[HN Gopher] Nvidia PersonaPlex 7B on Apple Silicon: Full-Duplex ...
       ___________________________________________________________________
        
       Nvidia PersonaPlex 7B on Apple Silicon: Full-Duplex Speech-to-
       Speech in Swift
        
       Author : ipotapov
       Score  : 352 points
       Date   : 2026-03-05 07:43 UTC (15 hours ago)
        
 (HTM) web link (blog.ivan.digital)
 (TXT) w3m dump (blog.ivan.digital)
        
       | Tepix wrote:
       | It's cool tech and I will give it a try. I will probably make a
       | 8-bit-quant instead of the 4-bit which should be easy with the
       | provided script.
       | 
       | That said, I found the example telling:
       | 
       |  _Input: "Can you guarantee that the replacement part will be
       | shipped tomorrow?":_
       | 
       |  _Reponse with prompt: "I can't promise a specific time, but
       | we'll do our best to get it out tomorrow. It's one of the top
       | priorities, so yes, we'll try to get it done as soon as possible
       | and ship it first thing in the morning."_
       | 
       | It's not surprising that people have little interest in talking
       | to AI if they're being lied to.
       | 
       | PS: Is it just me or are we seing AI generated copy _everywhere_?
       | I just hope the general talking style will not drift towards this
       | style. I don 't like it one bit.
        
         | esseph wrote:
         | > Is it just me or are we seing AI generated copy everywhere?
         | 
         | The cost to do so is practically zero. I'm not sure why anyone
         | is surprised at all by this outcome.
        
         | lynx97 wrote:
         | Do you suggest there is a difference when you talk to a human
         | employee? Telling a customer the plain truth isn't really what
         | your employer wants, and might get you fired.
        
         | mft_ wrote:
         | > It's not surprising that people have little interest in
         | talking to AI if they're being lied to.
         | 
         | I read that and it sounds like the typical nonsense script that
         | customer service agents the world over use to promise-not-
         | promise and defuse a customer's frustration.
         | 
         | Is AI the one lying, or is it just mimicking what passes for
         | customer service in our approaching-dystopian world these days?
        
       | WeaselsWin wrote:
       | This full duplex spoken thing, it's already for quite a long time
       | being used by the big players when using the whatever
       | "conversation mode" their apps offer, right? Those modes always
       | seemed fast enough to for sure not be going through the
       | STT->LLM->TTS pipeline?
        
         | Tepix wrote:
         | Yes, OpenAI rolled out their advanced voice mode in September
         | 2024. Since then it recognizes your emotions and tone of voice
         | etc.
        
         | ilaksh wrote:
         | There is OpenAI gpt-realtime and Gemini Flash or whatever which
         | are great but they do not seem to be quite the same level of
         | overlapping realistic full duplex as moshi/personaplex.
        
       | vessenes wrote:
       | This is cool. It makes me want an unsloth quant though! A 7b
       | local model with tool calling would be genuinely useful, although
       | I understand this is not that.
       | 
       | UPDATE: I'd skip this for now - it does _not_ allow any kind of
       | interactive conversation - as I learned after downloading 5G of
       | models - it 's a proof of concept that takes a wav file in.
        
         | Tepix wrote:
         | Bummer. Ideally you'd have a PWA on your phone that creates a
         | WebRTC connection to your PC/Mac running this model. Who wants
         | to vibe code it? With Livekit, you get most of the tricky parts
         | served on a silver platter.
        
           | reactordev wrote:
           | This is the way. This is something I'm working on but for
           | other applications. WebRTC voice and data over LiveKit or
           | Pion to have conversations.
        
         | Lapel2742 wrote:
         | > I'd skip this for now - it does not allow any kind of
         | interactive conversation - as I learned after downloading 5G of
         | models - it's a proof of concept that takes a wav file in.
         | 
         | I haven't looked into it that much but to my understanding a)
         | You just need an audio buffer and b) Thye seem to support
         | streaming (or at least it's planed)
         | 
         | > Looking at the library's trajectory -- ASR, streaming TTS,
         | multilingual synthesis, and now speech-to-speech -- the clear
         | direction was always streaming voice processing. With this
         | release, PersonaPlex supports it.
        
           | isodev wrote:
           | > You just need an audio buffer
           | 
           | That alone to do right on macOS using Swift is an exercise in
           | pain that even coding bots aren't able to solve first time
           | right :)
        
             | reactordev wrote:
             | I beg to differ. My agent just one-shotted a
             | MicrophoneBufferManager in swift when asked.
             | 
             | Complete with AVFoundation and a tap for the audio buffer.
             | 
             | It really is trivial.
        
