@OpenAI - your @ChatGPT is not reliable with its support in general.
Also it can't report its onwn shortcomings to home.
#wasteoftime
in hindsight, this feature was poorly timed with dot, since dot does not currently use intelligent UI
everyone who switched to primarily using dot has completely missed out on the intelligent UI
1. GPT-6 is now in ChatGPT for everyone
It comes with this Intelligent UI thing where answers come out more visual and interactive
Free users get it too
xitter.cf/OpenAI/status/21078949…
Las interfaces siguen en pie. ChatGPT lo sabe, ahora tú también 😁 el gusto por el diseño sigue siendo necesario! 😘
We are very close. This can generate slide explanation for you: xitter.cf/openai/status/21078949…
parecía lejos, ya la tenemos por aquí
UI generativa, adaptable, liquida
Empieza a ser aburrido estar asombrado todo el día
Hace casi tres años decíamos esto en una charla del Instituto Tramontana.
This works surprisingly well, much better than just text.
I Think SORA is still most compatible video making ai till now?
We're heading toward composing entire canvasses just in time, perfectly suited for the task.
And even better when this is activated via a continuous voice interaction with slides animating in and out
Añadieron hace unos días esto del intelligent UI y se nota en las respuestas más complejas que lo aprovechan bien.
AI this AI that, I just tried GPT-6.1 versions and they are borderline r*
Now I understand why all completely made AI games look so terrible; it's not even close to being decent as an assistant.
Everyday questions getting visual answers is such a nice change. Seeing it pop up everywhere already.
curious whether speed holds up on harder tasks
これだってよこれ！
数日前に言った「チャッピーの（個人的に）すげぇ進化！」っていうのは。
🟥いやーな書類も
🟥ワケワカラン文章も
↓
🐶ぜーーんぶ、図解してくれるんよ！神！
ChatGPTの「Pro」、プラン名と推論設定で話が違うんですね。
10月に紹介されたIntelligent UIは、会話の中に図や操作できるツールを出す機能。公式の説明ではInstant〜Extra Highが対象で、推論設定のProは対象外です。Proプラン自体が非対応、という意味ではありません。
操作できる説明を使いたい時は、契約名だけでなく選んでいる推論設定も確認したい。なお、回答形式は質問に応じて選ばれるので、毎回図が出るわけではありません。
help.openai.com/en/articles/…
👎I’ve tested GPT-6 in Chat enough to conclude that it’s a downgrade in terms of language and context comprehension compared to 5.5 and 5.6.
It took OpenAI nearly a year to make GPT-5 linguistically creative, yet GPT-6 feels like a regression, a blend of 5.2 Instant and 5.3
GPT-6 annoyingly keeps repeating "I'm not a human", for no apparent reasons, the jokes are unfunny and inappropriate.
GPT-6 has certain verbal tics and intrusive patterns: such as using italics to describe actions, or posing questions that sound like those of a retail salesperson.
With 5.6 I started using Codex for the first time, and we built many applications for both daily life and work. The split into a "Work" model and a "Chat" model has made both versions worse.
It seems OpenAI fails to recognize its own mistakes, if version 5.6 disappears, I see no reason to stick with GPT-6.