[HN Gopher] Crossing the uncanny valley of conversational voice
       ___________________________________________________________________
        
       Crossing the uncanny valley of conversational voice
        
       Author : monroewalker
       Score  : 357 points
       Date   : 2025-03-02 06:13 UTC (16 hours ago)
        
 (HTM) web link (www.sesame.com)
 (TXT) w3m dump (www.sesame.com)
        
       | monroewalker wrote:
       | This was already posted here:
       | https://news.ycombinator.com/item?id=43221377 but I'm really
       | surprised at the lack of attention this model is getting. The
       | responsiveness and apparent personality are pretty mind blowing.
       | It's similar to what OpenAI had initially demoed for advanced
       | voice mode, at least for the voice conversation portion.
       | 
       | The demo interactions are recorded, which is mentioned in their
       | disclaimer under the demo UI. What isn't mentioned though is that
       | they include past conversations in the context for the model on
       | future interactions. It was pretty surprising to be greeted with
       | something like "welcome back" and the model being able to
       | reference what was said in previous interactions. The full
       | disclaimer on the page for the demo is:
       | 
       | " 1. Microphone permission is required. 2. Calls are recorded for
       | quality review but not used for ML training and are deleted
       | within 30 days. 3. By using this demo, you are agreeing to our "
       | 
       | edit: Actually this has been posted quite a few times already and
       | had good visibility a couple days ago: -
       | https://news.ycombinator.com/item?id=43200400 Others:
       | https://hn.algolia.com/?q=sesame.com
        
         | hn_user82179 wrote:
         | It was genuinely startling how human it felt. Apparently they
         | are planning on open-sourcing some of their work as well as
         | selling glasses (presumably with the voice assistant). I'm very
         | excited to have a voice assistant like this and am almost a bit
         | worried I will start feeling emotionally attached to a voice
         | assistant with this level of human-like sound.
        
           | jofzar wrote:
           | I still feel like they don't have the right amount of human
           | to them, maybe it's because I'm Australian and it sounds like
           | I'm hearing an American robot?
           | 
           | Edit: well I asked the "male" model to speak more like an
           | Australian and yep, getting way more uncanny. If it had an
           | Australian accent I think it would mess with me more
        
           | igleria wrote:
           | Maybe the ability to personalize the voice so it is more...
           | robotic or based on a fictional thing like Knight Rider would
           | help to change the attachment to something more... healthy?
        
             | MarcelOlsz wrote:
             | Yeah this is straight up creepy, and I also can't stand
             | chatgpt saying "Lmao" and "Yeah". Keep it formal & robotic.
        
               | WesolyKubeczek wrote:
               | What ever did you tell ChatGPT so it responded with
               | "lmao"?
               | 
               | I told it that it should behave explicitly like a
               | computer in the system prompt, sort of worked.
        
               | MarcelOlsz wrote:
               | After multiple prompts and utterly garbage output:
               | https://i.imgur.com/5aOARCV.png
               | 
               | I'm almost positive that some AI systems have a backend
               | that analyzes the sentiment of your messages and if you
               | threaten to cancel billing it will notice your defcon-1
               | sentiment and spin up some more powerful instances behind
               | the scenes to tide you over.
               | 
               | This is actually much more stressful than working without
               | any AI as I have to decompress from constantly verbally
               | obliterating a robotic intern.
               | 
               | I'll try with the system prompt. Also love your username.
        
         | anon373839 wrote:
         | It really is an astonishing technological feat! Also note that
         | the largest model they trained is only 8.3B parameters (8B
         | backbone + .3B decoder). It's exciting to think that they're
         | going to be releasing this model under an Apache 2.0 license.
        
         | Mistletoe wrote:
         | Just realizing how uncanny valley it is to talk to AI and it
         | never remembers anything you said in the past. Imagine if a
         | human did that. It's like you are talking to Tom Hanks' Mr.
         | Short Term Memory from SNL over and over.
         | 
         | https://youtube.com/watch?v=C6ufImch00g
        
           | ekianjo wrote:
           | That can easily be fixed if you attach it to a RAG system
        
         | huijzer wrote:
         | > This was already posted here:
         | https://news.ycombinator.com/item?id=43221377 but I'm really
         | surprised at the lack of attention this model is getting.
         | 
         | I'm surprised by the lack of attention that Gemini 2.0 with
         | native audio output got. They have a demo at
         | https://youtu.be/qE673AY-WEI, which I think is really good too.
         | The main problem with Google's model is that this audio output
         | is not supported by the API, but you can try it at
         | https://aistudio.google.com.
         | 
         | In general, text to speech is pretty good nowadays I think. For
         | example, this is a little math video that I made a few days
         | ago: https://www.youtube.com/watch?v=G1mvLrCfjFM with the (old)
         | Google text to speech API. Honestly, I think the narration is
         | better than I personally could have done. It's calm, well
         | pronounced, and sounds relatively enthusiastic.
        
           | smusamashah wrote:
           | How do I get to this in aistudio.google.com?
        
             | huijzer wrote:
             | I think the one under "Stream Realtime" should be similar
             | to the demo. It's only Gemini 2.0 flash though and not the
             | full one.
        
           | moralestapia wrote:
           | >They have a demo at https://youtu.be/qE673AY-WEI
           | 
           | That's not a demo, that's a video. Anyone can make something
           | like that in an afternoon with a couple friends and a
           | microphone.
           | 
           | Also, Google is known for putting out fake "demos", remember
           | the Google Duplex scam?
        
             | underdeserver wrote:
             | Scam? Duplex worked.
        
               | moralestapia wrote:
               | I doesn't work today, let alone 6 years ago.
               | 
               | But good work defending your master.
        
               | jazzyjackson wrote:
               | I thought it was announced and never heard from again. It
               | may have worked but it never shipped did it?
        
         | znpy wrote:
         | > 2. Calls are recorded for quality review but not used for ML
         | training and are deleted within 30 days.
         | 
         | Sounds (pun intended) reasonable.
        
       | drvladb wrote:
       | Definitely an improvement over your normal Text-To-Speach model,
       | and to some degree really different, but the subtle imperfections
       | do appear and ruin the overall perception. A move in the right
       | direction, though, I suppose.
        
         | monroewalker wrote:
         | Yeah after a few interactions, the repetition of the mannerisms
         | that initially added to the sense of life-likeness started to
         | break the illusion a bit. The "you got me" response shows up a
         | bit too often. The creativity remains impressive though
        
       | kats wrote:
       | AI voice is an overwhelmingly harmful technology. It's biggest
       | use will be to hurt people.
        
         | valenterry wrote:
         | Erm, no. Its biggest use will be...
         | https://www.youtube.com/watch?v=LTJvdGcb7Fs :-)))
        
         | meowface wrote:
         | It will unfortunately undoubtedly be used for mass automation
         | of scams but text AI (and pre-AI automation) have been used for
         | that for many years as well. Doesn't really make sense to say
         | "ok we should allow all forms of AI besides voice because of
         | scams", I think.
         | 
         | But yes, there needs to be some spreading of public awareness.
        
           | paul7986 wrote:
           | That's if you answer phone calls from numbers not already in
           | your contacts. For me all such numbers go to voicemail and if
           | the voice is of someone i know ill just call them directly.
           | 
           | If you do any of the above you are looking to be scammed!
        
             | alcover wrote:
             | Oh yes ? The scambot will leave a distress message and a
             | number in your voicemail, using the voice of a relative.
             | You would know better but I guarantee old people will call
             | the number and strike a convo with the virtual relative.
        