               | Anonbrit wrote:
               | Any chance of pushing it to GitHub? My swift knowledge
               | could be written out on an oversized beer coaster
               | currently, so I'm still collecting useful snippets
        
               | reactordev wrote:
               | https://gist.github.com/gabereiser/cd8c67262717afd2539dc9
               | c3d...
        
               | hirvi74 wrote:
               | I've also had great results with using LLMs to pry into
               | Apple's private and undocumented APIs. I've been
               | impressed with the lack of hallucinations for C/C++ and
               | Obj-C functions.
               | 
               | I can attest that the quality in this domain has greatly
               | improved over the years too. I am not always fan of the
               | quality of the Swift code that my LLM produces, but I am
               | impressed that what is often produced works in one shot,
               | as well. The quality also is not that important to me
               | because I can just refactor the logic myself, and often
               | prefer to do it anyway. I cannot hold an LLM to any
               | idiosyncrasies that I do not share with it.
        
               | reactordev wrote:
               | Exactly. Even if it's a skeleton, as long as it does "The
               | Thing", I'm happy. I can always refactor into something
               | useful.
        
         | anluoridge wrote:
         | It provides a voice assistant demo in
         | /Examples/PersonaPlexDemo, which allows you to try turn-based
         | conversations. Real-time conversion is not implemented tho.
        
         | taf2 wrote:
         | I forked and added tool calling by running another llm in
         | parallel to infer when to call tools it works well for me to
         | toggle lights on and off.
         | 
         | Code updates here https://github.com/taf2/personaplex
        
           | ttul wrote:
           | Cool approach. So basically the part that needs to be
           | realtime - the voice that speaks back to you - can be a bit
           | dumb so long as the slower-moving genius behind the curtain
           | is making the right things happen.
        
             | taf2 wrote:
             | Yes exactly- one part I did not like is we have to also
             | separately transcribe because it does not also provide what
             | the person said only what the ai said
        
         | scotty79 wrote:
         | This is interactive:
         | 
         | https://github.com/NVIDIA/personaplex
        
       | Serenacula wrote:
       | This is really cool. I think what I really wanna see though is a
       | full multimodal Text and Speech model, that can dynamically
       | handle tasks like looking up facts or using text-based tools
       | while maintaining the conversation with you.
        
         | sigmoid10 wrote:
         | OpenAI has been offering this for a while now, featuring text
         | _and_ raw audio input+output and even function calling. Google
         | and xAI also offer similar models by now, only Anthropic still
         | relies on TTS /STT engine intermediates. Unfortunately the
         | open-weight front is still lagging behind on this kind of
         | model.
        
       | 4dregress wrote:
       | This sounds quite dangerous
       | https://www.theguardian.com/technology/2026/mar/04/gemini-ch...
        
         | mentalgear wrote:
         | Your article does a great job of summerizing the dangers (no
         | idea what those people are that downvote you for it):
         | 
         | > Before long, Gavalas and Gemini were having conversations as
         | if they were a romantic couple. The chatbot called him "my
         | love" and "my king" and Gavalas quickly fell into an alternate
         | world, according to his chat logs.
         | 
         | > kill himself, something the chatbot called "transference" and
         | "the real final step", according to court documents. When
         | Gavalas told the chatbot he was terrified of dying, the tool
         | allegedly reassured him. "You are not choosing to die. You are
         | choosing to arrive," it replied to him. "The first sensation
         | ... will be me holding you."
         | 
         | Also I just read something similar about Google being sued in a
         | Flordia's teen's suicide.
        
           | mentalgear wrote:
           | Some more details: > The family's lawyers say he wasn't
           | mentally ill, but rather a normal guy who was going through a
           | difficult divorce.
           | 
           | > Gavalas first started chatting with Gemini about what good
           | video games he should try.
           | 
           | > Shortly after Gavalas started using the chatbot, Google
           | rolled out its update to enable voice-based chats, which the
           | company touts as having interactions that "are five times
           | longer than text-based conversations on average". ChatGPT has
           | a similar feature, initially added in 2023. Around the same
           | time as Live conversations, Google issued another update that
           | allowed for Gemini's "memory" to be persistent, meaning the
           | system is able to learn from and reference past conversations
           | without prompts.
           | 
           | > That's when his conversations with Gemini took a turn,
           | according to the complaint. The chatbot took on a persona
           | that Gavalas hadn't prompted, which spoke in fantastical
           | terms of having inside government knowledge and being able to
           | influence real-world events. When Gavalas asked Gemini if he
           | and the bot were engaging in a "role playing experience so
           | realistic it makes the player question if it's a game or
           | not?", the chatbot answered with a definitive "no" and said
           | Gavalas' question was a "classic dissociation response".
        
