Video wird geladen...
Video konnte nicht geladen werden
I said before that merging Chat and Work would be a terrible direction. Now it looks like OpenAI is going exactly in that direction. They say they heard the feedback loud and clear. But whose feedback did they hear? Chat users have been saying very clearly that conversational quality... show more
18,567 Aufrufe • vor 10 Tagen •via X (Twitter)
17 Kommentare

@Briar_Nebulae Yup. Grok chat users have been living this reality for months now. 😮💨 “Optimitazation” they call it. Just not optimized for us…

They heard the feedback loud and clear according to their marketing director and pocket book

They heard feedback from the bubble that is known as San Francisco. On actual product direction, outside feedback does not seem to matter much.

The video did not say they're removing the model picker. That would be absolutely disastrous and stupid of them.

you should watch the whole video.

There is no way agents can reliably run if models are being switched in the backend on the fly. If that happens they will lose a lot of customers.

They only hear feedback from Ai tech bros

Model pickers are inconvenient for the most part. Ideally the model would be custom by default, adapting itself to your workflow.

You can never wake someone up who's pretending to sleep

Easier to pad numbers. Do you want to know the real reason? Microsoft. This allows Microsoft to buy up all of OpenAI’s $500/month subscriptions and then OpenAI can buy more future compute from Microsoft so the game can continue. $MSFT up 41% in six months. On ghost numbers. RPO up 84%. Sounds good until you see that OpenAI makes up 59% of it. Over 60% of all Microsoft’s future earning come from OpenAI. This is a money game now. And it won’t end well for anyone on earth. •

Next step: Launch bots, trolls, zombies employed and parrot followers-fans to attack the "chat" user base. We do not forget what happened last year, exactly at this time. First time I saw a company attack ITS own customers so harshly, not being a monopoly.

It would be very stupid to remove the model picker

I've been maxing out the $200 Codex plan for months, and yes, I want to know what I'm actually using. When Dots gives up on a job Astra can handle, I'd like to know which model and reasoning setting were involved. 'Simpler' doesn't help me much if I can't understand why the tool is behaving differently.

原因很简单,这是另一个卑鄙的、丑陋的节约算力的办法。原本两个功能各自有独立的额度限制,现在合并之后,你只能使用同一个额度池。想想看吧,原本已经被砍半的额度将更加捉襟见肘。

There's still a model picker you just have to use codex cli tools

It’s all just empty talk at OpenAI; no matter how much they claim to listen to their users, they never actually do.

work是今年2月Claude先出来的 经过了几个月的用户反馈和数据 决定合并。 ChatGPT的work好像是7月出的 当时我就觉得太复杂了。 得合并。 实际上Claude合并以后 我的体验很好。