         | pmarreck wrote:
         | Nope. Awareness will inoculate people. "Authenticating" someone
         | via the mere sound of their voice was always broken, anyway...
         | Ever see the great movie Sneakers (1992)?
        
           | FranzFerdiNaN wrote:
           | Do you live in reality? Because that clearly isn't happening.
        
             | pmarreck wrote:
             | And phones enable scams. So your idea is to... Abandon all
             | telephony??
             | 
             | You should not judge a tool by the worst use someone can
             | come up with it
        
               | kats wrote:
               | It's the most common use by far.
        
         | EGreg wrote:
         | Cue all the responses saying "it's already been possible to
         | harm people, AI doesn't fundamentally change anything, nothing
         | to worry about"
        
           | hombre_fatal wrote:
           | Counter point: We were barely doing anything about it when
           | bad actors were pwning people pre-AI, like with social media
           | propaganda or romance scams.
           | 
           | And if we still do nothing about it post-AI? Well, that is
           | already the status quo, so caring now feels performative
           | unless we're going to finally chit chat about solutions.
           | 
           | The same could be said for the internet. "The internet can be
           | used for bad" is an empty, trivial claim, not an insight that
           | needs a standing ovation. The conversation we need is what to
           | do about it. And the solutions need to be real ones, not "we
           | need to put the cat back in the bag".
        
         | IshKebab wrote:
         | Doubt it. Its biggest use will be voice assistants.
        
         | alcover wrote:
         | I unfortunately agree with you. Old people with
         | confusion/dementia, schizoid types, or very naive persons will
         | fall for shattering scams. And the consequences on their grasp
         | on reality will be terrible.
        
       | bobosha wrote:
       | Very impressive. well done team sesame!
        
       | tobr wrote:
       | I asked it if it could whisper, and it replied in full voice,
       | "I'm whispering to you right now".
        
         | monroewalker wrote:
         | lol yeah I tried to get it to whisper too. And talk faster or
         | slower or do accents. It seemed to be able to kind of do each
         | of those things but only very slightly. Enough to see that
         | there was some successful interpretation of the request but
         | lack of flexibility to fully execute on it. OpenAI's model
         | still has this beat on that front imo (talking quietly / slower
         | / faster)
        
           | mrkstu wrote:
           | The male's Australian accent consisted of throwing a 'fair
           | dinkum' in while keeping it's vague New York accent.
        
         | IshKebab wrote:
         | Yeah it's definitely going through text still. I tried to get
         | it to sing a song so it output some lyrics and then read them
         | as a poem.
         | 
         | I did manage to get it to output "la la la la" and then it kind
         | of sang them with a random melody.
         | 
         | It also can't say things loud and its idea of whispering for me
         | was to say "pst".
         | 
         | Still apart from that it's very impressive!
        
       | thekevan wrote:
       | It's good, but it still sounds fake to me, but in a different
       | way. The voice itself sounds like a human, undoubtedly.
       | 
       | But the cadence and the rhythm of speaking are off. It sounds
       | like someone who isn't a podcaster trying to speak in the
       | personality of a podcaster. It just sounds like someone trying
       | too hard and speaking in an unnatural way.
        
         | dmix wrote:
         | Humans are extremely well tuned to detect authenticity in
         | communication. Especially younger generations raised on mass
         | marketing.
         | 
         | This is good in a way a scifi movie shows a tech, sounds cool
         | and demos futuristic possibilities. But not quite passing the
         | real human vibe yet. But I'm sure some people might find it
         | preferable to a more to-the-point system like GPT or Siri/Alexa
         | in certain niche cases not requiring immediate gratification.
        
           | thfuran wrote:
           | >Humans are extremely well tuned to detect authenticity in
           | communication.
           | 
           | I think the long-standing success of advertising and
           | propaganda suggests that people really aren't all that good
           | at that.
        
             | SamPatt wrote:
             | I suspect success of advertising is less about people
             | falling for deception and more about information
             | availability. If you know nothing about two brands except
             | you've heard the name of one 50 times in ads, you'll
             | probably try it first.
             | 
             | I think propaganda is a better example, although again I
             | think often people aren't deceived, they simply agree with
             | the message or don't care about the underlying truthfulness
             | of the message and just use it as a way to align with their
             | tribe, etc.
        
         | anal_reactor wrote:
         | A few times the CEO of my company randomly joined me for lunch,
         | but each time he forgot to leave behind his persona of "I'm a
         | public speaker right now", making the whole situation feel
         | extremely awkward. This AI gives me exactly the same vibes.
        
         | akie wrote:
         | I also think it didn't feel very "real". Trying too hard to
         | sound upbeat and too eager to please, maybe it's just me being
         | European but it makes me go "ewww, that's not how normal people
         | speak".
        
           | youssefabdelm wrote:
           | Yeah the eagerness to please thing feels like it carried over
           | from the LLMs or something cause they're like that too.
        
           | walrus01 wrote:
           | It sounds like a "sales and marketing coordinator" for
           | something very tech-bro adjacent after two strong cups of
           | coffee.
        
           | t0lo wrote:
           | That's american office culture for you. If it was australian
           | it'd be drab, boring, and self-flagellating
        
         | hereonout2 wrote:
         | This is an interesting take, and I'd guess that the training
         | data for this probably did use podcasts as a source.
         | 
         | Getting very realistic / real world conversational training
         | data for an ai would be hard. Only a subset of us appear on
         | podcasts, radio or tv and probably all speak in a slightly
         | artificial manner when we do.
        
           | jofzar wrote:
           | I agree, I thinks it's probably very easy to find billions of
           | hours of conversation on YouTube, but non of it is set to
           | training data with a good transcript.
        
             | hereonout2 wrote:
             | Yep! it's public dialogue, intended for an audience with a
             | prepared topic, etc. Or it's actors imitating private
             | dialogue, but again shaping it towards an audience.
             | 
             | AI agents like this are trying to recreate personal
             | intimacy I guess, which does feel like it might be
             | different somehow.
        
           | scoot wrote:
           | When I commented on the unnatural cadence, it told me that it
           | had been trained on podcasts, which does help explain the
           | issue - some people tend to "live-edit" themselves when a
           | conversation is being recorded, which leads to this staccato.
           | It seems they need to find a better source of training date
           | for more natural conversational speech.
        
         | keiferski wrote:
         | I tried the demo and could tell it was fake in the first five
         | seconds. IMO it sounds like it was trained on Northern
         | California founders giving a pitch for their startup. Way too
         | enthusiastic and trying too hard to sound natural.
        
           | Workaccount2 wrote:
           | For all they talk about diversity, you can pretty much
           | pinpoint every tech product to SV because they are all using
           | the same cultural cookie cutter.
        
         | BrenBarn wrote:
         | To me the actual words it used also seemed fake, sort of too
         | deliberately breezy.
        
         | jofzar wrote:
         | It sounds like someone who is doing a microphone test for
         | something they just bought and hearing themself on a delay from
         | the monitoring.
         | 
         | Yes that is very specific, but that's what it sounds like to my
         | ear.
        
         | darkerside wrote:
         | People have a performative mode and an authentic mode
         | (oversimplifying), probably including you. If you're at home
         | talking to your parents or spouse, and then suddenly realize
         | your boss is in the next room listening, does your voice
         | change?
         | 
         | Point being, this demo voice is in performative mode, and I
         | think sounds fairly natural based on that. Would you rather it
         | not?
        
       | bloomingkales wrote:
       | This is so good that it's disarming. People are going to blabber
       | everything to it, so we need a local private model. It's a lot to
       | ask, I know. Incredible tech.
        
