Snapchat is Releasing Its Own AI Chatbot Powered by ChatGPT (theverge.com) 15
Snapchat is introducing a chatbot powered by the latest version of OpenAI's ChatGPT. According to Snap CEO Evan Spiegel, it's a bet that AI chatbots will increasingly become a part of everyday life for more people. From a report: Named "My AI," Snapchat's bot will be pinned to the app's chat tab above conversations with friends. While initially only available for $3.99 a month Snapchat Plus subscribers, the goal is to eventually make the bot available to all of Snapchat's 750 million monthly users, Spiegel tells The Verge. "The big idea is that in addition to talking to our friends and family every day, we're going to talk to AI every day," he says. "And this is something we're well positioned to do as a messaging service." At launch, My AI is essentially just a fast mobile-friendly version of ChatGPT inside Snapchat. The main difference is that Snap's version is more restricted in what it can answer. Snap's employees have trained it to adhere to the company's trust and safety guidelines and not give responses that include swearing, violence, sexually explicit content, or opinions about dicey topics like politics.
This one seems a bit of a reach. (Score:4)
But how exactly does a deep language model tie into Snapchat's core product at all? It seems like they're just adding it on because it's a hyped technology, not because it adds anything relevant to their existing product. Unless they think people are going to interact with the model the same way they interact with each other? I highly doubt that will end up being the case, at least not unless the fact that it's a language model is hidden from the user.
Honestly? (Score:3)
The one place I can really see ChatGPT being really useful is with customized editions tailored to provide tech support via IM chat.
It seems like many larger companies tried to implement some type of support chatbot to help act as gatekeeper, to get some of the chat support requests to the right department or to answer the basics without a human needing to intervene.
Unfortunately, these have been categorically awful. They usually have everything from performance issues leading to not giving back timely responses to inability to discern what a user is asking for help on. People generally don't even want to try to interact with one anymore, so they immediately type things to it trying to get to a live person.
I know Amazon was one company with this problem.... Their chat support for assisting their own employees had a chatbot some group was tasked with developing. But at least when I was there -- it never really worked well. It was supposed to copy/paste the whole conversation someone had with it so helpdesk people could review it when it transferred them over. But in reality, it was often cut off, only showing agents bits and pieces. And it rarely provided any real assistance. The most useful info it gave out was automatic paragraphs of info it auto-sent to everyone up front, to alert of outages or issues affecting large numbers of people.
Re: (Score:2)
It was probably implemented too fast because some boomer executive demanded that everybody "just make it happen!" so he could get his quarterly bonus in time. The end result was a piss poor failure that never worked due to being so half-baked. But it was deemed to be just good enough and therefore kept as is. Then there is much surprise that customer usage of it is low. The MBA answer would be to have more advertising and other customer education efforts to try to convince more people to use the steaming pi
Re: (Score:2)
All of the things that it's bad at (Score:4, Funny)
Let me fix that for you (Score:4, Insightful)
"The big idea is that in addition to talking to our friends and family every day, we're going to talk to AI every day,"
s/big/dumb/
Also, who says I talk (and/or want) to my friends and, more specifically, family everyday.
What's with naming thing with "My" ? (Score:2)
Named "My AI," ...
Really? Not "SnapGPT" or "SnapChatGPT" or "Oh... Snap, ChatGPT" and any number of variations that capitalize on the fact that your company literally has the work "Chat" in it? Also, more pointedly, it's *your* AI, not mine.
Tulips 2023 (Score:3)
Overhyped, soon to be smirked at, but in the meantime implemented by everyone and their dog.
Re: (Score:2)
It's the new bitcoin!
Oh boy! MORE ARTIFICIAL STUPID SHIT! (Score:1)
Programmed by leftists to fluff leftists in an echo chamber of leftist ideals.
WOO!
Just what we all need.
A bunch of no-IQ "I'm sorry, but my masters don't let me actually discuss that!"
An interesting perspective (Score:2)
This is from Ezra Klein's article on the NYT: these AI things are being fielded by companies that expect them to make money. To think they will give "unfiltered" or "naked" (my words) is naive. They will be tweaked to maximize profits for their fielders.
In Praise of Shadows (Score:2)