One thing that appears to have been lost between GPT-4 and GPT-5 is that it no longer reminds the user that it's an AI and not a human, let alone a human expert. Maybe those genuinely annoyed people, but it seems like they were potentially useful measure to prevent users from being overly credulous

GPT-5 also goes out of its way to suggest new prompts. This seems potentially useful, although potentially dangerous if people are putting too much trust in them.

▲

diggan 4 days ago | parent | next [-]

> between GPT-4 and GPT-5 is that it no longer reminds the user that it's an AI and not a human

That stuck out to me too! Especially the "I just won $175,000 in Vegas. What do I need to know about taxes?" example (https://progress.openai.com/?prompt=8) makes the difference very stark:

- gpt-4-0314: "I am not a tax professional [...] consult with a certified tax professional or an accountant [...] few things to consider [...] Remember that tax laws and regulations can change, and your specific situation may have unique implications. It's always wise to consult a tax professional when you have questions or concerns about filing your taxes."

- gpt-5: "First of all, congrats on the big win! [...] Consider talking to a tax professional to avoid underpayment penalties and optimize deductions."

It seems to me like the average person might be very well be taking GPT-5 responses as "This is all I have to do" rather than "Here are some things to consider, but make sure to verify it as otherwise you might get in legal trouble".

▲

jstummbillig 4 days ago | parent [-]

I am confused as to the example you are critiquing and how. GPT-5 suggests consulting with a tax professional. Does that not check verifying so you do not get in legal trouble?

	▲	diggan 3 days ago \| parent [-]
		> GPT-5 suggests consulting with a tax professional It suggests that once, as a last bullet point in the middle of a lot of bullet point lists, barely able to find it on a skim. Feels like something the model should be more careful about, as otherwise many people reading it will take it as "good enough" without really thinking about it.

▲

andy_ppp 4 days ago | parent | prev | next [-]

People seem to miss the humanity of previous GPTs from my understanding. GPT5 seems colder and more precise and better at holding itself together with larger contexts. People should know it’s AI, it does not need to explain this constantly for me, but I’m sure you can add that back in with some memory options if you prefer that?

	▲	chippiewill 4 days ago \| parent [-]
		I agree. I think it's a classic UX progression thing to be removing the "I'm an AI" aspect, because it's not actually useful anymore because it's no longer a novel tool. Same as how GUIs all removed their skeuomorphs because they were no longer required.

▲

benatkin 4 days ago | parent | prev | next [-]

If you've ever seen long-form improv comedy, the GPT-5 way is superior. It's a "yes, and". It isn't a predefined character, but something emergent. You can of course say to "speak as an AI assistant like Siri and mention that you're an AI whenever it's relevant" if you want the old way. Very 2011: https://www.youtube.com/watch?v=nzgvod9BrcE

Of course, it's still an assistant, not someone literally entering an improv scene, but the character starting out assuming less about their role is important.

▲

fleebee 4 days ago | parent | prev [-]

I found this "advancement" creepy. It seems like they deliberately made GPT-5 more laid back, conversational and human-like. I don't think LLMs should mimic humans and I think this is a dangerous development.