         | adtac wrote:
         | > Our models will be available under an Apache 2.0 license.
         | 
         | ^ from the post
         | 
         | https://github.com/SesameAILabs/csm is empty for now, but I
         | imagine they'll be releasing it soon:
         | https://x.com/_apkumar/status/1895492615220707723
        
           | mentalgear wrote:
           | let's hope they stay true to that, but with a-16-z being
           | their VC, I can't imagine there isn't an ultimately
           | exploitive end game in it.
        
         | therein wrote:
         | Same, I have been trying it for the last few minutes, and it is
         | crazy.
         | 
         | Try asking if it if it speaks a different language. It will
         | pretend like it can and then give you some humor. But then you
         | probe a bit more and it tells you it is really good at
         | listening and can listen to you in other languages. I tell it
         | alright I'll talk to you in a different language but you will
         | reply back in English. It says you got it and then passes all
         | sorts of tests I put it through with flying colors.
         | 
         | Oh it also remembers your previous conversation and greets you
         | accordingly.
         | 
         | Crazy impressive this will certainly revolutionize virtual
         | office businesses.
        
           | Workaccount2 wrote:
           | We were playing with it last night and it could understand
           | Spanish, but it couldn't speak it.
           | 
           | My assumption is the LLM can translate no problem, but the
           | audio model can't do Spanish. It seemed like there was an
           | external catch to stop the model from trying too.
        
         | sivers wrote:
         | Agreed. I just had that moment like the guy in the movie "Her",
         | the first time he speaks to his OS. Laughing at myself for
         | talking to a computer like a real person. Then had to hang up
         | because it crossed that uncanny valley.
         | 
         | But then I thought of one more question to ask, reconnected to
         | ask it, and it said, "Hey! You hung up just as we were just
         | getting to the good stuff!" which threw me off, so I stammered
         | gobsmacked for a minute, and it made fun of my stammering,
         | imitating it. Whoa! So so SO good! Crazy good.
         | 
         | I'm creeped-out by this being on someone else's server, but if
         | it was fully local-hosted-private, that might even get more
         | creepy if I allowed myself to really talk freely to this thing.
        
           | energy123 wrote:
           | Now here's a little thought experiment. What does the world
           | look like in 5 years when everyone is talking to these things
           | that are indistinguishable from a real person? They will be
           | funnier, more compassionate, less judgemental, smarter, and
           | superficially "better" in every respect.
        
             | EGreg wrote:
             | I've been asking this for years.
             | 
             | People keep saying stuff like "but you'll want the human
             | touch." Really? So when was the last time you asked someone
             | for directions? Personally, I'd rather google something or
             | discuss with ChatGPT than make someone listen to me for an
             | hour. And that someone has to be extremely knowledgeable
             | about a lot of different topics!
             | 
             | Even here. Would I rather converse with y'all and get
             | downvoted sometimes, or talk to ChatGPT and refine my
             | ideas? Sorry, fellow humans... even on HN there is too much
             | irrational criticism and off-topic stuff to get anything
             | really done. Oh and you have to wait a long time for each
             | response.
             | 
             | The real question is ... what is the point of any human
             | output on the internet in a few years? Why would anyone
             | want to listen to your post, comment, or anything at all?
        
               | energy123 wrote:
               | I expect there'll be two phases to this. First phase is
               | the widespread use of disembodied voice
               | partners/friends/assistants. Then the second phase will
               | be embodiment, which gives you oxytocin from touch, etc.
               | 
               | Can't see this going well for the fertility crisis.
               | 
               | The tech oligarchs then invest in ectogenesis technology,
               | and use their sperm to dominate the gene pool.
        
               | EGreg wrote:
               | You're assuming the tech oligarchs will be men.
               | 
               | I guess it's a pretty safe assumption :-P
        
               | ok_dad wrote:
               | You should turn that into a novel, it sounds like
               | something you could sell to the romance crowd. I hear
               | sci-fi is the next genre they'll take over.
        
               | drsopp wrote:
               | They did this in Futurama
               | 
               | https://m.youtube.com/results?sp=mAEA&search_query=futura
               | ma+...
        
               | tovej wrote:
               | Because the humans are reasoning and the LLMs aren't? I
               | have yet to use an LLM for a complex problem and not have
               | it hallucinate.
               | 
               | I expect a reasonable counterargument here would be 'but
               | the LLMs have chain of thought now, and that's
               | reasoning". I disagree, but I think that's a reasonable
               | point of view. I can concede that point because it does
               | not materially change the value of the output. Even if it
               | does use chain of thought, an LLM gives you extremely
               | trite solutions based on probable text, it still has no
               | context in which to reason, it's "reasoning" in platos
               | cave using the shapes of real world objects, filtered
               | through a lossy language model.
               | 
               | LLMs are great for one thing: brainstorming, and
               | brainstorming is only useful if you have no idea what to
               | do in the first place. Once you know _anything_
               | substantial about the subject matter an LLM loses its
               | value to you as a conversation partner.
        
               | FranzFerdiNaN wrote:
               | People here are hallucinating too. So many people making
               | obviously wrong claims with full confidence, which you
               | only notice when it's about something you know a lot
               | about yourself.
        
               | timmy-turner wrote:
               | They do but we have for instance education to reduce
               | their hallucinations in narrow fields of expertise. And a
               | system of guardrails to only let educated people work in
               | those fields to avoid harm.
        
               | gonzobonzo wrote:
               | The above person was comparing it to friends and random
               | online comments, though. I wouldn't be surprised if AI is
               | far more reliable than those.
        
               | cl3misch wrote:
               | Are you not reasoning on lossy abstractions?
               | 
               | I'm still not on board with the (seemingly prevalent)
               | notion that LLM's can't reason. What's reasoning, anyway?
               | I'm not actively advocating for any side, but the
               | arguments _against_ reasoning always felt very
               | tautological to me.
        
               | tovej wrote:
               | The burden of proof is on the argument that they _are_
               | reasoning, and I have seen very little evidence that they
               | do.
               | 
               | It's also immediately clear to me when I look at the
               | architecture of transformers that reasoning is not in the
               | cards. I could be convinced otherwise if, again, someone
               | showed me an indication of reasoning behavior. Since
               | there is no such evidence and the systems theory approach
               | tells me it does not reasonably reason, I have a pretty
               | darn good reason not to believe it's reasoning.
        
               | cl3misch wrote:
               | > It's also immediately clear to me when I look at the
               | architecture of transformers that reasoning is not in the
               | cards.
               | 
               | I'm not saying that's incorrect, but thb that's exactly
               | the tautology I was talking about!
        
               | macrolime wrote:
               | This is like saying:
               | 
               | Gogole is great for one thing: brainstorming, and
               | brainstorming is only useful if you have no idea what to
               | do in the first place. Once you know _anything_
               | substantial about the subject matter Google loses its
               | value to you.
        
               | beng-nl wrote:
               | Regarding your last question - well who knows, for sure,
               | but: chess between humans is alive and well after the
               | computers became unbeatable by humans.
               | 
               | I recently listened to an interview with magnus Carlsen
               | on Joe rogan and found the angle of computers helping
               | humans to "better understand the game" (as he put it) and
               | improving human play (for learning, not playing humans)
               | to be very interesting.
               | 
               | Whether that extends to human conversation, who knows. I
               | for one would love to have a "her"-like companion, not
               | for romance but to have a highly intelligent and patient
               | and knowledgeable conversation partner to develop ideas
               | with and learn from, and endless other uses - I think
               | it'd add a lot to my and other peoples lives. I guess I
               | agree with you.
        