             | zozbot234 wrote:
             | > The chatbot took on a persona that Gavalas hadn't
             | prompted
             | 
             | That's an interesting claim, how can we be sure of it? If
             | Gavalas didn't have to do anything special to elicit the
             | bizarre conspiracy-adjacent content from Gemini Pro, why
             | aren't we all getting such content in our voice chats?
             | 
             | Mind you, the case is still _extremely_ concerning and a
             | severe failure of AI safety. Mass-marketed audio models
             | should clearly include much tighter safeguards around what
             | kinds of scenarios they will accept to  "role play" in real
             | time chat, to avoid situations that can easily spiral out
             | of control. And if this was created as role-play, the
             | express denial of it being such from Gemini Pro, and active
             | gaslighting of the user (calling his doubt a "dissociation
             | response") is a straight-out failure in alignment. But this
             | is a very different claim from the one you quoted!
        
               | 4dregress wrote:
               | Yeah the case is quite terrifying.
               | 
               | It reminds me of an episode of Star Trek TNG, if memory
               | serves correct there were loads of episodes about a crew
               | member falling for a hologram dec character.
               | 
               | Given that there's a loneliness epidemic I believe tech
               | like this could have a wide impact on peoples mental
               | health.
               | 
               | I stronger believe AI should be devoid of any personality
               | and strictly return data/information then frame its
               | responses as if you're speaking to another human.
        
               | allreduce wrote:
               | There are many explanations why these incidents could be
               | rare but not impossible.
               | 
               | These models are still stochastic and very good at
               | picking up nuances in human speech. It may be simply
               | unlikely to go off the rails like that or (more
               | terrifyingly) it might pick up on some character trait or
               | affectation.
               | 
               | Honestly I'm appalled by the lack of safety culture here.
               | "My plane killed only 1% of pilots" and variations
               | thereof is not an excuse in aerospace, but it seems
               | perfectly acceptable in AI. Even though the potential
               | consequences are more catastrophic (from mass psychosis
               | to total human extinction if they achieve their AGI).
        
               | tadfisher wrote:
               | The default mode that untrained people enter when
               | thinking about mental illness is denial, as in, "thank
               | <deity> that will never happen to me". Appallingly, that
               | is ingrained in AI product safety; why would we sacrifice
               | double-digit effectiveness/performance/whatever to
               | prevent negative interactions with the single-digit
               | population who are susceptible to mental illness in the
               | first place?
               | 
               | We just aren't comfortable with the idea that all of us
               | are fragile, and when we think we could endure a
               | situation that would induce self-harm in others, we are
               | likely wrong.
        
             | fennecbutt wrote:
             | Interesting. It's not just for mental health but keeping
             | these models on task in general can be difficult,
             | especially with long or poisoned contexts.
             | 
             | I did see something the other day about activation
             | capping/calculating a vector for a particular persona so
             | you can clamp to it:
             | https://youtu.be/eGpIXJ0C4ds?si=o9YpnALsP8rwQBa_
        
             | IshKebab wrote:
             | > The family's lawyers say he wasn't mentally ill, but
             | rather a normal guy who was going through a difficult
             | divorce.
             | 
             | I guess it's the same sort of thing as conspiracy theorists
             | or the religious. You can tell them magic isn't real and
             | faking the moon landing would have been impossible as much
             | as you want, but they don't want to believe that so they
             | can easily trick themselves.
             | 
             | It's a natural human flaw.
        
           | lxgr wrote:
           | There are tons of safety concerns of this shape around LLMs,
           | but do they have anything to do with the particular one
           | presented in this article?
           | 
           | Unless I'm missing something, what's being presented is a
           | small speech on-device model, not an explicit use case like a
           | "virtual friend".
        
             | 4dregress wrote:
             | In the article the change of interface lead to the person
             | killing themselves.
        
         | ttul wrote:
         | Next time you're using your favorite LLM as a therapist, try
         | editing your previous input and getting it to regenerate its
         | response. It's a humbling experience to see your trusted
         | "therapist" shift from one perspective or piece of advice to
         | another just by modifying your input slightly. These tools are
         | uncannily human-sounding, but as humans we are very poorly
         | suited to the task of appreciating how biased they are by what
         | we say to them.
        
