窥基
694 posts


ok! quick updates from me:
- i left anthropic (who does that?!)
- had a baby (i love her)
- and am joining @cursor_ai as head of design (eep!)
it's been a low-key dream of mine to nurture a team that cares so deeply about craft, quality, and building great tools. very excited!
English

我很困惑,您或者其他人没有注意到,ChatGPT Plus 和 Go 的这些订阅中并不提供最前沿的模型,以及即使提供前沿模型,它也并不像 Pro 那样提供最佳的努力程度。
而一个关键的问题是,在通用对话中,理论上回复并没有质量上限。它不像编程或者智能代理,只要把工作完成就可以了,所以回复质量完全取决于模型愿意付出的努力程度。
现在 Plus 订阅中的模型思考通常只有十几秒。与之相对照的是,Claude 的 Pro 订阅就提供了最强的模型,尽管使用量极为有限,但是在通用对话中,遇到复杂的问题它有可能思考十几分钟。而相同的问题,ChatGPT Plus 的模型却只思考十几秒,这是非常令人遗憾的事情。
希望能够给 Plus 用户试用 Pro 模型的机会。
中文

Do you agree that ChatGPT should offer its most powerful model in the $20 subscription tier, even if it's a very limited allowance, similar to what Claude does, for fairness and equal access to the model?
@thsottiaux
English

@morumbizero @thsottiaux 但是您是否注意到,Claude Pro 就提供了最强的模型,而 ChatGPT Plus 却没有提供?如果您是因为预算有限而从 Claude 转向 ChatGPT,那反而无法使用最强的模型啊。
中文

@thsottiaux Usei Claude Fable 5 por 10 minutos e já cheguei no limite. Infelizmente sou do Brasil e a minha moeda é extremamente desvalorizada. Impossível continuar neste modelo.
O Opus já era difícil.
Detalhe: $1 = R$5 da moeda brasileira.
Português

I think GPT 5.6 is pretty good
Claude@claudeai
We're extending Claude Fable 5 access on all paid plans, as well as keeping Claude Code’s weekly rate limits 50% higher, through July 19.
English

Claude 延长了 Claude Fable 5 的访问到 7/19!
从来没见过这么傻逼的公司,跟儿戏一样!
Claude@claudeai
We're extending Claude Fable 5 access on all paid plans, as well as keeping Claude Code’s weekly rate limits 50% higher, through July 19.
中文

Stop joking. Google isn't focused on providing the most advanced, high-performance models anymore; instead, they're just trying to provide the most widely available ones.
If you look at the current version, it's performing significantly worse than it used to. It's a classic case of laziness: the model just doesn't want to do the work anymore.
English

Gemini 3.5 Pro benchmark leak just dropped and the numbers are turning heads.
> Reportedly outperforming Claude Fable 5 and GPT-5.6 in internal evals
> Significant zero-shot performance improvements over 3.1 Pro
> Currently in private validation and testing
- Public rollout preparations appear to be underway
- July 17 is the targeted launch date
If the leaked numbers hold, this isn't Google catching up ,it's taking the lead.
Fable 5 just became the model to beat. Gemini 3.5 Pro might already have beaten it.

English

@typelessdotcom @furyfrog_hu 不稳定,经常输入的预期效果都不太一样,导致使用这个输入法的摩擦感太强了。
所以,优化的方向并不是一味地提速,而是保持稳定。哪怕每次思考的时间比现在稍微慢一点,却能够始终提供符合质量、符合预期的输出。
而这种风格,恰恰像 Typeless 最初的那样。
中文

@furyfrog_hu Please email us! :)
📩hello@typeless.com
Thanks again for your patience and support.💙
English

Typeless 最近经常要重试几次才能成功,一直会处于“连接比平常慢”的状态。
本来尝试去换了一下豆包输入法,但豆包输入法在体验上对比起 Typeless 还是有一些差距。比如:
1. 在开启语音输入的时候,Typeless 会自动屏蔽掉背景音(比如 Mac 播放器里播放的声音),让人能够更专注地输入。
2. 对于说话内容的总结和“去废话”处理,也是Typeless好的多
Typeless 算是我为数不多年付费的 AI 产品,但最近的体验实在是不太好,能不能优化一下呢? @huang_song_

中文

@babayagatwt @CaptainInsightX 你是不是没有用过 Gemini?
Gemini 不是落后于其他模型,而是落后于过去的自己。现在它的模型非常的懒惰,这说明什么呢?说明谷歌想提供普惠型的模型,而不是提供最佳模型。
它已经放弃了战斗。
中文

@CaptainInsightX Am I missing something? why are they working on their own model?
English

@thsottiaux ChatGPT should follow Claude's lead and offer a trial of its most powerful model within the $20 subscription tier.
English

小县城婆罗门们作威作福的好日子,彻底到头了。
原因极其残酷:年轻人跑光了,没人再留下来求他们了。
曾经挤破头才进得去的县一中,如今生源流失;县医院被彻底边缘化,老百姓生病一脚油门直奔省城。那些靠地方权力倒腾的基建、绿化工程,现在做完了也根本要不到钱。最让这群特权阶层绝望的是,他们引以为傲的“人情关系网”彻底瘫痪了——年轻人转头去了南方打工,你盘根错节的本地土权力,再也卡不住任何人的脖子。
没有了源源不断供其收割、买房、托关系的年轻一代,县城的资产、编制和资源瞬间失去了全部溢价。曾经在熟人社会里高人一等的县城精英,最终只能守着无人接盘的房子和厂子,看着那点可怜的优越感,在人口抽干的废墟中一同贬值。
#认知觉醒 #县城现状 #社会观察 #看清本质 #时代变局
中文