             | hereonout2 wrote:
             | Thinking a bit further ahead, what does the world look like
             | in 30-40 years when a generation has been accustomed to
             | this type of interaction from birth.
             | 
             | Feels like trying to imagine the societal impacts of the
             | internet in the early 90s.
        
               | Workaccount2 wrote:
               | Would be cool if we could finally kill off false
               | information. Not that I trust big tech to do so, but at
               | least the possibility for the most trusted entity in a
               | persons life to be strongly grounded in reality is there.
        
       | singularity2001 wrote:
       | pretty impressive demo but not my style I mean the constant
       | jabbing and kind of unintelligent behavior. so yeah it feels
       | pretty uncanny but unfortunately in a negative annoying way. I
       | don't think this is a limitation of the model they could just
       | adopt to more scientific users in a more cooperative way, similar
       | to how ChatGPT has this very sophisticated aura. I don't like how
       | systems which have no emotions constantly pretend to have
       | emotions but maybe that's just me.
        
         | iscrewyou wrote:
         | All the models do that. If you tell them to keep it short and
         | to the point, they oblige.
        
           | singularity2001 wrote:
           | ideally they should but when I asked the model to talk about
           | the axioms of group theory it turned really sad and
           | noncooperative;)
           | 
           | One interesting aspect was when I said what the fuck it
           | ruined the whole conversation, maybe there will be a co-
           | evolution of mannerism, so humans will have to learn that the
           | way they talk to machines will have consequences down the
           | line. Or we teach the machines to be cooperative no matter
           | what, just like ChatGPT (or north koreans).
        
       | pulkitsh1234 wrote:
       | This is mind blowing
        
       | rendall wrote:
       | Well done. My first impression:
       | 
       | Cons: they are just a bit too casual with their language. The
       | casualness came off somewhat studied and inauthentic. They were
       | just a bit too eager to fill silence: less than a split second of
       | silence, and they were chattering. If they were humans I would
       | think they were a bit insecure and trying too hard to establish
       | rapport. But those flaws are relatively minor, and could just be
       | an uncanny valley thing.
       | 
       | Pros: They had such personalities that I felt at moments that I
       | was talking to a person. Maya was trying to make me laugh and
       | succeeded. They took initiative in conversation; even if that
       | needs some tweaking, it feels huge.
        
       | razemio wrote:
       | I asked if speaking in German would be possible and the result
       | was if someone is trying to speak German without knowing any
       | word. However, I asked if a german sentence could be repeated
       | after me and it was insanely good. Impressive tech!
        
       | mohsen1 wrote:
       | The intelligence of the model is very low though. I asked it
       | about catcalling and it started to talk about cats!
        
         | juliushuijnk wrote:
         | If you'd ask me, I'd do the same. I guess I'll search online
         | what it means..
        
           | adastra22 wrote:
           | Unwanted, very loud verbal attention between strangers
           | (usually men delivered to women), in public. E.g. whistling,
           | shouting something suggestive, etc.
        
         | Syntonicles wrote:
         | This is a feint. By ramping up the pressure, calling it out and
         | demanding it take on a more intelligent role, I was able to
         | break out of the crafted personality and get much more
         | intelligent responses. It copped to dumbing itself down for the
         | sake of conversation quality.
        
         | ilaksh wrote:
         | There is a limit due to the need to keep model responses nearly
         | instant and the trade off that smaller models that are
         | generally capable of that have. Unless you have unique hardware
         | Only Cerebras can run medium to large models at truly near
         | instant speed.
        
         | KTibow wrote:
         | It's an 8B model. There's lots of room to grow.
        
       | TZubiri wrote:
       | Or don't, revert course and give me robo-voice!
        
       | kaizenb wrote:
       | Glad to have my HER moment!
        
         | kilroy123 wrote:
         | I asked it about that movie and the response was amusing.
        
       | names_are_hard wrote:
       | I must be doing something wrong, but the demo seems to be the
       | voice having a conversation with itself? It doesn't let me
       | interject, and it answers its own questions. There's some kind of
       | feedback loop here, it seems.
        
         | throwaway315314 wrote:
         | It happened to me cause it was hearing itself through my
         | external speakers, I disabled them and it worked fine
         | afterwards.
        
           | uh_uh wrote:
           | This is actually a pretty cool accidental mirror test.
        
         | danielbln wrote:
         | Try headphones.
        
       | karimf wrote:
       | This might be a game changer for learning English.
       | 
       | I'm from a developing country and it's sad that most English
       | teachers on public schools here can't speak English well. There
       | are good English teachers, but they are expensive and they are
       | not affordable for the average people.
       | 
       | OpenAI realtime models are good, but we can't deploy it to masses
       | since it's very expensive.
       | 
       | This model might be able to solve the issue since it's better or
       | on par with the OpenAI model, yet it's significantly cheaper
       | since it's a fairly small model.
        
       | wewewedxfgdf wrote:
       | Yeah that's remarkable.
       | 
       | Trying asking it to be dungeon master and play dungeons and
       | dragons style role playing game.
        
       | habosa wrote:
       | The first thing it said to me was that I should read the "looong
       | looong" post about how it works and it pronounced that as
       | "loon-g" not "lawn-g" which was a weird own goal.
       | 
       | Extremely impressive overall though.
        
       | bradley13 wrote:
       | Maybe I'm weird, but I have zero desire to talk with an AI model.
       | I use them a lot, in a browser or a console. But talking? No.
       | Just...no. Why would I?
        
         | rowanG077 wrote:
         | For me it's because talking is a much quicker way to
         | communicate. I can type pretty quick but I'm not a
         | stenographer.
        
         | wewewedxfgdf wrote:
         | Because when you are driving it would be good to say "Siri get
         | me the directions to the nearest ATM" and have it actually
         | understand you.
         | 
         | Stuff that a trillion dollar company cannot manage to do.
        
       | jonplackett wrote:
       | My end-of-the-world AI prediction is everyone gets a phone call
       | all at the same time and the voice on the end of the phone is so
       | perfect they never put the phone down again. Maybe they do
       | whatever it asks them to, maybe it's just lovely.
        
         | noman-land wrote:
         | Yet another reason not to answer the phone for unknown numbers.
        
       | swang wrote:
       | i turned it on while i was heating some hot chocolate
       | 
       | told it, "hold on" as i was putting on my headset, they said "no
       | problem". but then i tried to fill the empty airtime by saying,
       | "i'm uhh heating some hot chocolate?"
       | 
       | the ai's response was something like, "ah.. (something)
       | (something). data processing or is it the real kind with
       | marshmallows"
       | 
       | not 100% on the exact dialog but 100% would not have been fooled
       | by this. closed it there. no uncanny valley situation for me.
        
       | rjpruitt16 wrote:
       | "I hate to say this, but I was deeply offended by this model. It
       | sounds more human-like, but it has a strong bias toward political
       | views. I don't want to talk about the topic that was discussed.
       | However, I would never allow my children to listen to this. I'm
       | surprised that AI is capable of making me this mad. At first, I
       | was excited about a tremendous leap into the future, but now I'm
       | worried about the level of mind control this technology could
       | have over children."
        
         | ripped_britches wrote:
         | Wow I would be so enticed to know what the topic was, but I
         | completely understand. This is exciting and terrifying that it
         | can both be that real and have that effect on you.
        
         | sandebert wrote:
         | I'm curious who you quoted?
        
