AI Engineer Europe 2026
How to talk to statues — Joe Reeve, ElevenLabs
Read the talk
How to talk to statues
A photograph becomes a speaking museum object, then opens a harder design problem: how to give voice agents useful knowledge, conversational cues and interfaces people can steer.
From a talk by Joe Reeve
Before you start: Basic familiarity with APIs and chat-message history helps with the implementation discussion; no audio-model background is required.
What would a statue sound like?
What would it take to photograph a statue and have a conversation with it? The necessary pieces already exist as audio services. Joe Reeve, who works in ElevenLabs’ growth organization, introduces text-to-speech, transcription, sound effects, voice creation and editing, and a managed agents platform. He singles out voice design as an underused capability: a developer can describe a voice instead of selecting an existing one. He also presents music generation as commercially licensed; the Eleven Music API announcement describes broad commercial use subject to plan restrictions, rather than unrestricted clearance for every use.
The statue app makes that collection of capabilities tangible. Its introductory slide pairs an interface identifying Artemis the Huntress with a photograph of the British Museum entrance. Reeve then plays a museum visit in which objects introduce themselves: Pharaoh Amenhotep, Demeter, Hoa Hakananai'a, Hans Sloane, the Guardian Lion, the Young Rider and the War Horse. A collection that visitors normally look at becomes a collection they can address.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.
From a photograph to a conversation
The app connects the services in a short sequence:
- Photograph the object. Use the image to identify the statue.
- Research its identity. OpenAI supplies historical knowledge and descriptions of what the statue’s voice might sound like if it were alive.
- Generate a voice. The Voice Design API turns a textual description into a matching synthetic voice.
- Create an agent and start a call. Combine the persona’s knowledge and voice in an ElevenLabs agent, then open the conversation.
Reeve reports roughly 30 seconds from taking the photograph to talking with the agent. That is his end-to-end example, without a supplied timing distribution or model configuration.
The QR code leads to Reeve’s companion build tutorial. His spoken description calls the research step deep research and the resulting interaction a phone call. The subsequently updated tutorial uses GPT-5.2 with web_search_preview and a browser WebRTC conversation; it does not establish the exact historical research model or a telephone-network connection for the demonstration. The enduring architecture is image identification, research, voice generation and agent creation.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.
A fast prototype meets institutional demand
Reeve reports building the prototype in Cursor in two hours on a Sunday and publishing its build prompt through the ElevenLabs blog. The first video attracted interest from museums and travel businesses, including Tripadvisor competitors. He says three museums or museum-group representatives contacted him; one CEO told him a team of ten had spent a year on a similar idea. These were accounts of interested organizations, not evidence that the prototype had already met their requirements.
The next post emphasized how quickly he had built it. Reeve reports that impressions grew from 50,000 on the first day to 1.5 million on the second after the vibe-coding repost. The response expanded into the arts: Bonhams and Christie’s wanted people to converse with items offered for sale. The experience also helped inspire ElevenHacks. What began as an API experiment had become a question about cultural experiences, distribution and what people could build for themselves.
An audience member asks the obvious engineering question: how does this become a production service for millions of users? Reeve’s answer separates the underlying services from the application around them.
| Layer | Prototype’s approach | Remaining responsibility |
|---|---|---|
| Voice generation and agents | Managed APIs | Depend on provider capacity and operation |
| User accounts | Potential third-party service | Implement and check authentication |
| Product integration | Glue between services | Finish the museum experience |
Reeve expects the voice and agent APIs to have ample capacity for this app. He does not demonstrate a million-user deployment. His point is that useful integration and a compelling explanation can create substantial value without rebuilding the underlying infrastructure.
Production work still remains. He is discussing a possible ElevenLabs-supported offering with museums, and suggests Supabase logins and magic links as conventional application pieces Cursor could help implement. Generated code still needs checking. Managed APIs remove major infrastructure tasks; they do not finish the product or decide what a museum should put into it.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.
Talk to the right agent, with a screen alongside it
Voice interfaces often force a choice between speaking and using another interface. Reeve wants them to work together. His coding example is a Lovable-like app in which the user talks to a product-manager agent, which then delegates implementation to a coding agent. The conversational partner and the worker need not be the same component: one helps establish what should happen, while another carries out the task.
A companion interface could display information extracted from the conversation and let the user edit it while speaking. That makes the conversation inspectable and correctable without requiring every correction to become another spoken exchange. There is also a social obstacle: people hesitate to interrupt an agent. Reeve finds that interrupting more aggressively improves his experience, but has no settled way to give other users permission to do it.
The next question concerns skills and prompt guidance. In the platform Reeve describes, ElevenAgents has no native skills concept, but it can load knowledge files and make MCP calls. Those are possible routes for supplying knowledge or exposing capabilities. More generally, voice can be attached to an existing coding agent rather than requiring the voice platform to reproduce its entire working environment.
An audience member makes the desired interface concrete: speaking to Claude Code in a human voice. Reeve reports that people have connected ElevenLabs to OpenClaw and some Claude Code setups, including phone numbers they can call and agents that call back. He also describes adding skill-loading capability through the agent itself as a possibility. These are examples of integration patterns, not a live implementation shown in the session.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.
Make the technology part of the experience
Museum engagement remains an open design problem even for museum leaders. Reeve describes discussions with Science Museum leadership and its willingness to experiment with interactive tablets for children. To him, a screen beside an exhibit can still feel like technology attached to an existing experience. He reports commissioning a statue with embedded technology, a speaker, a phone and a microphone so visitors can speak directly to it. At the venue, a K6 red British telephone booth offers another physical interface: pick up the handset and talk with a Sir Michael Caine agent.
An audience member extends the idea from consuming an exhibit to creating an experience: a child could arrive at a science museum and build something in response to what they encounter. Reeve thinks consumer creation has not yet reached that point. Even tools aimed at nontechnical users can feel oriented toward B2B SaaS applications, with Supabase and familiar design components determining the shape of the result.
Evening vibe-coding events offer a different picture. Reeve meets people who have never considered that someone makes the apps on their phones, much less learned what an accordion or hamburger menu is. They describe what they want, and an LLM attempts combinations an engineer might immediately translate into familiar components. Their lack of software vocabulary can allow unusual ideas to survive longer. He wonders what an Instagram-filters or TikTok moment for social software creation would look like.
