[AKSF_Ackermann]

They stepped on every single rake possible, didn't they? 1. Why are you making a phone call in the first place, your agent probably got the number from the internet, just keep using that. 2. If you insist on initiating the conversation over a phone call, why not immediately terminate the call and again, go over the internet once you realize that it is an ai to ai conversation. 3. You did in fact re-invent a modem but worse, the quoted speed on that library is 8-16 bytes/sec, and i would like to point out that the Bell 103 did ~37 bytes/sec, and was released in 1963.

[bibimsz]

It comes down to the problem statement and what the constraints are. This is solving for using the phone-only scenario, which is perfectly valid.

If you want to address a phone-with-internet-backchannel, that's valid too - but it assumes different problem statement and constraints.

[AKSF_Ackermann]

Please pay more attention to the point 3 in my original post. To reiterate: their encoding is hilariously bad, and is easily outcompeted by a modem from the 60s.

[0xDEFACED]

youre missing the forest for the trees. the library this demo is using for audio encoding (ggwave) was not made by the creators of this demo. speed (or lack thereof) aside, having a direct audio<->text encoding is much more computationally efficient than speech<->text generation.

on the subject of the encoding efficiency, the ggwave depo mentions the use of reed-solomon error correction to make transmission more reliable. im struggling to find any info on error correction used by bell 103 or other modems, but if they aren't as robust that could partially explain the discrepancy you're describing

[swexbe]

Sounds more futuristic than old dial-up sounds though

[godelski]

I think the most important part is the bitrate. As you said elsewhere: "time is money". Seems like you're not saving that much money

[rglullis]

4. If you are an agent receiving a call, why not announce it right away?