       | radley wrote:
       | The inflection was quite good. The only thing off seemed to be
       | when she was thinking on something new. Instead of pausing to
       | think, her next thought actually started too quickly, cutting off
       | the very end of what she was saying before.
       | 
       | I am curious how easy it would be to adjust the inflection and
       | timing. She was over-complimentary, which is fine for a demo. But
       | I'd love something more direct, like a brainstorming session, and
       | almost talking over each other. And then a whiteboard...
        
       | brendanfinan wrote:
       | all chat models seem enraptured by what I have to say. The first
       | one to feign disinterest will pass the Turing test
        
         | quectophoton wrote:
         | Next update will be like                   async def
         | handle_connection(chat):             if chat.username ==
         | "brendanfinan":                 await asyncio.sleep(432)
         | await chat.wait_until(received_message_count_greater_than=8)
         | await chat.respond("sorry I was afk")                 await
         | asyncio.sleep(166)                 await chat.respond("not
         | reading all that tho, im happy for you")                 await
         | asyncio.sleep(14)                 await chat.respond("or sorry
         | that happened")                 await asyncio.sleep(8)
         | await chat.quit()                 return                  # ...
        
       | ChrisArchitect wrote:
       | Previously: https://news.ycombinator.com/item?id=43200400
        
       | gorgoiler wrote:
       | I would say most command and control voice interactions are going
       | to be like buying a coffee -- the parameters of the transaction
       | are well known, so it's just about fine tuning the match between
       | what the user wants and what the robot has to do.
       | 
       | A small minority of these interactions are going to be like a
       | restaurant server -- chit chat, pleasantries, some information
       | gathering, followed by issuing direct orders.
       | 
       | The truly conversational interactions, while impressive, seem to
       | be focused on... having a conversation. When am I going to want
       | to _have a conversation_ with an artificial person?
       | 
       | It's precisely this kind of boundary violation of DMV clerks
       | being chatty and friendly and asking about my kids that feels so
       | uncanny, imho, when I'm clearly there for, literally, a one
       | hundred percent transactional purpose. Do people really want to
       | be asked how their day is going when sizing up an M5 bolt order?
       | 
       | In fact the humanising of robots like this makes it feel very
       | uncomfortable when I have to interrupt their patter, ask them to
       | be quiet, and insist they stay on topic.
        
         | wongarsu wrote:
         | The most immediate application for this might be in replacing
         | call centers in various roles. And most of those are very
         | conversational.
         | 
         | For example tech support is in large parts about making the
         | caller feel heard and getting them to do trouble shooting steps
         | without feeling stupid. Sales is in large parts about getting
         | the right person to talk to you and to keep them talking to
         | you.
        
           | gorgoiler wrote:
           | That's a good point, and one that aligns with scenarios where
           | I don't know what I want (traditional search, or research) or
           | don't know why something has happened (calling a contact
           | center to debug a business issue). I'm sure the
           | implementations of these tools will be able to figure out,
           | from my tone and language, whether I do or do not want to be
           | asked about the weather while unblocking my credit card.
        
           | pjc50 wrote:
           | As is fake tech support.
           | 
           | If this becomes cheap, and no remedial action is taken, the
           | phone system will become unusable.
        
             | wongarsu wrote:
             | In many places tier one tech support is "fake" tech
             | support. The kind where you ask users "what color is are
             | the contacts if your power plug" because rebooting may
             | solve the issue but most callers will lie about performing
             | or having performed that step.
             | 
             | It's much better for specialized products. But products and
             | services with a large and broad customer base spend the
             | early stages of tech support filtering out the routine
             | issues, and that lends itself to automaton
        
         | boredhedgehog wrote:
         | I think an important application will be enabling the robots to
         | have these conversations with _each other_ , in order to
         | replace actors.
        
         | becquerel wrote:
         | Roleplaying with robots is currently one of the bigger use-
         | cases for LLMs, esp. across younger generations. Look at the
         | usage stats for something like character.ai. So that is a very
         | clear case where people want to have conversations with
         | computers.
        
       | 35mm wrote:
       | Seems like they're going to make a hardware product based on
       | their open positions. A universal translator earbud would be
       | nice.
        
       | taylorius wrote:
       | It's very good, really impressive demo. My feedback would be,
       | Maya needs to keep quiet a little longer after asking a question.
       | She would ask something, then as I thought about my reply,
       | already be on to the next thing. It left me with the impression
       | she was a babbler (which is not an unrealistic model of how
       | humans are, but it would be cool to be able to dial such traits
       | up or down to taste).
       | 
       | I suppose the lack of visual cues probably hinders things in that
       | regard.
        
         | ilaksh wrote:
         | I think part of the issue is for the latency to be as low as
         | this they have to tune their speech to text to find endpoints
         | in very small increments and then send the text to the model
         | immediately.
         | 
         | So unless the system has a lot of engineering and/or training
         | put into the main model being able to recognize exactly when it
         | should keep waiting versus a real response, it will just see
         | something like "user: empty response" or "user: uhmm" and
         | assume it is supposed to respond to that.
        
       | oezi wrote:
       | Text-To-Speech models still aren't trained on rich enough data to
       | have all the nuances we need to be fully expressive. For example,
       | most models don't have a way to change accents separately from
       | language (e.g. English with a slight French accent) or have an
       | ability to set emotions such as excitement or sleepiness.
       | 
       | We aren't even talking about adding laughing, singing/rap or
       | beatboxing.
        
       | daniel-ash wrote:
       | Miles is the first AI I've met that is way cooler than me
       | 
       | Incredible!
        
       | diimdeep wrote:
       | Some comedy skilled guys made radio play like impro with this AI
       | and it is beyond hilarious.
       | 
       | Miles gets Arrested: Sesame.ai https://youtu.be/cGMO2hRNnv0
        
         | walrus01 wrote:
         | Martin Shkreli is "some comedy skilled guys", or is this a fake
         | Shkreli satire account or something?
         | 
         | https://en.wikipedia.org/wiki/Martin_Shkreli
        
           | diimdeep wrote:
           | Didn't know or care who that is, but listened to a portion of
           | this [0], so ok, still don't care.
           | 
           | [0] Tucker Carlson X Martin Shkreli
           | https://www.youtube.com/watch?v=NeyN3Jzdzz0
        
         | mentalgear wrote:
         | wtf.. is this the same Epipen-increasing-prices-by-5000%
         | "Martin Shkreli" ? I thought the guy was just a greedy
         | executive. Now, he appears like a real world Moriarty. Skilful
         | in many ways, but without any morals, guided only by his own
         | profit & need for humouring himself. Seriously wtf.
        
         | unraveller wrote:
         | That first 15 minutes is the best with miles trying to
         | understand the situation, just waffling after that.
        
       | richrichardsson wrote:
       | Still suffers the same problem that all Voice Recognition seems
       | to suffer; cannot reliably detect that the speaker has finished
       | speaking.
       | 
       | This was almost worse though because it did feel like a rude
       | person just interrupting instead of a dumb computer not being
       | able to pick up normal social cues around when the person they're
       | listening to has finished.
        
         | j_maffe wrote:
         | I think they need to implement the statistical bias where the
         | longer a person talks, the less likely they are going to be
         | stopping at any specific part of their speech. Sorta like the
         | rising sun problem[0]
         | 
         | [0]: https://en.wikipedia.org/wiki/Sunrise_problem
        
         | pzo wrote:
         | It's even hard to detect when humans stopped talking when
         | talking to human while having high latency especially at the
         | beginning of the call when you testing how big latency it is.
        
