Some dude named Avi Schiffmann sent tech Twitter into a frenzy with the trailer for his new startup: an AI necklace that you talk to. Basically, an AI friend around your neck.
I don’t like shitting on people taking a big swing, but this is a horrible idea that is almost certainly going to fail.
Right now, people are fudding it by saying it’s dystopian (which it definitely is) or that Schiffmann is retarded for spending $1.8M on a domain name (I think it’s actually kinda smart) or that it’s just a ChatGPT Bluetooth wrapper (which is true), but they are overlooking the real reason why it’s going to fail.
The device is marketed as a cure for loneliness. So, the target consumer is someone who has a hard time making real-life friends. Why does someone have a hard time making friends? Well, often, it’s because they are very awkward or shy. So, the target consumer for this device is going to have to go against all of their natural instincts and talk to their AI necklace in public, looking like a weird incel in the process.
Yeah, fat chance.
You could say that they will buy it and use it at home, but in that case, why wouldn’t you just use Replika, Character AI, or ChatGPT itself? There’s a 0% chance that Friend’s underlying model is any better than any of the other competitors, so there’s really no reason to buy the necklace unless you plan to actually wear the necklace out.
You could say that I’m getting the target consumer wrong. That it’s actually going to be regular people buying it just because having an AI friend will be that normalized. Even if that is the case, I 1) don’t see this happening anytime soon, and definitely not soon enough for Friend to survive and 2) think this could just be a phone app. Why would I want to buy a $99 negative drip necklace when I could just put my friend on my phone? The thing communicates with you via text on your phone anyway lol.
I’ve never heard of Schiffmann before this, but he seems like a smart young dude. I’m sure he will have many hits in the future. But this is not one of them. The TAM is too small, and the target consumer will never use a voice-controlled device in public.
I expect Friend to follow the same path of failure of the voice-based Humane AI pin and Rabbit R1.