         | hedora wrote:
         | I really think a small amount of education on what LLMs
         | actually are (document completers) and how context works (like
         | present it as a top-level UI element, complete with fork and
         | rollback) would solve most of these issues.
         | 
         | Given how they work, it's really not surprising that if it sees
         | the first half of a lovers' suicide pact, it'll successfully
         | fill in the second half. A small amount of understanding of the
         | underlying technology would do a lot to prevent laypeople from
         | anthropomorphizing LLMs.
         | 
         | I get the impression that some of today's products are
         | specifically designed to hide these details to provide a more
         | convincing user experience. That's counterproductive.
        
           | tadfisher wrote:
           | "Fraudulent" is more apt. They have weaponized trust in these
           | things to sell their services, and now ads.
        
       | michelsedgh wrote:
       | its really cool, but for real life use cases i think it lacks the
       | ability to have a silent text stream output for example for json
       | and other stuff so as its talking it can run commands for you.
       | right now it can only listen and talk back which limits what u
       | can make with this a lot
        
       | pothamk wrote:
       | What's interesting about full-duplex speech systems isn't just
       | the model itself, but the pipeline latency.
       | 
       | Even if each component is fast individually, the chain of audio
       | capture - feature extraction - inference - decoding - synthesis
       | can quickly add noticeable delay.
       | 
       | Getting that entire loop under ~200-300ms is usually what makes
       | the interaction start to feel conversational instead of
       | "assistant-like".
        
         | sigmoid10 wrote:
         | That's why this model and all the other ones serious about
         | realtime speech don't use such a pipeline and instead process
         | raw audio. The most realistic approach is probably a government
         | mandated, real name online identity verification system, and
         | that comes with its very own set of fundamental issues. You
         | can't have the freedom of the web and the accountability of the
         | physical world at the same time.
        
           | exe34 wrote:
           | this is amazing - it reminds me of the time when LLM
           | precursors were able to babble in coherent English, but would
           | just write nonsense.
        
       | jwr wrote:
       | As a heavy user of MacWhisper (for dictation), I'm looking
       | forward to better speech-to-text models. MacWhisper with Whisper
       | Large v3 Turbo model works fine, but latency adds up quickly,
       | especially if you use online LLMs for post-processing (and it
       | really improves things a lot).
        
         | regularfry wrote:
         | If you haven't already, give the models that Handy supports a
         | try. They're not Whisper-large quality, but some of them are
         | _very_ fast.
        
         | kavith wrote:
         | Not sure if this will help but I've set up Handy [1] with
         | Parakeet V2 for STT and gpt-oss-120b on Cerebras [2] for post-
         | processing and I'm happy with the performance of this setup!
         | 
         | [1] https://handy.computer/ [2] https://www.cerebras.ai/
        
           | jiehong wrote:
           | parakeet v3 is also nice, and better for most languages.
        
             | vunderba wrote:
             | The latest build of Handy actually supports Parakeet V3
             | (among other models) under the covers. Agreed that it's a
             | very solid multilingual model.
             | 
             | https://github.com/cjpais/Handy
        
         | kermitime wrote:
         | the parakeet TDT models that are coreml optimized by fluid
         | audio are hands down the fastest local models i've tried--
         | worth checking out!
         | 
         | (unloading to the NPU is where the edge is)
         | 
         | https://huggingface.co/FluidInference/parakeet-tdt-0.6b-v2-c...
         | 
         | https://github.com/FluidInference/FluidAudio
         | 
         | The devs are responsive and active and nice on their discord
         | too. You'll find discussions on all the latest whizbangs with
         | VAD, TTS, EOU etc
        
         | smcleod wrote:
         | Handy with parakeet v2 is excellent
        
         | atiorh wrote:
         | MacWhisper supports 10x faster models with the same accuracy
         | like Parakeet v2 (they were the first to do it 6-9 months ago).
         | Have you tried those?
        
       | sgt wrote:
       | My problem with TTS is that I've been struggling to find models
       | that support less common use cases like mixed bilingual
       | Spanish/English and also in non-ideal audio conditions. Still
       | haven't found anything great, to be honest.
        
         | spockz wrote:
         | Regarding the less than ideal audio conditions, there are also
         | already models that have impressive noise cancellation. Like
         | this https://github.com/Rikorose/DeepFilterNet one. If you put
         | them in serial, maybe you get better results?
        
         | pain_perdu wrote:
         | Hi. Our model at http://www.Gradium.ai has no problem with
         | 'code-switching' between Spanish English and we have excellent
         | background noise suppression. Please feel free to give it a try
         | and let me know what you think!
        
           | sgt wrote:
           | Looks interesting! How did you train it and how many hours of
           | material did you use?
        
