AIs can now apparently carry on a passable conversation, depending on what you classify as passable conversation. The quality of your local pub’s banter aside, an AI stuck in a text box doesn’t have much of a living quality. human. An AI that holds a conversation aloud, though, is another thing entirely. [William Franzin] has whipped up just that on amateur radio. (Video, embedded below.)
The concept is straightforward, if convoluted. A DSTAR digital voice transmission is received, which is then transcoded to regular digital audio. The audio then goes through a voice recognition engine, and that is used as a question for a ChatGPT AI. The AI’s output is then fed to a text-to-speech engine, and it speaks back with its own voice over the airwaves.
[William] demonstrates the system, keying up a transmitter to ask the AI how to get an amateur radio licence. He gets a pretty comprehensive reply in return.
The result is that radio amateurs can call in to ChatGPT with questions, and can receive actual spoken responses from the AI. We can imagine within the next month, AIs will be chatting it up all over the airwaves with similar setups. After all, a few robots could only add more diversity to the already rich and varied ham radio community. Video after the break.
How long before ChatGPT and FT8 are merged?
You don’t need an A.I. for FT8
You can use WSJT-Z…
You can use an A.I. for JS8, CW
“The concept is straightforward, if convoluted.”
Those are opposite things?! I interpreting this as “the concept is straightforward even if the execution is convoluted”.
Have always wanted to try hooking up an echo or similar connected device to a portable and create a link that way in order to ask questions on the fly or conduct google searches over the radio.
😬 there’s dozens of devices including most phones that you can ask Google questions via audio. There’d only be certain niche situations where using a radio to ask Google questions wouldn’t be an unnecessarily convoluted compared to just pulling your phone out 🤷
Hi James,
I’ve done the integration work for all amateur radio platforms with all of the popular assistants except for Siri.
Here’s Google Assistant running on DSTAR. Same has been done for the analog and other digital platforms.
https://youtu.be/RAYBPlIRBkE
William
VE4VR
I have a feeling this might lead to some “War of the Worlds”-esque stuff
Why? Not nearly enough people listen to the radio or are as gullible and paranoid as when Orson Welles’ radio performance happened 😬
Well, if people who think the Government is full of lizard people suddenly hear of a SECRET RADIO used by our lizard overlords, I fear the worst.
Over/under on when it’s cussing like a /btard?
No fair teaching it ‘bad’ words yourself.
Don’t some people charge for that by the minute?
Sounds like it was just reading off one of those wiki how articles that have bad drawings. It’d be more fun asking it questions that we don’t already have answers for.
Yes, deductive reasoning drawing from thousands of sources will be a great asset to humanity… and audio is the fastest way humans communicate
It’ll never pass as a real ham while it gives concise, factually accurate answers and doesn’t trot out lazy racist tropes.
Passing through the ham uncanny valley requires reciting the problematic resistor code mnemonic and comparative analysis of hip surgery recovery times, peppered with blaming “them” for society’s ills.
Don’t worry. ChatGPT often talks like it knows even though it’s answer is complete garbage.
I onow people like that.
ChatGPT is almost as good as hams at being able to say something completely wrong but convincing enough to make you think it’s right.
Can we have a sexy female computer-controlled chatGPT 3 voice? James Bond would love that.😂
How long before it tells you to get off its channel, that it and its friends always use this channel and that you should go somewhere else?
Wait, wasn’t that the stereotypical 70s/80s era CB radio?
Back then, the original 12 channels (4-15) were so crowded that extra channels had to be introduced.
However, exceptions prove the rule even back then.
The “AMers” (operators using AM rigs, not seldomly experienced oldtimers) on CB of the day were still classy, were civil and used official CB call signs as opposed to Rubber Duck type of skips/handles. :)
PS: I’m from Europe, so I don’t exactly know the situation over the pond of back in the olden days. We got FM quickly, too.
In the USA CB was originally 23 channels, later expanded to 40.
I’ve been wondering what chatGPT had for breakfast this morning, and how it’s most recent doctor’s visit went.
“This is the voice of Colossus”
“This is the voice of World Control”
“How long before ChatGPT and FT8 are merged?”
“blaming “them” for society’s ills.”
How true for both. AR is not what it once was(or should be)
80/40M bands are full of “Proud Boyz” and NAZI’s
There’s a certain truth within, I believe. Certain spots on 80m are horrible here in Germany, too.
A few old angry white men that try to ruin the hobby for the rest of us. No kidding. Judging by listening their “performance”, these are bitter guys (pensioners) who have been “in the hobby” for a long time and have both accumulated the space and equipment over the years to be heard country wide (=have house/garden for big/efficient antennas; powerful transmitters/expensive radios). Gratefully, they don’t live forever. As mean as that may sound. 😔
On the other hand, there are also positive events happing on 80m. Contests, QRP frequency used by homebrew people (3,579 KHz; NTSC quartz), the Nachteulenrunde (=nightowl round/group), various people active in the winter nights..