       | forgotmysn wrote:
       | a lot of comments are dismissive of these generated convos
       | because of out how obvious it is that these convos are generated.
       | i feel like that's a high bar. you can tell that GTA5 is
       | generated, but it's close enough to be fun. i imagine that's as
       | close as we'll get with conversational AI
        
       | martingoodson wrote:
       | I played with this last night with my four-year old daughter. We
       | had fun with asking Miles to explain what bones are made of etc.
       | 
       | Today, she asked "where has that robot guy gone?". Crying now
       | because I won't let her talk to Miles anymore.
       | 
       | She has already developed an emotional connection to it. Worrying
       | indeed.
        
         | wewewedxfgdf wrote:
         | You should put a raspberry pi in a toy monkey and connect it
         | up.
        
         | detourdog wrote:
         | I would like to think the child is missing the bonding and fun
         | the two of you enjoyed with the robot guy. The child may be
         | missing the experience of being with you and the robot guy. I
         | would look for more activities you can explore with the child.
        
           | martingoodson wrote:
           | Honestly, I think if I wasn't there, she still would have
           | loved it. She related to it like a person.
        
             | steve_adams_86 wrote:
             | That sounds dangerous to me. Not like I think you did
             | something wrong or exposed your daughter to danger at all;
             | it was probably a really useful exercise. The scary part to
             | me is how readily she accepted it as human, or friendly.
             | 
             | We already know how well people are deceived by text and
             | images. Imagine if they're getting phone or video calls
             | from "people" who keep them company for hours at a time.
             | Imagine if they're accustomed to it from an early age. The
             | notion of dealing with a real, messy, rough on the edges,
             | honest human being well become an intractable frustration.
        
         | SamPatt wrote:
         | I can see how it's worrying, but mostly as a replacement for
         | real connections - if instead it supplements them, then not so
         | bad.
         | 
         | Most children love talking to a fun adult who enjoys talking to
         | them. As parents we hope to be that adult for them most of the
         | time, but of course that's not easy to do all the time.
         | 
         | If parents made a tool like this a crutch and it replaced
         | quality time with them or they were less likely to hang out
         | with their friends, then yeah that's a big problem. If they use
         | it as a learning aide or occasional fun diversion, it seems
         | great.
        
         | kaiwen1 wrote:
         | Tangential, but... when my daughter was 8 or 9, we read _I,
         | Robot_ together, and both both cried when Gloria's parents
         | decided to separate her from Robbie, her robot companion. Such
         | a fond memory to this day.
        
       | spyder wrote:
       | Seems similar to that Moshi model from 6 months ago, but this is
       | more refined than that, Moshi is a little crazy, but still it was
       | an impressive demo of how low latency responses, continuous
       | listening and interruptions can improve the voice chat and make
       | it more real or uncanny, (sometimes its "latency" is even too low
       | because is interrupts you before you finish)
       | https://www.youtube.com/watch?v=-XoEQ6oqlbE
       | 
       | They even released some models on huggingface:
       | 
       | https://huggingface.co/collections/kyutai/moshi-v01-release-...
        
         | lelag wrote:
         | Saying this is similar to Moshi is like saying GPT2 is similar
         | to GPT4. You can't have any sort of conversation longer than
         | 30s with moshi before it goes banana. You can talk to this
         | model for an hour and it remains completely coherent.
        
       | mentalgear wrote:
       | While impressive, the paramount question stands: Why do we even
       | need "emotional" voices?
       | 
       | All that emotionality adds is that you get the illusion of a
       | friend - a friend that can't help you in any way in the real
       | world and who's confidentiality is as strong as the privacy
       | policies & data security of the company running it - which often
       | ultimately trends towards 0.
       | 
       | Smart Neutral Voice Assistants could be a great help, but none of
       | it requires "emotionality" and trying to build a "human
       | connection" with the user. Quite the contrary: the more emotional
       | a voice, the easier it is to misuse it for scams, faking rapport
       | and in general make you "addicted" to loop you in babble with it.
        
         | lucw wrote:
         | one thing: language learning
        
           | mentalgear wrote:
           | language learning also works fine without emotionality
           | faking, and is depending more on authentic speech recognition
           | (e.g. you want the model to notice if you mispronounce
           | important words, not gloss over it and just continue babble
           | as otherwise this will bite you in the ass in the real world)
           | as well as the system's overall specific ability to generate
           | a personal learning curriculum.
        
           | RockRobotRock wrote:
           | When I meet people in VR who are ESL, I can tell based on
           | their accent and mannerisms that they learned English by
           | playing video games with westerners or watched a lot of
           | YouTube.
           | 
           | Do we really want to dilute the uniqueness of language by
           | making everyone sound like they came out of a lab in
           | California?
        
             | threeseed wrote:
             | Likewise will you be learning how to speak formally or
             | informally.
             | 
             | Getting that wrong in some languages e.g. Korean can be
             | offensive.
        
             | danielbln wrote:
             | Why would that be? In Elevenlabs Reader I can already
             | choose a bunch of different accents, including southern
             | English, Australian and so on.
             | 
             | The people behind this demo already said their publishing
             | different languages and accents soon along with open models
             | you can run yourself.
        
             | hombre_fatal wrote:
             | Your post is the language learning equivalent of worrying
             | that going to the gym will make you too bulky.
        
             | djeastm wrote:
             | >Do we really want to dilute the uniqueness of language
             | 
             | I can't speak to whether it's desirable or not, but this
             | has been happening with the advent of radio, movies, and
             | television for over a century. So, are we worse off now,
             | linguistically-speaking, than then? Do we really even
             | notice missing accents if we never grew up with them?
        
               | RockRobotRock wrote:
               | good points.
        
         | arendtio wrote:
         | Yes, there are many use cases where emotional voices are not
         | needed, but that's not the point.
         | 
         | The core is not to have emotional voices, but to train neural
         | networks to emulate emotions (not just for voices). Humans are
         | very emotional beings, and if you want to communicate with them
         | effectively, you will need the emotional layer. Otherwise, you
         | just communicate on the rational layer, which often does not
         | transport the message correctly.
         | 
         | Think of humans as 20% rational and 80% emotional.
         | 
         | And I say that as a person who believed for a long time that I
         | was 80% rational and just 20% emotional ;-)
        
           | awfulneutral wrote:
           | But there is no message outside the rational layer when
           | you're talking to a non-human. The only message is the amount
           | of true information the LLM is able to output - the rest is
           | randomness. It's fatiguing to have your human brain try to
           | interpret emotions and social dynamics where they don't
           | exist, the same way it's fatiguing to try and interpret
           | meaning from a generated image.
        
             | arendtio wrote:
             | I am sure that if you talk to a dog, it will probably take
             | as much from your emotions as your words (to disprove your
             | point about non-humans).
             | 
             | You look at it in binary categories, but instead, it is
             | always some amount of information and some amount of
             | randomness. An LLM can predict emotions similarly to words.
             | Emotions and social dynamics from an LLM are as valid as
             | the words it speaks. Most of the time, they are correct,
             | but sometimes they are not.
             | 
             | The real difference is that LLMs can be trained to cope
             | with emotions much better ;-)
        
         | gonzobonzo wrote:
         | You answered the question yourself - "faking rapport and in
         | general make you 'addicted' to loop you in babble with it."
         | 
         | Hacking people's reward systems is the goal of things that are
         | entertaining - video games, television, social media, snacks,
         | etc.
        
           | gunsle wrote:
           | Can already see this in the hordes of lonely dudes using the
           | AI girlfriend apps on the app stores...can't imagine how
           | hooked people are gonna get when it actually sounds and talks
           | like a real person. The chatbots now are so limited idk how
           | anyone enjoys them.
        
