 🐿️🦋@song.work
    @songkeys
    16h
    Haiku 5.5 强过 GPT 6 Luna Sonnet 5.5 强过 GPT 6 Sol Opus 5.5 强过 GPT 6 Astra Grok bot 强过 dots 每天在和土豆姐你来我往的 Tibo:我们在侧边栏加个什么功能好呢 回来吧我的openai,我最骄傲的信仰😭
    @claudeai
    Claude
    Anthropic
    @claudeai
    18h
    Introducing Claude Haiku 5.5: the cheapest, fastest, and most capable small model we’ve ever released. On average, it costs around 75% less to run than Claude Haiku 4.5.
    00:00
    9
  • @songkeys
    songkeys 🐿️🦋@song.work
    @songkeys
    Oct 6
    通过对多家 AI 公司的订阅计划进行实际限额测试,得出结论:在 API 等价价值上,Anthropic 的订阅约为 OpenAI 的 5 倍以上。🤯
    @SemiAnalysis_
    SemiAnalysis
    @SemiAnalysis_
    Oct 5
    Anthropic Subscriptions Offer 5x+ More Value Than OpenAI Limit testing every AI subscription plan from Anthropic, OpenAI, Meta, SpaceXAI, MiniMax, Moonshot, Zdotai, Cursor, and Cognition newsletter.semianalysis.com/p/anthropic-su…
    2
  • @songkeys
    songkeys 🐿️🦋@song.work
    @songkeys
    Oct 6
    别的国家不清楚,中国一半以上的算力都消耗在华强买瓜上面了。
  • @songkeys
    songkeys 🐿️🦋@song.work
    @songkeys
    Oct 4
    腾讯给了 @OpenAI 多少钱?codex 拒绝通过 computer use / browser use 访问微信的任何页面,包括微信开发者平台 / 微信文档等。
    53
  • @songkeys
    songkeys 🐿️🦋@song.work
    @songkeys
    Oct 3
    仔细想想,GPT 这种 只顾结果、不顾过程 的工作方式也未尝不可。反正99%公司的老板也只看你的最终表现,把活抓紧干完就好了。也只有同事会喷你代码。 在老板/资本看来,程序是实现商业的手段,不是目的。我认识的老板口头禅现在都是: 越快越好 / one shot 一下 / 先上线,以后再说
    @Gegam245074
    Gegam
    @Gegam245074
    Oct 2
    If you think GPT-6 Astra and GPT-6.1 Sol are good models, just look at this. And the problem isn't that the code is ugly. It's much more fundamental. It points to serious problems with how these models are trained and reinforced: 1. The model was optimized for measurable
    00:00
    00:00
    2