starry

A student

Fond of AI and stars

[starry @ODPQUsLP5V69068]

codex的上下文被改成80万了？说他两句就开始玻璃心了哈哈

[Tibo @thsottiaux]

@rohit3a Folks. Why do you think we picked 300k for context length for Codex.

We ran the numbers and it’s better. We can run it at 1M too but it would just cost more of the usage. Good defaults are important.

[starry @ODPQUsLP5V69068]

🤣

[Yuchen Jin @Yuchenj_UW]

The guy who built this digital whip for Claude Code is definitely getting a lifetime Claude ban for “abusive or cruel behavior.”

[starry @ODPQUsLP5V69068]

@CodeByNZ Yes,because Claude is the god Anthropic worship for

[starry @ODPQUsLP5V69068]

@ScarletKc 我觉得opus5.5更像是agi

[starry @ODPQUsLP5V69068]

@ScarletKc 以后让AI直接写机器码会不会更快一点

[starry @ODPQUsLP5V69068]

AI is >10% likely to kill all humans？

[Mikita Balesni @balesni]

i am at OpenAI and i think AI is >10% likely to kill all humans

this proposal is among the top things we should do as an industry to lower that risk (it’s not enough though!)

[starry @ODPQUsLP5V69068]

理解，这是他们的神明

[Hayden Field @haydenfield]

Exclusive: Anthropic is updating its usage policy for the first time in over a year. The new rules prohibit sustained "abusive or cruel behavior" towards Claude & add new restrictions about propaganda campaigns, surveillance & weapon development. https://t.co/GLsBeSR0fN

[starry @ODPQUsLP5V69068]

@kamilbaranek_ smarter and cheaper

[starry @ODPQUsLP5V69068]

@ScarletKc 嗯嗯现在说话感觉精简很多

[starry @ODPQUsLP5V69068]

@ScarletKc 开启目标模式就不会停下来了

[starry @ODPQUsLP5V69068]

@thsottiaux Just shipping 2.1 is ok enough, these useless small things will make people feel you are perfunctory instead.

[starry @ODPQUsLP5V69068]

I don’t believe Tibo can really either ship meaningful improvements or giving resets in the continuous 28 days.

Once he doesn’t achieve what he promised, he will really destroy users trust as he hyped
so much before.

He is playing fire.

[starry @ODPQUsLP5V69068]

@thsottiaux Stop shipping these useless things,we want 26 resets in the next 26 days.

[starry @ODPQUsLP5V69068]

@EatMyTarts17 @kimmonismus @notquantized @thsottiaux then you are so selfish

[starry @ODPQUsLP5V69068]

I smell love

[Tibo @thsottiaux]

@poteto @bot I smell fear

[starry @ODPQUsLP5V69068]

要是OpenAI再不抓紧发布一个新的前沿模型能够超过Opus5.5的话真的要完了。

这个公司一直在加速透支用户的信任，从模型降智降速到削额度，再到Dev day的各种hype。

从使用体验上来讲，Opus5.5是完爆Sol6.1甚至是Astra。claude的20美元订阅都有很宽裕的额度，再反观astra，plus用户基本上两三轮对话就达到限额。ChatGPT的模型用起来就是比Claude要差很多，很多时候用O的模型，明明很明显的地方，但你必须要说出来，你不说他就不做，Claude用起来就是你不用和他说太多，他会自己把很多因素考虑进去。

还有一点，看起来Tibo经常会去重置很大方的样子，但他们的重置会刷新重置时间，这对那些想要保持自己节奏的人反而是副作用；Cluade的重置虽然给的少，但是不刷新时间，也不会在你用的时候开始计时周额和五小时额，然后当5小时额度刷新的时候他会自动进行对话，用起来都很舒服。

O的定价现在要比A要更高了，但他最好的旗舰模型却比不上别人的中档模型，更何况Fable5.5马上要来了，那凭什么用户再继续订阅付费O，就凭他一直推出的那些花里胡哨又没用的功能吗。

[starry @ODPQUsLP5V69068]

@ScarletKc 是的，可能他们自己也意识到模型追不上Ant，想靠其他补平，但用户不买单…