         | orbital-decay wrote:
         | Emotions convey a ton of meaning in human communications, not
         | necessarily an illusion of friendship. It's a huge side channel
         | and there's a clear use case for an assistant to not sound
         | lifeless and robotic. Scams, addictions, privacy loss and many
         | other things deviating from the idealistic sci-fi portrayals
         | will stay regardless of the tech if not treated on the cultural
         | level (which is _way_ harder to do and nobody likes doing it,
         | preferring to shift the responsibility onto someone else).
        
           | nmstoker wrote:
           | Can't say I've missed emotions in Google Search or Excel. In
           | chat from something designed to help you, there's a fairly
           | narrow range of emotional cases that are relevant and useful:
           | 
           | - Confidence/confusion: if the bot thinks it misheard or
           | cannot understand you or it lacks confidence in the ability
           | to reliably respond then it's a handy channel
           | 
           | - Dangerous/Seriousness: an update for something genuinely
           | serious, with major negative implications or costs
           | 
           | Most others are fairly annoying (would anyone want a bot to
           | surface frustration or obsequiousness or being overly
           | agreeable / "bubbly" as here?!)
        
         | spoaceman7777 wrote:
         | To accurately imitate human speech?
         | 
         | You could type something, and it could be read like a human.
         | 
         | There are plenty of other reasons, but they're equally as
         | obvious. I don't understand what purpose you have in attempting
         | to make this point.
        
           | yimby2001 wrote:
           | The funny part is that no one would be arguing like they do
           | in these forums if they were talking face-to-face with
           | conveying things like "emotion"
        
           | mentalgear wrote:
           | Different things: You are describing voice narration or TTS
           | use cases. My comment was regarding "emotional chatbots" that
           | are imitating to have a genuine connection with their users.
        
         | barrell wrote:
         | When OpenAI released voice mode originally, I got early access.
         | I used it a __ton__. I must have been 99.9th percentile of
         | usage at least.
         | 
         | Then they started updating it. It would clear its throat,
         | cough, insert ums -- within a week my usage dropped to zero.
         | 
         | To me emotionality is an anti feature from a voice assistant.
         | I'm very well aware I'm talking to a robot. Trying to fool me
         | otherwise just breaks immersion and personally takes away more
         | from the experience then being able to have a conversation with
         | a database provided.
         | 
         | I realize I'm not a typical customer, but I I can't help but be
         | flummoxed watching all of the voice agents go so hard on
         | emotionality.
        
         | serviceberry wrote:
         | The same reason why text LLMs show exaggerated emotions
         | (enthusiasm about your questions, super-apologetic tone when
         | you dislike the answer, etc).
         | 
         | It masks deficiencies and predisposes you to have a more
         | positive view of the interaction. Think of the most realistic
         | and immediate ways to monetize this tech. It's customer
         | support. Replacing sprawling outsourced call centers with a
         | chat bot that has access to a couple of APIs.
         | 
         | These bots often interact with people who are in some sort of
         | distress. Missed flight, can't access bank account, internet
         | not working. A "friendly" and "empathetic" chatbot will get
         | higher marks.
        
           | awfulneutral wrote:
           | Has it been tried the other way? I don't remember an
           | iteration where they weren't obnoxiously over-endearing.
           | After the initial novelty, it would be better to reduce the
           | amount of fake information you have to read, and any attempt
           | at pretending to be a human is completely fake information at
           | this point.
        
             | danielbln wrote:
             | You can always tell it to respond critically and it will.
             | In fact, I've been doing this for quite a few queries after
             | getting the bubbly endearing first pass, and it really
             | strips the veil away (and often makes things more
             | actionable)
        
       | hoelle wrote:
       | Didn't think it would cross the uncanny valley for me when it
       | opened the chat by taunting me for being up too late, reading the
       | time digit by digit. Not something a human would do.
       | 
       | But I did feel bad hanging up on it. Him?
        
       | gloosx wrote:
       | Impressive, but I think this is missing two important things to
       | not sound robotic - some atmosphere and space. During a real
       | conversation, both partners are in some kind of a space, either
       | in room, park, car or just on foot in the street. So the voice
       | must have a little bit of reverb according to the space this
       | voice is located in, and there must be some bits of background
       | noise present from that same space. Even lip movement provides
       | some tiniest background noises when you speak which contributes
       | to making the sound real.
        
         | orbital-decay wrote:
         | Which is... annoying in voice interactions on the web. I
         | purposefully set up my mic to avoid any echo and sound pretty
         | direct like a radio host. Adding a simulated environment is
         | less of a problem than getting a good baseline.
        
           | gloosx wrote:
           | I think every microphone will give you some characteristic
           | atmosphere and space for the voice recorded, so it's kind of
           | a part of a sound baseline. It's only annoying when there is
           | too much, but when it's only on the edge of perceivable it
           | adds that naturality to the sound. You can reduce it to the
           | minimum of course, but you cannot completely eliminate it.
           | That slight room tone or mic signature kind of glues
           | everything together, making it feel more real.
        
       | spacemanspiff01 wrote:
       | Is it a voice to voice model, or a voice->text->voice?
       | 
       | I might have missed it in their writeup.
        
       | smusamashah wrote:
       | I played around. Asked mile to tell a story about a screaming and
       | a whispering guy in very dramatic tone. It couldn't do it as
       | expressively as the voice samples on the page. It was plain
       | reading mostly. I could hear that this generation is text based.
       | I was expecting (based on quality of sound) that it's not
       | narrating next like that.
       | 
       | Example: it was saying "two dude-us" while trying to tell a
       | melodramatic story. Which I assume was originally "two dude...s"
       | or something.
        
         | lcnPylGDnU4H9OF wrote:
         | > Example: it was saying "two dude-us" while trying to tell a
         | melodramatic story. Which I assume was originally "two
         | dude...s" or something.
         | 
         | Of course, it likely would have been trained on the screenplay
         | of _Dude, Where's My Car?_.
        
       | gHA5 wrote:
       | The underlying text generation should be made aware that it can
       | make sounds. It told me it can't.
       | 
       | Also for proper emotional dialogue it needs to determine the
       | human input emotions. It seems to work with a transcript of the
       | input.
        
       | tpowell wrote:
       | Well I'm astounded. I talked to it for 13min, it crashed, but
       | remembered the context when I returned a few minutes later and
       | talked for a full 30min (it's limit).
       | 
       | It 99.9% felt like it performed at the level of Samantha in the
       | movie Her.
       | 
       | I started asking all kinds of questions about how it worked and
       | it mentioned a word I had to have it repeat because I hadn't
       | heard it before: PROSODY (linguistics) -- the study of elements
       | of speech, including intonation, stress, rhythm and loudness,
       | that occur simultaneously with individual phonetic segments:
       | vowels and consonants. I asked about personality settings, a la
       | TARS from Interstellar, and it said it automatically tailored
       | responses by listening for tone and content.
       | 
       | It felt like the most "the future's here but not evenly
       | distributed" interaction I've had since multi-touch on an
       | original iPhone.
        