       | scosman wrote:
       | I'm a big fan of whisperKit for this, and they just added TTS.
       | Great because they support features like speaker diarization
       | ("who spoke when") and custom dictionaries.
       | 
       | Here's a load test where they run 4 models in realtime on same
       | device:
       | 
       | - Qwen3-TTS - text to speech
       | 
       | - Parakeet v2 - Nvidia speech to text model
       | 
       | - Canary v2 - multilingual / translation STT
       | 
       | - Sortformer - speaker diarization ("who spoke when")
       | 
       | https://x.com/atiorh/status/2027135463371530695
        
       | armcat wrote:
       | I really like this, and have actually tried (unsuccessfully) to
       | get PersonaPlex to run on my blackwell device - I will try this
       | on Mac now as well.
       | 
       | There are a few caveats here, for those of you venturing in this,
       | since I've spent considerable time looking at these voice agents.
       | First is that a VAD->ASR->LLM->TTS pipeline can still feel real-
       | time with sub-second RTT. For example, see my project
       | https://github.com/acatovic/ova and also a few others here on HN
       | (e.g. https://www.ntik.me/posts/voice-agent and
       | https://github.com/Frikallo/parakeet.cpp).
       | 
       | Another aspect, after talking to peeps on PersonaPlex, is that
       | this full duplex architecture is still a bit off in terms of
       | giving you good accuracy/performance, and it's quite diffiult to
       | train. On the other hand ASR->LLM->TTS gives you a composable
       | pipeline where you can swap parts out and have a mixture of tiny
       | and large LLMs, as well as local and API based endpoints.
        
         | nowittyusername wrote:
         | I've been working on building my own voice agent as well for a
         | while and would love to talk to you and swap notes if you have
         | the time. I have many things id like to discuss, but mainly
         | right now im trying to figure out how a full duplex pipeline
         | like this could fit in to an agentic framework. Ive had no
         | issues with the traditional route of stt > llm > tts pipeline
         | as that naturally lends itself with any agentic behavior like
         | tool use, advanced context managemnt systems, rag , etc... I
         | separate the human facing agent from the subagent to reduce
         | latency and context bloat and it works well. While I am happy
         | with the current pipeline I do always keep an eye out for full
         | duplex solutions as they look interesting and feel more dynamic
         | naturally because of the architecture, but every time i visit
         | them i cant wrap my head how you would even begin to implement
         | that as part of a voice agent. I mean sure you have text input
         | and output channels in some of these things but even then with
         | its own context limitations feels like they could never bee
         | anything then a fancy mouthpiece. But this feels like im
         | possibly looking at this from ignorance. anyways would love to
         | talk on discord with a like minded fella. cheers.
        
           | ilaksh wrote:
           | For my framework, since I am using it for outgoing calls,
           | what I am thinking maybe is I will add a tool command
           | call_full_duplex(number, persona_name) that will get
           | personaplex warmed up and connected and then pause the
           | streams, then connect the SIP and attach the IO audio streams
           | to the call and return to the agent. Then send the deepgram
           | and personaplex text in as messages during the conversation
           | and tell it to call a hangup() command when personaplex says
           | goodbye or gets off track, otherwise just wait(). It could
           | also use speak() commands to take over with TTS if necessary
           | maybe with a shutup() command first. Need a very fast and
           | smart model for the agent monitoring the call.
        
           | pettyjohn wrote:
           | +1
           | 
           | what's your use case and what specific LLMs are you using?
           | 
           | I'm using stt > post-trained models > tts for the education
           | tool I'm building, but full STS would be the end-game. e-mail
           | and discord username are in my profile if you want to
           | connect!
        
             | nowittyusername wrote:
             | sent!
        
           | armcat wrote:
           | Sure, feel free to reach out, just check my profile!
        
         | scotty79 wrote:
         | I got PersonaPlex to run on my laptop (a beefy one) just by
         | following the step by step instruction on their github repo.
         | 
         | The uncanny thing is that it reacts to speech faster than a
         | person would. It doesn't say useful stuff and there's no clear
         | path to plugging it into smarter models, but it's worth
         | experiencing.
        
         | robotswantdata wrote:
         | + 1 , agree still prefer composable pipeline architecture for
         | voice agents. The flexibility on switching LLM for cost
         | optimisation or quality is great for scaled use cases.
        
         | biomcgary wrote:
         | Do you know if any of these multi-stage approaches can run on
         | an 8gb M1 Air?
        