He mentions a mobile game-creation app he calls Spielwork, with TikTok-style swiping, and an unnamed London game-building tool. Games may themselves be too complex an entry point, he suggests, and many offerings still resemble Lovable on a phone. The audience broadens the goal to content creation, building and interacting with digital systems—including spawning game objects by voice. A mainstream creator should be able to express an intention without first adopting concepts such as databases.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.
Freeform speech in, dense information out
An audience member identifies an asymmetry: speaking is a convenient way to get thoughts out, but listening can be an inefficient way to receive the result. They want to express raw intent and see it transformed into diagrams, text or application state. Email is one desired setting: speak freely, then inspect an information-rich result rather than wait through a spoken explanation.
Reeve wants the same arrangement: one voice input, parallel outputs. A little speech might accompany a generated interface, a diagram or changes in the current application. Another audience contribution adds a different benefit of spoken output: companionship. A voice can make tinkering feel less lonely and sustain motivation even when text or diagrams would convey more information. The choice of output therefore depends on both the information being delivered and the experience surrounding it.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.
Navigate spoken answers by concept
The discussion briefly reaches for an evolutionary explanation of visual processing, but the practical usability observation is more immediate: concise text and concise speech can feel different. One participant is happy to request a short written answer, yet finds the equivalent spoken answer abrupt or rude. Simply reducing word count does not solve the listening problem.
Reeve proposes skim listening. Forward and backward controls could initially jump through pieces of an answer, much as podcast speed controls or an old iPod’s scrubbing interface let listeners move through audio. The discussion then refines the unit of navigation: a listener wants to move between concepts or topics, not necessarily sentence boundaries.
That distinction matters because a spoken navigation request can accidentally become a generation request. If an agent is reading a long answer and the user asks it to move to the next point, it may treat the interruption as a new prompt and generate another long explanation of that point. The intended operation was to advance through an existing answer, not expand it.
The room suggests several ways to expose the answer’s structure:
- Topic navigation: move directly to the next concept.
- Summary and expansion: show a short paragraph summary, then hold a control to hear more.
- Separate visual and spoken content: display higher-level sections while speech develops each one.
Reeve describes Claude’s voice interface as exploring the last pattern with tappable sections. These controls move beyond the conventions of ordinary human conversation, but they give a listener direct access to information that otherwise unfolds only over time.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.
Make turn-taking visible
Human conversations carry signals that audio-only agents lose. A listener can often see that another person is about to interrupt, and a speaker can accelerate, change direction or notice disengagement. With an agent, the listener may not even know whether the answer will take ten seconds or a minute. Advance information about its length could help them decide whether to continue or cancel. The interaction resembles a phone call in its missing visual cues.
Even incoming speech has different meanings. Acknowledgments such as “uh-huh” or “yeah, yeah” can encourage a speaker to continue. A request to go back is steering. Neither is necessarily a demand to stop the current turn and begin a completely new exchange. A voice interface needs to preserve those distinctions.
An audience developer describes a game in which a player can request a getaway vehicle or helicopter while participating in a shooter-style encounter, then use the generated object in the world. Automatic voice interruption proved unreliable for their work, so they adopted a Wispr Flow-style push-to-talk boundary: hold to speak, release to finish. A physical input supplies a clear signal that the audio stream alone was failing to provide.
For the opposite direction, the developer proposes a small visual indicator showing that the agent wants to ask a question. The user could finish speaking or invite it to contribute, without being interrupted immediately. Reeve extends the idea: show the topic the agent wants to raise. That would communicate more than a human’s preparatory gesture—the user could decide whether that particular contribution is worth taking now. These are proposed interface behaviors, not features demonstrated in the session.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.
Conversation history should reflect what was heard
How would an agent decide that it has something to add? An audience member asks whether this requires a tool call, RTC audio handling, a separate channel or changes to the system prompt. Reeve’s proposed approach is an asynchronous transcript monitor: repeatedly inspect the evolving conversation and ask whether there is a useful contribution to make. He explicitly describes this as something still to build, rather than an existing transport or tool-calling design.
The follow-up question is how to steer an agent that already has a plan. Reeve separates the agent into next-message logic and a malleable conversation transcript. In the behavior he describes, a platform generates an answer’s text and then plays its audio. If the user interrupts partway through playback, the platform may append the interruption after the entire generated answer—even though the user heard only its beginning.
Generated text and heard speech are different states. Reeve suggests using playback timestamps to locate the interruption and remove the unheard suffix from conversation history. For a concrete example, suppose an assistant generated First, consider the material. Next, consider the carving. Finally, consider the museum. The listener hears only the first sentence and says Go back to the material. The proposed history repair keeps that first sentence, removes the unheard remainder and appends the interruption.
Repair history at the interruption point
Constructed example: The message IDs, system instruction, assistant text, user interruption and first-sentence playback boundary are teaching values. The transcript-repair mechanism is Reeve's unbuilt proposal.
Go back to the material.
Operation: Proposed repair: retain the heard prefix of assistant-1, remove its unheard suffix and append interruption-1.
System message
Explain the object’s history.
Explain the object’s history.
Assistant message
First, consider the material. Next, consider the carving. Finally, consider the museum.
First, consider the material.
User interruption
Not present
Go back to the material.
The following JavaScript expresses that history operation once playback alignment has supplied heardTextEnd, the exclusive character boundary of the text heard. It requires the interrupted assistant message to be the latest history entry, so unrelated messages cannot be silently removed.
javascript
function recordInterruption(history, messageId, heardTextEnd, userText) {
const last = history.at(-1);
if (!last || last.id !== messageId || last.role !== "assistant") {
throw new Error("Expected the interrupted assistant message last");
}
if (!Number.isInteger(heardTextEnd) ||
heardTextEnd < 0 || heardTextEnd > last.content.length) {
throw new RangeError("Invalid heard-text boundary");
}
const heardText = last.content.slice(0, heardTextEnd).trimEnd();
return [
...history.slice(0, -1),
...(heardText ? [{ ...last, content: heardText }] : []),
{ id: "interruption-1", role: "user", content: userText }
];
}
const history = [
{ id: "system-1", role: "system", content: "Explain the object’s history." },
{
id: "assistant-1",
role: "assistant",
content: "First, consider the material. Next, consider the carving. Finally, consider the museum."
}
];
const repairedHistory = recordInterruption(
history,
"assistant-1",
"First, consider the material.".length,
"Go back to the material."