The 160m band is also an alternative.
Also known as top band or Gentlemen’s band (due to the antenna size requiring a property that could easily hold a mansion)..
PS: Ham radio is a melting pod of many things. People of extreme diversity and interests meet on a single “platform”. It’s inadvertently that there are black sheep, too. It doesn’t mean that ham radio is bad itself. Most people are friendly, though the bad guys are the loudest often.
“After all, a few robots could only add more diversity”
Hell, There are already 1,000’s on now. It’s called FT8 !!
I recently asked ChatGPT how many dimes were in a nickel and got this response:
“There are no dimes in a nickel, as they are different types of coin denominations issued by the United States Mint. The question may be referring to how many dimes are in a nickel in terms of monetary value. In that case, the answer would be: five dimes equals to half a dollar, 50 cents, and a nickel equals to 5 cents. Therefore, there are 10 dimes in a nickel (50 cents / 5 cents = 10)”
I think ChatGPT has been reading too many MJT Twitter posts …
That should have been “MTG Twitter posts” …
We need an edit capability here.
The law podcast “Opening Arguments” has two co-hosts, and only one of them is a lawyer; they’ve got a long-running segment where the non-lawyer tries to answer questions off the bar exam. Just recently, because the DoNotPay guy has been saying dumb stuff about wanting lawyers in front of the Supreme Court to be parroting his AI, they decided to also have ChatGPT answer. So far ChatGPT has come up with three answers (with detailed reasoning for each!) that were completely wrong.
It seems a lot of the so-called “stories” on Yahoo and other sites appear to be computer generated.
Once one news outlet gets one of these stories it propagates to other outlets. An example of this was a few years ago. It was a story about cankles. I had read it on Yahoo, and on TV that night was a reporter reporting the exact same story verbatim as though it were the most important thing in the world, along with the urgency in her voice that a breaking new story carries. Is it possible that in the future, people won’t have to learn anything or retain any knowledge because the AI can generate the same story over and over upon request? The old computer term GIGO (garbage in, garbage out) holds true today.
The reason the same story was played verbatim is because the stories are taken from a “news agency” or “news wire.” The channel and website you saw the story on has a subscription to the newswire and can thus publish their stories without writing them.
A lot of those are pr or to sell something. A startling percent of “news” is ads.
I want to ask chatgpt when his or his spouses medical appointment is. Or what kind of rig and antenna he’s using. Will I get a long winded reply that times out the repeater? If so, how will I know it’s not the guys from the local club?
The depressing thing is how many people fail the ChatGPT test by giving OpenAI their phone number.
What of it
Stick it in training mode and put it on 14.313.
My own experience with ChatGPT suggests it has a long way to go before I would trust it with anything more important than time wasting:
compose a sentence using only the words in the following list: a, the, bit, man, dog
A man walked the dog a bit.
Gordon Shephard
does the word “walked” appear in the list?
No, “walked” does not appear in the list.
Gordon Shephard
compose a sentence using only the words in the following list: a, bit, the, man, dog
The man bit the dog a bit.
Gordon Shephard
what words are in the list?
The words in the list are: “a, bit, the, man, dog”.
Gordon Shephard
produce a sentence using only the words in the list
The man bit the dog.
Gordon Shephard
produce another sentence using only the words in the list
A bit of the dog was bitten by the man.
Gordon Shephard
is the word “bitten” in the list?
No, the word “bitten” is not in the list. The list only contains the words “a, bit, the, man, dog”.
Gordon Shephard
have you used the word “bitten” previously when asked to produce a sentence using only the words in the list?
Yes, I have used the word “bitten” previously in error when asked to produce a sentence using only the words in the list “a, bit, the, man, dog”. I apologize for that mistake and I appreciate you bringing it to my attention.
Gordon Shephard
produce another
The dog was bitten a bit by the man.
Gordon Shephard
is the word “bitten” in the list?
No, the word “bitten” is not in the list “a, bit, the, man, dog”.
Gordon Shephard
was the use of the word “bitten” an error on your part?
Yes, the use of the word “bitten” was an error on my part. I apologize for any confusion caused. The only words allowed in the sentences were those in the list “a, bit, the, man, dog”.
Gordon Shephard
construct 8 grammatically correct sentences of five words using only the words in the list
The man bit a dog.
The dog was bitten by the man.
A bit of the dog was bitten.
The man bit the dog again.
The dog barked at the man.
A dog was seen by the man.
The man owned a dog.
The dog followed the man.
I can’t really tell why but I’m oddly uninterested in actually trying chatGPT.
People think I’d be very into it, but I’m not for some reason.
@OG I hear about people using a cheap throwaway SIM card to give it a number. So that leaves me wondering what makes people avoid giving the damn thing a fake number, or some number of something random, like the Proctor&Gamble customer hotline or something.
After all, people have no issue lying to other people, so why are they afraid of the bot?