The features stand to be a significant upgrade to the current experience for ChatGPT free users, who were until now stuck on the text-only GPT-3.5 model, lacking the powerful capabilities of GPT-4 to analyze images and documents uploaded by users.
Now free ChatGPT users will have access to a significantly more intelligent model, web browsing, data analysis & chart creation, access to the GPT Store, and even memory so the app can store information the user wants about them and their preferences simply by typing or asking it audibly.
GPT-4o reasons across voice, text, and vision” Murati said. That includes accepting and analyzing realtime video captured by users on their ChatGPT smartphone apps, though this capability is not yet publicly available. In a blog post, OpenAI wrote: “GPT-4o is much better than any existing model at understanding and discussing the images you share.” OpenAI also noted that while it would eventually be available to free ChatGPT users, GPT-4o would first roll out to paying subscribers.

Does the freemium model scale?

Looks like the freemium model of giving-up your privacy & data is still their preferred way of keeping this "accessible" to most. Is this a model for mass-adoption? Time will tell. It's clear however this is capable of processing way more data for tech giants than even the biggest privacy-pooper would feel comfortable with. I'm personally still betting on private FOSS AI winning-out comfortably.
I didn't find the demo authentic. It felt pre-recorded and even if it's authentic the voice is too nice and becomes annoying after few minutes.
reply
The voice annoyed me too. But then I had a thought that you will be able to customize it to whichever sex, whatever age, accent, and even personality you want in the future. Imagine the impact on the call center industry. Pretty soon we will be all be interacting with voice or chat bots that match our preferences. Not sure if utopia or dystopia but either way the future looks to be an inexorable slide into an atomized, entirely curated existence for each individual plugged into the matrix, and even for those who try to opt out, any interaction with government or corporations is likely to be mediated by some AI trying to tailor the interaction to best engage you.
reply
The bots at call center are limited by programmers. I don't think these bots will be given as much access as human agents.
reply
that's somebody's girlfriend in a couple of months. Be respectful!
reply
It should be able to automatically or manually adapt the level of expression or totally different mood, depending on the context of the conversation. Otherwise yeah it would be annoying if it is all the time like that. A demo is a demo. We will have to test it and see.
reply

Absolutely dope.

reply
21 sats \ 1 reply \ @gmd 13 May
Excited about the real-time translation capabilities. Half of my patients frustratingly don't speak English and it takes 4-5x the time of a normal English encounter.
The ipad/iphone human translation services are so bad (take 1-5m to get one online, then they waste so much time introducing themselves.
Sad to know jobs will be lost because of this but I literally despise using the human services.
reply
Jobs will be lost but the need for human input and supervision will never go away. We just got new tools.
reply
Thanks for the heads-up. As a teacher who relies on AI to plan his lessons faster, I m intrigued.
I think many people would happily give up their privacy n data. Just look at World Coin. People queued up for hours to have their irises scanned for ten bucks
reply
Is GPT the leading ai model? Or are there other ones that are better?
reply
The AI won't stop talking so the guy needs to interrupt it. lol
Also: I see why they chose this voice for their totally-not-a-girlfriend AI product.
reply