);
The timestamp-to-text alignment is the essential input; a fraction of elapsed audio time is not necessarily the same fraction of the text. This example implements the proposed transcript edit, leaving playback cancellation and subsequent response generation to the surrounding application.
An audience member suggests treating such user messages as a distinct form of steering. Reeve’s response is to invite a prototype at the ElevenLabs booth. The practical experiment is small but consequential: make the next response depend on the conversation that actually reached the listener, while giving navigation and acknowledgment their own meanings.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.
Make the demo worth watching
The final question returns to distribution: when building becomes easier, how do people discover what you made? Reeve learned video production through political campaigning before joining ElevenLabs, and that experience fed into ElevenHacks. He finds virality unpredictable. Practice and iteration are more dependable than confidence that a particular video will succeed.
He applies an 80/20 heuristic to editing: reaching the statue video’s level of polish was relatively easy, while going further would introduce substantially more work. Reeve reports spending about 20–25 minutes editing the statue video on his phone. He finds desktop CapCut much harder to use and describes its subscription as three times the phone price in his experience. The practical choice is to keep the workflow simple enough to repeat.
Captions help, and the interesting part needs to arrive early. Reeve reports median viewing times of 6–12 seconds across his videos. His advice is to inspect the platform’s analytics after publishing several examples, then put the hook before the point where viewers tend to leave. A compelling demonstration buried late in the edit will never reach much of its potential audience.
Music can be developed in either direction:
- Narrative first: edit the speech and story, generate different musical genres, then adjust musical sections to fit the spoken sections.
- Music first: choose the feeling, generate a track, and build the speech and edit around it.
For the statue video, Reeve chose the music first. Its imperial feeling suited the British Museum exterior and served as an attention grabber. Even quiet background music, he finds, strongly changes the feeling of a scene.
The equipment stayed simple: a mobile phone, CapCut and a borrowed DJI Bluetooth lapel microphone that improved the recorded audio. Reeve says the microphone cost £200. He deliberately planned the shot outside the British Museum, but the finished video came from a second recording trip. The first visit produced footage he found boring; he went back, took a collection of silly photos, and those became the video people saw. The iteration was in the demonstration itself as much as in the software.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.
Resources
From the talk
Joe Reeve's build prompt and tutorial connecting image recognition, historical research, synthetic voices and a WebRTC conversation.
Generate voice previews from a description and save a selected voice for use with other ElevenLabs APIs.
Access museum collection metadata and images, with documentation and worked examples for exploring the records.
Further reading
- Eleven Music API launchArticle
The August 2025 announcement explains music generation capabilities and plan-dependent commercial usage.
Read the complete timestamped transcript
- 0:00
[upbeat music] Can I get a vibe check of the room?
- 0:17
How's everyone feeling? Are we like, "We really want to hear what Joe has to say about statues,"? Or are we like, "I just wanna like chill and just sit in a r- quiet room"?
- 0:26
How are people feeling?
- 0:27
Statues.
- 0:28
Statues. [laughs] Okay, that's good to hear. You're much more lively than statues, which I've spent a surprising amount of my, my life with recently, so. [laughs] Um,
- 0:36
this, I'm gonna, uh, well, actually, I'll quickly introduce myself. I'm Joe. I work in the growth organization at ElevenLabs. Hands up if you've... Actually, I've got a slide for that.
- 0:45
Hands up if you've ever used/heard of ElevenLabs. Okay. Um, so th- this next slide, you'll probably be familiar with a lot of it. Um, ElevenLabs does, basically we're an audio AI foundation model company, so everything from text-to-speech, you put some text in, you get some speech out, um, transcription, the other direction.
- 1:04
Music, we've got the first- [coughs] ... um, commercially legal, uh, AI music generation. You know, we licensed all the, all the training data, uh, behind an API. Sound effects. Create voices.
- 1:15
This one, by the way, just like pro tip in case you're ever using ElevenLabs at a hackathon or something, this section, creating and editing voices, is like the thing that people really just don't use ElevenLabs enough for, and this is the thing that the Statue app that I built, that we'll talk about, um, is built, built on.
- 1:30
And then agents. Uh, this is our fully managed agents deployment platform, which sounds like a mouthful of, of SaaS, and it sort of is, but it's also very cool in, in various ways.
- 1:40
So who saw the Statue app? It went quite viral, at least here in London. Raise your hands if you saw it. Okay, that's fine, because I'm gonna play you a video, 'cause, I guess people like me love playing videos- [laughs] ...
- 1:52
of our own voice out loud. Um, it's-- I'll just play the first 30 seconds or so. I made an app that lets you talk to any statue you want using AI.
- 2:02
So we've come to the British Museum to see how it works. [trumpets playing]
- 2:09
I am Pharaoh Amenhotep.
- 2:11
I am Demeter.
- 2:12
Hoa Hakananai'a.
- 2:14
I am Hans Sloane.
- 2:16
The Guardian Lion. I am the Young Rider.
- 2:19
And I am the War Horse.
- 2:22
First, take a picture. Okay, so now I'm about to explain that bit of me explaining. So we, uh, basically what this statue does is it lets you take a picture of a, of a statue, sorry, this app.
- 2:32
It lets you take a picture of a statue. Um, it then does a, an OpenAI deep research, uh, on, on the identity of the statue, generates a bunch of the sort of historical knowledge and prompts for what it thinks the, the, the voices of those individual statues would've been if they were alive.
- 2:50
Um, it then uses our Voice Design API, that really underutilized API where you can put in a description of a voice and it will go and generate something that matches.
- 2:59
And then it creates an ElevenLabs agent, and then starts a phone call. And that whole thing works in like 30 seconds. So you take a picture of something, you get all the like search research back from OpenAI, generate a voice, and start talking to an agent, to a statue within 30 seconds, which is pretty fun.
- 3:15
Um, what sort, uh, if you, if you're interested in like reading more about the details of how it was all built, you can scan that QR code. It's just a blog post.
- 3:22
Um, this was attached to the, the initial tweet that I made, um, which basically had the prompt for one-shotting. This whole thing I built in Cursor in two hours.
- 3:33
Um, it, it sort of, sort of was. So I, I built this, um, in, in two hours on a Sunday 'cause I was, I was sort of tired and bored, and then published the, the prompt through the ElevenLabs blog.
- 3:44
Made this video that you just saw, posted it on a Tuesday or something, and got 50,000 impressions. It's like pretty, pretty good, not bad. Um, people on, on Twitter kinda liked it.