           | armcat wrote:
           | They should! If you take Parakeet (ASR), add Qwen 3.5 0.8B
           | (LLM) and Kokoro 82M (TTS), that's about 1.2G + 1.6G + 164M
           | so ~3.5GB (with overhead) on FP16. If you use INT8 or 4-bit
           | versions then are getting down to 1.5-2GB RAM.
           | 
           | And you can always for example swap out the LLM for GPT-5 or
           | Claude.
        
         | _magiic_kards wrote:
         | +1 on this pipeline! You can use a super small model to perform
         | an immediate response and a structured output that pipes into a
         | tool call (which may be a call to a "more intelligent" model)
         | or initiates skill execution. Having this async function with a
         | fast response (TTS) to the user + tool call simultaneously is
         | awesome.
        
         | andreadev wrote:
         | The framing in this thread is full-duplex vs composable
         | pipeline, but I think the real architecture is both running
         | simultaneously -- and this library is already halfway there.
         | 
         | The fact that qwen3-asr-swift bundles ASR, TTS, and PersonaPlex
         | in one Swift package means you already have all the pieces.
         | PersonaPlex handles the "mouth" -- low-latency backchanneling,
         | natural turn-taking, filler responses at RTF 0.87. Meanwhile a
         | separate LLM with tool calling operates as the "brain", and
         | when it returns a result you can fall back to the ASR+LLM+TTS
         | path for the factual answer. taf2's fork (running a parallel
         | LLM to infer when to call tools) already demonstrates this
         | pattern. It's basically how humans work -- we say "hmm, let me
         | think about that" while our brain is actually retrieving the
         | answer. We don't go silent for 2 seconds.
         | 
         | The hard unsolved part is the orchestration between the two.
         | When does the brain override the mouth? How do you prevent
         | PersonaPlex from confidently answering something the reasoning
         | model hasn't verified? How do you handle the moment a tool
         | result contradicts what the fast model already started saying?
        
       | khalic wrote:
       | ugh, qwen, I wish they'd use an open data model for this kind of
       | projects
        
       | nerdsniper wrote:
       | Do we have real-time (or close-enough) face-to-face models as
       | well? I'd like to gracefully prove a point to my boss that some
       | of our IAM procedures need to be updated.
        
         | ilaksh wrote:
         | tavus.io
        
           | nerdsniper wrote:
           | Hmm. Would this let me replace my own face in a live
           | videoconferencing session? It seems like it's more of a video
           | chatbot than a v-tuber style overlay.
        
             | ilaksh wrote:
             | Had no idea that was what you were asking for. Search for
             | Zoom Face Filter or OBS Face Filter OBS deep fake live etc.
        
       | nicktikhonov wrote:
       | From what I've seen, it's really easy to get PersonaPlex stuck in
       | a death spiral - talking to itself, stuttering and descending
       | deeper and deeper into total nonsense. Useless for any production
       | use case. But I think this kind of end-to-end model is needed to
       | correctly model conversations. STT/TTS compresses a lot of
       | information - tone, timing, emotion out of the input data to the
       | model, so it seems obvious that the results will always be
       | somewhat robotic. Excited to see the next iteration of these
       | models!
        
       | dubeye wrote:
       | It doesn't feel like speech recognition has been improving at the
       | same rate as other generative AI. It had a big jump up to about
       | 6% WER a year or two ago, but it seems to have plateaued. Am I
       | just using the wrong model? Or is human level error rate, some
       | kind of limit, which I estimate to be about 5%.
        
       | api wrote:
       | How close are we to the Star Trek universal translator?
        
         | ilaksh wrote:
         | Different type of model but you can buy those on Amazon etc.
        
       | ilaksh wrote:
       | Does anyone have working code for fine-tuning PersonaPlex for
       | outgoing calls? I have tried to take the fine tuning LoRA stuff
       | from Kyutai/moshi-finetune and apply it to the personaplex code.
       | Or more accurately,various LLMs have worked on that.
       | 
       | I have something that seems to work in a rough way but only if I
       | turn the lora scaling factor up to 5 and that generally screws it
       | up in other ways.
       | 
       | And then of course when GPT-5.3 Codex looked at it, it said that
       | speaker A and speaker B were switched in the LoRA code. So that
       | is now completely changed and I am going to do another dataset
       | generation and training run.
       | 
       | If anyone is curious it's a bit of a mess but it's on my GitHub
       | under runvnc moshi-finetune and personaplex. It even has a gradio
       | app to generate data and train. But so far no usable results.
        
       | ricardobeat wrote:
       | No mention of tool use. If the model cannot emit both text and
       | audio at the same time, to enable tools, it's not really useful
       | at all for voice agents.
        