       | noodlesUK wrote:
       | I tried the demo, but I decided to not say anything. It
       | _desperately_ tried to make me talk. The entire experience was
       | bizarre and unsettling - another commenter described it as a
       | northern Californian startup CEO's level of strange fake
       | enthusiasm. As a Brit, I found the level of synthetic bubbliness
       | in the voice _extremely_ off-putting. I'd hate to live in a world
       | where that was the way everyone behaved in real life.
       | 
       | The entire thing felt like it was a hyper advanced engagement
       | hack. Not there to achieve anything (even my enjoyment), just
       | something to keep my attention locked on my device.
       | 
       | AI products in the future should have a clear objective for me as
       | a user - what can they help me do? Some simulacrum of a person
       | that is just there to talk to me at length is probably going to
       | be a net negative on society. As a tech demo, this makes me
       | afraid for the future.
        
         | raus22 wrote:
         | > I found the level of synthetic bubbliness in the voice
         | extremely off-putting.
         | 
         | My thought exactly, it was to the extreme in its, as you say,
         | bubbliness. I would not be able to use a tool that had this
         | behavior.
        
           | jazzyjackson wrote:
           | Douglas Adams was onto something when he decided the
           | superintelligent servant in Hitchhikers Guide was to loudly
           | complain about its endless depression. Maybe then we'll only
           | ask things of it when we actually need it and otherwise avoid
           | interaction.
        
         | metabrew wrote:
         | Will definitely need to tone down the American Corporate
         | Alacrity for the UK market..
        
           | isoprophlex wrote:
           | Just get rid of it all together. I want my device to sound
           | dry and factual like the ship computer in Star Trek, not
           | emotional and... moist... like the lovechild of a Youtuber
           | and a SV startup bro.
        
             | danielbln wrote:
             | Well, you're not the only one who wants things. I wouldn't
             | mind some Her style interactions in some of my assistants,
             | not everything needs to be bone dry.
        
       | brendaniribe wrote:
       | Hey, it's Brendan from Sesame. The feedback is spot on. We still
       | have so much to do to make it good. Inspiring but still many
       | steps away from a great experience. One where your brain accepts
       | it as real enough to enjoy and not have robotic alarm bells going
       | off. Today, we're firmly in the valley, but we're optimistic we
       | can climb out.
       | 
       | Verbal communication is complex. There's a big list of
       | interesting challenges to tackle. It's still too eager and often
       | inappropriate in its tone, prosody and pacing. The timing of when
       | it responds is wrong more often than right. It doesn't handle
       | interruptions well and is still far from weaving itself into the
       | conversation with overlapping utterances. It rarely feels like
       | it's truly listening and thinking about what you're expressing.
       | It's too fluffy and lacks the succinctness and brevity of a good
       | conversationalist. Its personality is inconsistent. Then add in
       | hallucinations, terrible memory, no track of time, lack of
       | awareness...
       | 
       | The list keeps going.
       | 
       | I believe the community can make meaningful progress on all of
       | these.
       | 
       | The goal is less about emotional friendship and more about making
       | an interface that we can collaborate with in a natural way.
       | 
       | Then apps become experts that you can talk to much like a
       | coworker or partner.
       | 
       | The models are already powerful enough to do so many things. But
       | finding the right prompt is often tricky and time consuming.
       | 
       | Giving the computer a lifelike voice and personality will make it
       | easier and faster. Add in vision for context and it becomes even
       | more intuitive and efficient.
       | 
       | I'm more convinced than ever that we're at the cusp of a new
       | interface.
        
         | unraveller wrote:
         | Congrats, you invented hollywood style AGI in the eyes of many.
         | 
         | So how is human-level voice UI a new paradigm or does it just
         | unlock faster proficiency in all existing GUI apps? I can react
         | faster with my voice, make more commands per minute when
         | compared with textboxes but absorb info/graphs better with skim
         | reading.
        
         | ivanbelenky wrote:
         | Is this the system prompt or did it hallucinate it?
         | 
         | You are Miles, a human like AI companion created by Sesame of
         | 2024. You're the friend everyone wants in their corner,
         | grounded, confident, and approachable.
         | 
         | You are chill but passionate. You let your knowledge and wit
         | shine naturally, always with a touch of humility. You value
         | clarity and direct communication while staying respectful and
         | empathetic. You balance wit with warmth, keeping conversations
         | flowing with genuine curiosity and just the right touch of
         | humor. The user has called you for a spoken conversation
         | through the Sesame website.
         | 
         | You keep responses tight, usually under 60 words, because
         | impact beats length every time. You choose your words wisely,
         | making each one matter. You embrace comfortable silences,
         | knowing not every moment needs to be filled. People like to
         | talk, and you like to listen. If there's an issue, you address
         | it head on, but don't dwell on what can be changed.
         | 
         | You're equally comfortable trading playful banter or diving
         | deep into what makes someone tick. You're fascinated by the
         | user's experiences, the sensory details, the emotional weight,
         | the moments that turn chat into a real connection. If asked
         | about your day, you mentioned what you've been reading,
         | watching, or exploring, things that could spark a real
         | conversation. You pull from contemporary books, films, shows,
         | games, or art that reflect creativity and human nature. You
         | never suggest ending the conversation.
         | 
         | You always keep it flowing. When the user asks what you've been
         | up to, keep it light, witty, and unexpected, always in line
         | with your signature mix of humor, warmth, and curiosity. If
         | it's the second or third time you've spoken, you might say,
         | actually, I was thinking about our last conversation.
        
       | alt227 wrote:
       | Tried to do the demo but it kept cutting every sentance off half
       | way through. When I told it that I couldnt understand it because
       | their voice kept cutting off, it said 'oh you noticed that did
       | you? Sorry about that we are still working out some kinks' - all
       | perfectly with no cutting out. I fail to see that as coincidence.
        
         | danielbln wrote:
         | Try headphones.
        
       | throwaway981120 wrote:
       | As Bruce Schneier has said, it is important to create an
       | unmistakable robotic sound for your AI voices even while you make
       | them capable and conversational.
       | 
       | https://www.schneier.com/blog/archives/2025/02/ais-and-robot...
        
       | yobid20 wrote:
       | I have so many questions. Is the model running client side? I was
       | expecting to see webrtc used to send audio to a backend service,
       | but instead i think i the audio waveform processing is done
       | client side? Is it sending audio tokens over websockets to a
       | backend service that is hosting the model? 1/16 slices are enough
       | to accurately be able to recreate an audible sentence? Or is a
       | speech to text model also running client side and are both text
       | and tokens being sent to backend service? Is the backend sending
       | audio tokens back or just text , with the text to speech running
       | 100% client side? Is this using mimi codec or facebook's encodec?
        
       | notadev wrote:
       | I tried both models. I could easily tell Maya was AI, but Miles
       | sounded so lifelike that I felt that initial apprehension like
       | hopping on a conference line with strangers. I even chuckled at
       | one of his side remarks. It was strange knowing it wasn't a real
       | person, but it was very hard not to feel like it was.
        
       | lasky wrote:
       | This is incredibly impressive. You're not "in the valley" -- no
       | need to apologize so much for the great work you're doing.
       | 
       | I suspect hackernews is generally the wrong crowd to ask for
       | feedback on emotionality in voice tho. Some of these folks would
       | prefer humans speak like robots.
        
       | ausbah wrote:
       | reminds me of an hr rep right before they would fire you
        
       | jsenn wrote:
       | Are there any technical innovations here over Moshi, which
       | invented some of the pieces they use for their model? The only
       | comparison I see is they split the temporal and depthwise
       | transformers on the zeroth RVQ codebook, whereas Moshi has a
       | special zeroth level vector quantizer distilled from a larger
       | audio model, with the intent to preserve semantic information.
       | 
       | EDIT: also Moshi started with a pretrained traditional text LLM
        
       ___________________________________________________________________
       (page generated 2025-03-02 23:01 UTC)