- 3:54
And then three museums, or people who represent groups of museums, and a bunch of other like businesses, including like TripAdvisor c- competitors and stuff, who were coming and saying, "We've been"...
- 4:05
Well, actually, one of them, the CEO called me. He found my WhatsApp phone number somewhere, and called me up and said, "I've had a team of 10 people working on this for a year.
- 4:12
How did you build this?" Um, and so then the next day I, I reposted saying, "I've had a bunch of like interesting people. I vibe coded this in, in like two hours."
- 4:22
Not as a brag, just as a, this is interesting, like vibe coding is so powerful, and if you are sort of experimenting with these, these interaction patterns, you can actually do something quite big.
- 4:31
Surprisingly big. Um, it then went completely viral and went from 50,000 on the first day to one and a half million on the second day. Uh, and it was in part because it got kicked off by the vibe coding, and then suddenly I got all these artists and creatives and everything from, from portrait, uh, museums to like
- 4:50
Bonham's and Christie's reaching out saying, "We want to have people be able to talk to the items we wanna sell." Um, I think what I would sort of... Oh, and then, and then this led to something called ElevenHacks, which we can talk about later maybe.
- 5:01
Um, there are loads of different things in this story we can talk about. We can talk more about the Statue app and about ElevenLabs. I sort of wanna get, get more from you.
- 5:10
Um, we can talk about ElevenLabs generally. We can talk about what it means to do sort of growth, particularly API growth, uh, at one of these companies, um, from the growth engineering point of view.
- 5:19
And, and, uh, we can talk about the sort of implications on culture, which I think is something that we're not, as a, as an industry, we're not really looking at as seriously as, as we could be.
- 5:29
Um, vibe coding generally, and again, what that impact is having on society, um, voice interaction patterns or making viral videos or anything else. So it's sort of... Yeah, please.
- 5:39
I mean, it's really easy to prototype these days, right?
- 5:43
Yeah.
- 5:43
But it's like pushing it into production that's the hard part. So maybe, I don't know, I'm curious about how you would take your really successful prototype to like how can you scale it to millions of users maybe?
- 5:53
Yeah. Well, that's one of the things, uh, this is gonna sound very ElevenLab salesy now. The nice thing is that pretty much all of, like what I've done is stitched together existing APIs which are designed to scale.
- 6:02
So yeah, if I wanted to start doing user management, that sort of thing, that's relatively, I think, well understood, and you can u- buy in other third parties for that.
- 6:10
But the, the hard bit, the maintaining the agents and the voice design, that's all APIs that there's no way I can make a dent in the API volume, even if this, this goes absolutely gangbusters.
- 6:20
So from that point of view, I think, and I think that's something that vibe coding is really showing, is that the glue pieces and, and telling a good story about the glue is the im- in part the most important thing of the project rather than the solving hard technical problems.
- 6:34
So yeah, obviously, there's a lot more work to do to make it actually production ready, uh, which is something that I've been talking to a bunch of the, the museums about.
- 6:41
Uh, we might be doing that as like a nice ElevenLabs gives something nice to the, to the museums. But it's-- the hard bit is actually not the user management, and like Cursor, again, can pretty much...
- 6:51
You need to, to check it all, obviously. Can pretty much one-shot that with Supabase or whatever else for, for logins and, and magic links and that sort of thing.
- 6:58
It's, it's mostly the relying on our, our APIs and our, our agents platform to do the heavy lifting, I suppose.
- 7:05
What about evals?
- 7:06
Evals, I get-- So that's, that's one of the big things. I think the, the take a photo and just get your research back is not really the long-term solution for the, for the museums.
- 7:15
Really, the important bit, the hard bit is going to the curators and saying, "Curators, can you figure out what is the actual narrative? Let's not just take random things you found from Google.
- 7:23
Let's actually, like, put some thought in, in, in design to the content." So that's, I think, the piece that's the longer tail. The nice thing is a lot of the museums have-- they see their core IP as these databases.
- 7:35
Um, so they have APIs often, um, and s- we can pull that information out. The V&A has a public API for, for their, their stuff, so...
- 7:45
Yeah. I guess on this topic, um, maybe related to AI culture and voice interaction patterns, uh, how-- what is the interface then that you enable the curator to design the experience?
- 7:59
Is it Sandra or?
- 8:00
Yeah. So right now, there's no, no... I haven't designed anything for that. I mean-
- 8:04
Oh
- 8:04
... at best, they could log into a, to a, a dashboard, the ElevenLab dashboard, and make, make edits to the system prompt and the knowledge base files that are in there.
- 8:12
Um, the-- I th- I think probably in, in this sort of d- information management, the best case or the best interaction pattern is probably editing text, uh, rather than speaking.
- 8:24
Although, obviously, if you're choosing a voice, you need to sort of manage that. There's an interesting question there. I've been talking to, uh, a chap, Jago, who used to be the head of the Americas at the British Museum and now runs the Sainsbury Centre, which amusingly is the location for the, uh, the Avengers headquarters in the Avengers
- 8:42
films. Um, he is going through this big, long academic process of figuring out what should a voice sound like for an inanimate object. So things like, where did the materials originally come from?
- 8:54
It came from some mountain in China, or it came-- And then, so it, it... The, the rock was shipped to Vietnam, and then it was carved in Vietnam, and then it spent the last two hundred years living in a British museum.
- 9:05
So like, what should it sound like? It would have a little bit of a maybe a Chinese origin with some Vietnamese twist in there, but then it's just lived around people with British accents, but maybe also not because it's, it's lots of tourists.
- 9:15
So actually thinking through from a, a much more philosophical point of view, what should objects sound like? And that's something that, from my point of view in, in, in ElevenLabs, is really interesting because I, I think we have the opportunity to give all sorts of things voices, like elevators, right?
- 9:29
A lift should probably-- I mean, they do have, have voices. They're often quite discordant with what a lift is, I, I find. But if-- It's quite likely, I think, that we start walking into lifts and saying, "I want to go to this floor, please," using voice to, to, to, to, to interact with them.
- 9:42
So what should they sound like? And it's beco- becoming a more important question. I, I don't know exactly what the answer is, but smarter people are doing that. So...
- 9:50
Sorry.
- 9:50
Sorry. Yeah.
- 9:50
Yeah. So it was re- really interesting to hear, uh, about your thoughts about, uh, voice as a, an interface to application.
- 9:58
Mm-hmm.