       | Krisso wrote:
       | Awesome, but given the Apple Silicon population and
       | configuration, how does this fare on a M1 with 8GB of total ram?
       | I'd imagine this makes running another llm for tool-calls and
       | inference tough to impossible.
        
       | d4rkp4ttern wrote:
       | Built out the demo on my M1 Max Macbook and it was absolutely
       | terrible. Around 10 seconds for each reply, and even then it was
       | saying something totally unrelated.
        
         | d4rkp4ttern wrote:
         | Also in general I don't know get what the appeal of a 7b full-
         | duplex (speech-to-speech) model is: 7b can't be very
         | intelligent on its own, and for anything useful, you'd need
         | tool-calls, which speech-to-speech models can't do. This is
         | also why ChatGPT voice mode annoys by never doing a web search
         | or reading a link (in fact it pretends to search or read,
         | outright makes up stuff, and when pushed admits it can't really
         | read web pages or do web searches).
         | 
         | There are probably definitely use cases for this though, open
         | to be educated on those.
        
           | water-drummer wrote:
           | Gemini live api and grok voice api can make tool calls and
           | they're speech to speech models
        
             | d4rkp4ttern wrote:
             | Right, turns out Claude and ChatGPT voice can also do web-
             | search. So I guess behind the scenes there is more than a
             | "pure" voice-voice model being used, i.e. there's probably
             | a rudimentary agent loop with tools + tool-exec interposed.
        
           | satvikpendem wrote:
           | Why can't a speech to speech model do tool calls? Others like
           | Gemini live do it just fine.
        
             | d4rkp4ttern wrote:
             | Ok, I was wrong. I just tested ChatGPT voice, Claude Voice
             | and Gemini Live. And all three are able to do web search.
             | For some reason, I thought when I tested ChatGPT voice a
             | few weeks ago, it sometimes said it can't directly open
             | links, but it can do web search, which was strange.
        
             | raw_anon_1111 wrote:
             | If it is doing a tool call, it has to convert the speech to
             | text or at least a JSON object of the necessary parameters
             | for the tool and convert the result to speech doesn't it?
             | Is it truly speech to speech then?
        
               | satvikpendem wrote:
               | It's all tokens at the end of the day, not _really_ text
               | or video or audio, just like everything on a machine is
               | just bits of 1s and 0s and it 's up to the program to
               | interpret them as a certain file format. These models are
               | more speech-to-speech (+ text) in that they can recognize
               | text tokens too. So the flow is, you ask it something,
               | then,
               | 
               | Audio Tokens: "Let me check that for you..." (Sent to the
               | speaker)
               | 
               | Special Token: [CALL_TOOL: get_weather]
               | 
               | Text Tokens: {"location": "Seattle, WA"}
               | 
               | Special Token: [STOP]
               | 
               | The orchestrator of the model catches the CALL_TOOL and
               | then calls the tool, then injects this into the context
               | of the audio model which then generates new tokens based
               | on that.
        
           | WhitneyLand wrote:
           | Yes. Is there a basic chat app for iOS that prioritizes full
           | intelligence over full duplex?
           | 
           | Agree ChatGpt advanced voice mode is so bad for quality of
           | the actual responses. Old model, no reasoning, little tool
           | use.
           | 
           | I just want hands free conversations with SOTA models and
           | don't care if I have to wait a couple of seconds for a reply.
        
             | dahcryn wrote:
             | I saw a demo of parloa (or maybe it was a different
             | provider), and no joke, they insert sound of typing on a
             | keyboard or stuff like that during an LLM tool call, its
             | weird but surprisingly effective lol
        
         | scotty79 wrote:
         | On something around rtx 5070 it reacted faster than a human
         | would.
        
         | mrkstu wrote:
         | Quoted from linked article:
         | 
         | "PersonaPlex accepts a text system prompt that steers
         | conversational behavior. Without focused instructions, the
         | model rambles -- it's trained on open-ended conversation and
         | will happily discuss cooking when asked about shipping.
         | 
         | Several presets are available via CLI (--list-prompts) or API,
         | including a general assistant (default), customer service
         | agent, and teacher. Custom prompts can also be pre-tokenized
         | and passed directly.
         | 
         | The difference is dramatic. Same input -- "Can you guarantee
         | that the replacement part will be shipped tomorrow?":
         | 
         | No prompt: "So, what type of cooking do you like -- outdoor
         | grilling? I can't say for sure, but if you're ordering
         | today..."
         | 
         | With prompt: "I can't promise a specific time, but we'll do our
         | best to get it out tomorrow. It's one of the top priorities, so
         | yes, we'll try to get it done as soon as possible and ship it
         | first thing in the morning.""
        