- 9:58
But in general, what, what, uh, problems arise? Uh, what do you think of how do users i- uh, uh,
- 10:06
interact or how, what, what do they expect from this voice agent, and when does it not work? Uh, what problems do we, do we expect if we are going to be embedding the voice interface?
- 10:17
Yeah. I think currently voice interfaces, voice interactions have quite a large range of problems currently, but a lot of them are solvable. One of them is you basically have a binary.
- 10:28
You're either interacting with voice, or you're interacting in some other way, and I still feel like the sort of interactive or generative UI plus voice is something that we still haven't seen.
- 10:38
Particularly, we've got... So sometimes it's, and this is something I've experimented with, is like if you think about a coding agent, you've got Lovable or something. I want to be able to talk to my app, talk to Lovable, uh, but not talk to the coding agent part.
- 10:51
I want to talk to like a product manager agent that then goes off and triggers my coding agent to go and do things. So what does that... So you-- The voice interaction there is not direct.
- 11:00
The, the thing that I'm talking to it or the thing that's doing the work is the thing I'm talking to. I sort of want to be talking to a, to a, a halfway house person.
- 11:08
So I don't-- What are the problems? Aside from often the thing you, you, you end up talking to is not the thing you actually want to be talking to.
- 11:15
There's also the, what are the parallel sort of interaction patterns, uh, in UI? And I think actually you were just showing me your app earlier, which does show the, the stuff that the, the voice agent's thinking and, and sort of extracting from the conversation and allows you to, to interact with that at the same time.
- 11:29
That's something that I, I think we're gonna see a lot more of, the sort of multimodal conversations where it's voice and visual. Um, the other thing is people don't like interrupting, uh, voice agents 'cause they're too polite.
- 11:41
People are too polite, and I'm starting to learn to just like interrupt agents much more aggressively, and that actually makes the experience much better. But we-- I don't know how to give people permission to interrupt.
- 11:50
Sorry.
- 11:50
How do you, how do you, uh, solve the problem of like prompt guidance or skill loading? So I mean, with a typical coding agent, you can give like skills, this, this, this.
- 11:59
You can throw anything at it, right? But I don't think it's the same interface or... Is it the same interface for-
- 12:06
So the, the ElevenAgents platform doesn't really support the concept of skills, though it could. It does support the concept of knowledge files, which do get loaded in. So you could do it in that way.
- 12:14
We also support MCP calling, so you can have knowledge em-embedded in those, um, or skills embedded in, in the MCPs. The, uh-
- 12:23
I, I don't think that's core to voice or not. Uh, I think that's mostly down to, like, the, the interaction patterns of, of coding agents are quite... l- lend themselves to skills.
- 12:34
Um, but you could have a voice agent that then has the ability to use skills, and you can add voice capabilities to an, to an existing coding agent that, that...
- 12:42
in, in that way. So I don't know if that's related directly to skills, but, um-
- 12:46
I think w- basically I want to talk to Claude code as a, as a human voice.
- 12:51
I see. I see. Well, actually, some people, some people have built the, particularly with OpenClau actually, there's ElevenLabs is, is quite a common interaction pattern for OpenClau. People have built phone, phone numbers they can call, and it'll call them back and that sort of thing.
- 13:01
Um, and then obviously you can, if you say, "Well, I want you to be able to load in skills," you, it just, like, learns how to do it and, and adds its, adds that capa-capability to itself.
- 13:10
So that's for sure possible. Um, and people are doing it with their sort of OpenClau setups and, and some Claude Code setups. But, yeah.
- 13:19
In, in this experience, one, the AI culture and the vibe coding-
- 13:22
Mm
- 13:22
... thinking about new interaction patterns to engage with history, with our built environment, like, where or how do you see that starting to... Again, like, I mean, I mean, we're so empowered now with this vibe coding, but it's still-
- 13:36
Mm
- 13:36
... there's a barrier, but how do we build that engagement and, and do it well? I don't know what the museums are telling you. I mean, I'm sure this is so new, but.
- 13:44
Yeah. I think, to be honest, the, the museums, I, I, I've met with the CEO of the science museum, or co-CEO of the science museum, and they're asking the same questions.
- 13:54
They don't really know the answers. They're saying, well... I mean, and, and they, the science museum, for example, is really good at going and trying stuff. So they've got all these big tablets that kids can go and interact with.
- 14:03
But in my personal opinion, a lot of the time that can sort of feel like sticking technology onto the, the th-
- 14:09
Yeah
- 14:09
... rather than it being a, a core part. So some of the stuff we're experimenting with here is, obviously, you've got to take a picture of a statue and you talk to it.
- 14:15
We're commissioning a, a statue to be made that has the technology inside of it and a speaker and phone, um, and microphone, so that you can then talk directly to the statue without having a piece of technology in the way, without it feeling tacked on.
- 14:27
And that's something like with the, the red phone booth you may have seen here, um, on floor three. Uh, there's a, a phone you can pick up inside of a, of K6, um, red London phone, phone booth or British phone booth and talk to an agent, talk to, um, Sir Michael Caine.
- 14:42
So that's trying to put it into the real world a little bit rather than having it go through a screen. Um-
- 14:46
So I almost imagine, like, with vibe coding, like you're where you see kids making their own games.
- 14:51
Yeah.
- 14:51
Like, you imagine a kid go to a science museum, they have their tool-
- 14:54
Mm
- 14:54
... and they just start creating whatever experience they want as they engage. Like that-
- 14:59
Yeah
- 14:59
... it, it just sort of explodes the possibilities.
- 15:01
Yeah. Yeah. I mean, I think even more generally there's a question of what's the vi- vibe coding I feel like still hasn't really gone consumer mainstream.
- 15:09
Right. No, it's still, yeah.
- 15:10
It's like even, even Lovable sort of feels like it's targeted at consumers for building, effectively building B2B SaaS apps. You know? It's like have Supabase and, you know-
- 15:20
Yeah
- 15:20
... standard, standard design components. But I... This is why I really love vibe coding, evening vibe coding events. They feel like the OG hackathons.
- 15:28
Mm.
- 15:28
Because people show up, and they have, they've never even thought about writing code before. Sometimes I go to them and I talk to people, and they're like... I say, "So what's your favorite app?"
- 15:36
Or like, "Who made the app?" They're like, "Wait, people make apps? I thought they were just there on my phone, or on the App Store." Right? They hadn't even thought through the fact that people have to make them.
- 15:44
So that's something that I, I find vibe coding events really fun because people come in and they type in. They have no idea what, they have no idea what a hamburger menu is or, or like a, an accordion is.
- 15:53
So they just say, "I want this and I want this, I want this," and they get something completely wacky because the LLM just says, "Yeah, okay, I'll try it."
- 15:59
Right.
- 15:59
Uh, when if they were talking to a software engineer, I would've said, "Ah, you want one of these and one of these."
- 16:02
Yeah.
- 16:03
Um, so I, I, I think at some point we're probably gonna have, like, what's the Instagram filters moment for vibe coding?
- 16:09
Yeah.
- 16:09
Or the TikTok moment for vibe coding. I think we're gonna have something like that that makes social vibe coding much more. But I d- I don't know what it's gonna look like, but worth experimenting.
- 16:18
Do you see anyone doing it well? Like in that sort of experimentation?
- 16:21
There's, there's Spielwork. There's sort of a, a, an app, a mobile app for vibe coding games, and it's TikTok swiping. Um, and there's, I think, I can't remember what it...
- 16:33
There's a London-based game vibe coding tool that is, again, focused on games. I don't know that games are really the thing because they're quite complex. Um, but it, the- there are a few people experimenting, but I don't think there are that many people really deeply pushing the boundaries of, of what's possible.
- 16:48
They're mostly like Lovable, but on your mobile phone.
- 16:51
Yeah.
- 16:51
Is my opinion. Um, does anybody else have any opin- any sort of, see anyone doing good consumer vibe coding?
- 17:01
Okay.
- 17:03
Vibe coding in what end? In terms of like creating things or copy or for their work or just in any, any sort of sense of-
- 17:10
I guess content creation
- 17:11
... it seems like maybe vibe building or something-
- 17:13
Yeah
- 17:14
... vibe working.
- 17:14
Yeah.
- 17:15
Vibe, you know, interacting with digital systems or firms, whatever it is. Like-
- 17:19
Well, well, well you, you, in your game, you can like spawn things in with voice, right? And that's like a great-
- 17:23
Yeah
- 17:23
... it's not quite vibe coding, but it's still interacting and-
- 17:25
Because I think that's, that's what I... Like vibe coding still means you have an understanding of certain primitives. Like even what you just said-
- 17:31
Mm
- 17:31
... thinking about databases. I think when this goes mainstream, you're not, people aren't thinking about it this way.
- 17:35
Yeah.
- 17:35
They're, they... And that's what's so exciting for me when you think about culture and what you build, is when it gets to a point where just things that us as engineers wouldn't even...
- 17:45
It's not even how we would approach the problem, and that's where some incredible creativity is gonna come.
- 17:48
Mm. So the, the pattern that I think is closest to being a winner in this space is the Facebook Instant Games API, which doesn't exist anymore.
- 17:57
Oh.
- 17:57
Um, they, they deprecated it, but it was in Facebook Messenger. You could play these games, and they, it had these primitives for social gaming. They, they tended to be quizzes or like Fruit Ninja, and you compete with your group chats and things.
- 18:09
I built, I bought for 15 pounds, um, a Fruit Ninja clone off a website, instrumented it with the, the Facebook Instant Games API, which was this beautiful like JavaScript async await.
- 18:21
Had a get user information, get friends, uh, create a leaderboard or get your position on the leaderboard. Um, some very basic data storage, key value storage. And, um- Async await at show an, uh, a rewarded ad and show an interstitial ad.
- 18:37
And so those things allowed you to make very quickly and easily a social graph enabled, ads enabled experience for consumers. So I bought this thing, 15 pounds, instrumented it with Facebook Instant Games API, and went to bed.
- 18:52
The next day, I woke up with 15 million users on this random... Yeah. I mean, it didn't make very much money, but there were like 15 million users in Vietnam.
- 19:01
Because obviously Facebook, they test everything out in, in the sort of lower value advertising regions and then rolls up. But that was an amazing-- Because you've got the social elements of people instantly just sharing it around, and that, I think, is probably the template that's gonna, something along those lines is probably gonna be the thing that wins
- 19:17
on, on the social vibe coding. But I don't know if that actually answers the question.
- 19:21
On the kind of like vibe working, vibe building kind of, uh, you know, thread, I kind of feel frustration whenever I get a, a response back in voice. Maybe it's just me, but like it's almost like the, the, the input, like the information density per second isn't quite high enough.
- 19:36
Like I-
- 19:37
Mm-hmm
- 19:37
... I use a lot of voice out to just get things out of my head, like far, far from the time or whatever-
- 19:41
Or voice input, I guess is the, yeah
- 19:43
... voice input. But then I, I still feel like I need, I don't know, diagrams or text or some like really high density, um, thing back from the system.
- 19:51
So I don't know whether, how you guys think about that, whether I should curiously get other people's views on that, whether they agree with me or disagree. It sounds like there are some nods.
- 19:58
But that's what I find myself leaning towards, where it's like information-rich input, and then I can just speak my thoughts as they come out, and it's almost semantically understood, put into that in-information rich format, and then my intent is like, you know, spawned out and, and across the system itself.
- 20:13
So that's the pattern that at least I'm seeing and feeling. And what I almost want to vibe interact with now, my email or like everything, is that I feel, I feel like that longing.
- 20:23
It's not quite there yet, but I, I, I, I feel like I, this inclination to build in that direction and to try to, um, get in-information rich, you know, get-
- 20:33
Response. Yeah
- 20:33
... get that back, but also sort of like the input being very freeform and just my sort of raw intent.
- 20:39
Yeah. That's something that I feel. I absolutely feel. I, you know, I want to speak, speak easily and quickly, and then receive maybe a little bit of voice. But mostly there's like generated, maybe it's a UI, maybe it's just diagrams, maybe it's whatever app I'm in context of.
- 20:53
But yeah, I want to have parallel input, uh, parallel output, but like single input with my voice.
- 20:59
I can really talk to that. So I, I find that the, what, I don't actually get that much information necessarily from voice, but what I do get is like companionship.
- 21:08
Mm-hmm.
- 21:08
So it kind of like triggers that. It lessens the loneliness feel somehow if I'm like talking to something and I'm getting some information. I don't feel... I might be learning more if I'm looking at diagrams or text, but I, I don't feel as, I feel more motivated to kind of like continue tinkering.