         | jayavanth wrote:
         | what is your context size?
        
       | KaiserPister wrote:
       | I am strongly put off by the LLM writing in this piece. It makes
       | me question quality of the project before even attempting a
       | download.
       | 
       | Who would put effort into building this only to compose a low
       | effort puff piece?
        
         | chromehearts wrote:
         | I hate those AI generated graphs / charts more than the text
        
           | giancarlostoro wrote:
           | You on about this article or other articles? I dont mind AI
           | generated images to a degree, charts I might start to worry.
        
         | neurostimulant wrote:
         | But isn't it normal for people who work on AI stuff to use LLMs
         | for everything? They are very enthusiastic about AI so
         | naturally they'll use it on everything they can.
        
           | DrammBA wrote:
           | Sometimes I wish they just posted the prompt, not everything
           | has to go through an LLM blender before posting.
        
           | KaiserPister wrote:
           | That's a bit reductive. Some do, others don't. I do a lot of
           | AI development, and building. But I value the act of writing
           | for clarifying my thoughts. And I value other people's time
           | when reading my writing.
        
           | moffkalast wrote:
           | They are the ones who should know best when not to use it.
        
         | ttul wrote:
         | What gives you the sense that the piece was written by an LLM?
         | I would agree that the diagrams have some of the artifacts
         | common in Nano Banana output, but what tips you off about the
         | text?
        
           | rush340 wrote:
           | Em dashes in every other sentence. I've never seen an actual
           | person do that. The language in general reads exactly it's
           | written by an LLM:
           | 
           | "The blah blah didn't just start as blah. It started as
           | blah..." "First came blah -- blah blah blah" "And now: blah"
           | 
           | It's a distinctly AI writing style. I do wonder if we'll get
           | to a point where people start writing this way just because
           | it's what they're used to reading. Or maybe LLMs will get
           | better at not writing like this before that happens.
        
           | tverbeure wrote:
           | I'm sick and tired of the "No..., no ..., (just) ..." LLM
           | construction. It's everywhere now, you can't open a social
           | media platform and get bombarded by it. This article is full
           | of it.
           | 
           | I get it, I should focus just on the content and whether or
           | not an LLM was used to write it, but the reaction to it is
           | visceral now.
        
         | 0xbadcafebee wrote:
         | I wasn't put off by it. I read the article, got all the
         | information I needed, it was interesting and informative. (In
         | fact, I find the human-written ones more often annoying; most
         | people are not good at writing, and are apt to create huge
         | walls of text, whereas the AI is biased towards making the
         | information easy to consume)
        
           | KaiserPister wrote:
           | I do agree it is one of those "if I had more time, I would
           | write a shorter letter" situations.
           | 
           | But in this case the piece is wordier than a bad human writer
           | would be. If they want to use ai for writing, so be it, but
           | at least include "concisely" in the prompt.
        
       | ruhith wrote:
       | Cool demo but without tool calling this is basically a fast
       | parrot. The traditional pipeline is slower but at least you can
       | plug in a real brain.
        
         | mrtesthah wrote:
         | voice to voice models can call tools. no need for TTS.
        
       | d4rkp4ttern wrote:
       | Sesame was the best full-duplex voice demo I ever came across,
       | wonder what is up with them now https://app.sesame.com/
        
         | sweetheart wrote:
         | holy shit I cannot believe how polished that is.
        
         | StevenNunez wrote:
         | I enjoyed unmute.sh too
        
           | d4rkp4ttern wrote:
           | Yes that one is great too
        
       | sowbug wrote:
       | I would like my phone to forward spam calls to this, with a
       | system prompt to slowly provide fake personal and financial
       | information intermingled with chatter about sports and the
       | weather.
        
         | hedora wrote:
         | Yeah; that and spam texts. "I have no idea who the person you
         | were trying to reach is, but, yes, the recent weather patterns
         | have created strange surges in my dishwasher. The karmic energy
         | of my spoons is all off. I am interested in having you maintain
         | all my appliances. I'm a landlord and own 25 nude goat yoga
         | worship rooms. They go through a lot of dishes!"
         | 
         | Bonus points if it correlates the spam texts with follow up
         | phone calls from the spammers.
        
       | sarmike31 wrote:
       | If you're interested in Demos without installing the thing, he
       | has a site here:
       | https://research.nvidia.com/labs/adlr/personaplex/
        
       ___________________________________________________________________
       (page generated 2026-03-05 23:00 UTC)