- 21:23
So there's like some interesting modalities where you feel different things if you get the information coming in from all these different modalities, at least I find. So curious to hear how other people think about whether they, they think differently or whether that's something that they also feel as well.
- 21:37
Um, yeah, this is at least my point of view.
- 21:39
Well, the visual cortex is much older than-
- 21:42
Yeah
- 21:42
... voice and, and text or whatever. So-
- 21:43
Yeah
- 21:44
... seeing something is much-
- 21:45
Yeah.
- 21:46
Also in text-
- 21:47
Yes
- 21:47
... when you ask it to be concise, I don't feel offended if it gives me a concise answer. But in speech, if it gives a concise answer, it's like, "Whoa, relax, man.
- 21:54
We're just asking a question here." [laughing] I asked it to be concise, and-
- 21:58
Well-
- 21:59
... it just sounds rude.
- 22:00
That's, that's an interesting-
- 22:01
Yeah.
- 22:01
Um, may-maybe this is possible, maybe it's not. But what does, what does skim listening look like?
- 22:06
Yeah, exactly.
- 22:06
May-maybe, maybe like listening actually should also have two buttons, like forwards and backwards, and I can just tap, tap, tap, tap, tap, go forwards half a sentence until I'm...
- 22:15
I don't know. Maybe that's, maybe we should build that. Who w- who wants to vibe code something with me straight after this?
- 22:21
How would it work? So you'd be, you'd, you'd just easily, yeah, back, back and forward probably with, with audio. Is that kind of what you had in mind?
- 22:27
It's like a speed dial like on a podcast or on 2X.
- 22:30
Yeah. We a- or like the old iPods where you can sort of spin forwards and backwards. That's-
- 22:34
Right.
- 22:35
I don't know. That's, that's probably quite a nice listening interaction. Co- And you sort of, I guess, you want to scroll forwards in concepts, not necessarily in sentences, right?
- 22:43
Yeah.
- 22:43
Like what is the thing that your eyes look at when you're skim reading? It's probably, it's not the sentence structure, it's like the words, I guess.
- 22:49
Yeah, but like you move on to the next thing.
- 22:50
Yeah. Yeah.
- 22:50
You get that topic, move on to the next topic.
- 22:52
Yeah. Yeah. And that's actually, oh, yeah, this is, sorry, this is very interesting and exciting for me. Like, because if, if I'm talking to an agent and it just starts rambling about, you know, sometimes you get back three paragraphs of stuff, and I'm like, "No, not this one."
- 23:04
But I want to interrupt it and say, "Go to the next one." But then it's effectively saying it's a new prompt, right? So then it's effectively saying, "Yes. Okay.
- 23:10
I'll focus on that next one," and it'll write me three paragraphs about the second paragraph. You know, that's not really at all what I wanted. Unless I say, "Be concise," then it says something rude to me. [laughing]
- 23:18
Or could you like make a summary for each paragraph, and then like if you hold the button-
- 23:23
Yeah. Yeah
- 23:24
... it will, it will expand and, or you can like...
- 23:26
Yeah. I think the Claude app has done some interesting stuff on the voice interactions because they, they show you something different to you here, and they show the higher level sections, and then it goes into each one, and you can tap on them.
- 23:40
So that's, I guess, getting a little bit closer to the... But I think that also means you're not interacting as though you would with a, a human conversat- I don't know.
- 23:48
Yeah. I, I, I'm thinking about that. Like why do we not have this issue when we're talking to humans, right? Like how do we-
- 23:56
Mm.
- 23:57
Well, there's so many other cues.
- 23:58
Yeah.
- 23:58
Like right there in-
- 23:59
There are visuals
- 24:00
... there, there vision is like a cue. We have a sense. There's just so many other things.
- 24:02
I think there are visual cues as well-
- 24:04
Yeah
- 24:04
... that you're lacking on voice. So that's something, 'cause it's like if I'm, as, as was mentioned before, it's like if there's an agent response coming in audio, I don't know how long that's gonna be.
- 24:12
Mm.
- 24:12
I just think what one was mentioned, like is it gonna be like, you know, a minute or is it gonna be ten seconds? I almost wanna know, and if I know it's like really long and waffly, I'm just gonna like cancel it.
- 24:21
I have, I have some other-
- 24:23
But also you can tell when I'm about to interrupt you, so you go faster, and you like move maybe... You, and you can tell if people are listening or not.
- 24:30
It's more like a phone experience, right?
- 24:32
Yeah. Yeah.
- 24:35
Because basically-
- 24:35
I guess, oh, there's another thing which is interesting here, which is the interrupting. Sometimes I don't want to interrupt. I just wanna say, "Yeah, yeah, yeah," or, "Oh, but no, go back."
- 24:43
And like, it's like you c- you're always listening for the, "Uh-huh. Yeah, yeah, yeah." But you can't do that with an agent.
- 24:49
I think, I sort of, I'm building sort of, you know... I mentioned this to Joe before, but, but briefly, she, I showed you this yesterday, but briefly, the, the ability to sort of interact with a, sort of a PS5 game, and also create what you want in that world as you're playing it.
- 25:02
So you're, you know, you're on like a, you know, a shooter battle. I'm playing you. We're trying to like, you know, kill each other, you know. Apologize language, but like if I'm, if I'm like trying to be creative and, you know, generate a new getaway vehicle or, or a helicopter, I could just sort of say that experience
- 25:18
and have that appear in the game straight away, and now I can sort of fly away. So that's sort of the work that I'm doing and how I met Joe, actually.
- 25:24
But I think the, the challenge that I faced is that relying on just the, like, voice interruptibility is quite unreliable. So I've just got round this with just a simple whisper flow type-
- 25:34
Push to talk type. Yeah
- 25:35
... where you push to talk, and then ho-hold to talk, and then let go to finish, which it, it augments, like, purely this audio stream with some other cue.
- 25:43
And in the same way that I think you almost need some very light nudge interface on top of the audio that you're receiving. And maybe as you say, like, you know, you're saying something, and then you see a little ye- um, sort of like circle appearing, being like the agent wants to ask you a question.
- 25:59
That would be an interesting experience to feel, right? You're, you're, you're not being interrupted-
- 26:02
Mm.
- 26:03
... but you're being kind of like, you know, the agent wants to talk, and then maybe you either stop or say, "Okay. What, what, what, what idea do you have?"
- 26:09
Because I even can feel you doing that to me right now, Joe.
- 26:12
Yeah.
- 26:12
You, you want, you want to say something.
- 26:14
Well, I'm just getting loads of ideas. This is great.
- 26:16
Yeah. I just sort of like... I, um, but we're, we're almost like having an information, you know, communication on that level.
- 26:21
Yeah.
- 26:21
But in audio, if I'm just listening to audio, I'm not, I don't think I'd, I'd know that as if I was an agent or if I was just like, just listening to the audio, not looking at the, at the, at the video stream.
- 26:30
Okay.
- 26:30
I'll let you talk. Sorry. [laughs]
- 26:32
Well, I'm just imagining, or I'm j- just imagining on that point that the agent wants to respond to you. It could be showing, "I want to interrupt and t- tell you about this thing or this one."
- 26:39
And, and then suddenly that becomes what we've just done, but with even more context than doing it with a human because it's signaling the topic it wants to talk about.
- 26:48
Yeah.
- 26:48
Yeah.
- 26:48
And-
- 26:49
I have just, just one question regarding the product. Because regarding this topic, uh, you need some sort of, uh, um, tool talk, tool calling in the background to be able to-
- 26:59
Mm
- 26:59
... actually understand that I want to interrupt. Uh, and, uh, and, uh, how would that, uh, tool call be, uh, handled? Will you, will you sort of be streaming the audio directly to back of the RTC to the, to the client, or through a back channel and then, uh, getting some information?
- 27:17
Do you update the, the system prompt? Or-
- 27:19
Well-
- 27:20
How can you work that?
- 27:21
Obviously, this hasn't been, as far as I'm aware, hasn't been built yet. The way I would probably approach it is looking at the transcript, and just keep analyzing the transcript over and over again, and say, "Do you have anything to, to add?
- 27:30
Do you have anything to add? Do you have anything to add? Or what would you add?" Um, rather than being p- a tool call or being part of it, I'd do this as an asynchronous looking at the transcript.
- 27:39
Are you, are you changing original prompt? Because the original prompt has sort of a plan. You want to do this. Um, and halfway you sort of change the plan, and you go in and sort of...
- 27:49
How do you steer the original a- agent then?
- 27:52
Well, this is actually an interesting thing with agents. They... Often we see agents as things that you can't really interact with the internals of. But effectively, an agent is, is the sort of logic that says, "What's the next message?"
- 28:05
But it's also a transcript, and that transcript of the conversation is completely malleable. So maybe some other thing we could experiment with is actually allowing the ability to do the interruptions where it's talking, and then I can say, "Uh, uh, uh, yeah, yeah, yeah."
- 28:18
And currently, if, if I have the, it, the way most agents platforms will work is it'll generate its full text thing and then start generating the audio. And then if I talk, even while it's partway through, it's still reading out the audio, uh, the, the text.
- 28:32
It will just append my message to the end of that full message. But we do have the timestamps. We know how far through the, the audio is played. So we could actually just go and edit the transcript that's coming back and say, "Well, no, they interrupted at this point, so we're gonna forget that the LLM even generated
- 28:45
more text."
- 28:46
Yeah. So if you sort of, you have, you have the system message and then the user message.
- 28:49
Yeah.
- 28:50
And sort of you need a, sort of a second type of user message that you are pro- uh, are passing in. I, I certainly would take the, of course, it's a user message, but-
- 28:59
Yeah
- 28:59
... it's also some other kind-
- 29:02
Yeah
- 29:02
... of steering here. Yeah.
- 29:04
Yeah. I mean, why don't you... Let's go to the ElevenLabs booth at the expo floor and just vibe code something and see. Like this, this is great. I, I can't wait.
- 29:11
Now it's getting easier and easier to make tools, and there's so many different ways to execute your ideas. And it seems like the main kind of differentiator now is getting your word out there.
- 29:21
Like, how did you approach making those videos, and like, um, how long did it take you for the statue video-
- 29:27
Mm
- 29:28
... to make it?
- 29:29
That's a good question. Totally se- totally separate.
- 29:31
Separate, yeah.
- 29:32
Um, so this is sort of the inspiration of this ElevenHacks thing. Um, I learned to make videos. I'm not particularly good at it, you know. It's like still relatively janky.
- 29:43
Um, I learned to make videos through doing other, like, politics related campaigning stuff before I worked at, before I even knew ElevenLabs existed. Um, the, it turns out that with videos, random things go viral, or with content, random things go viral and random things don't.
- 29:58
Like, things I think are gonna go viral don't, and the things that do, do. Um,
- 30:03
I think it's basically about practicing. Like, uh, it's, it's, editing a video, I find, to, it's like the t- 80/20 rule. Editing a video to the standard of the statue app is actually quite relatively easy.
- 30:16
But then to go beyond that, you suddenly-- 'Cause I edited that on my phone. The editing itself took about 20 minutes, 25 minutes. Going beyond the quality of that suddenly means thinking about using a desktop editing tool, which suddenly makes everything...
- 30:31
Even using CapCut on desktop, to me, is like three times harder than using it on my lap- on my phone. And it's also three times more expensive, the subscription on the laptop than on the phone.
- 30:39
So I think a lot of it is just- Doing stuff and trying and iterating. Big things that I found, adding captions really help, um, to the video. Having a hook in the first...
- 30:50
You can, I mean, you look at your various platform, um, uh, analytics once you've posted a few videos. My videos tend to get between 6 and 12 seconds is the aver- the median view time, and then people drop off.
- 31:04
So you need to get your hook in there because most people are gonna drop off if they don't buy the hook. Um, so that's, that's an important piece, front-loading the, the, like, interesting piece.
- 31:15
Uh, adding music makes a massive difference, and this is something that was, was much harder, but now with ElevenLabs music generation, I will make the video and then add...
- 31:26
Sometimes I'll make the video with the narrative and, and everything, and then just experiment with completely different genres until I find music, and I'll just put the music on.
- 31:32
Does it work? Yes. No. And then you can edit the, the different sections so that it times up with the sections of the speech, so you don't need to, like, figure out mu- music first, find a piece of music, and then match your speech to it.
- 31:44
The other way, sometimes I will have a vibe I want to get across, and I will generate the music first and then figure out what's my speech that matches that vibe.
- 31:54
Um, so it might be like an, uh, uh, an excited theme or a, a... I, I actually chose the music for, for the statue app before I, um, before I made the, the video itself because I thought this is a fun piece of music for a statue type thing.
- 32:10
It's like a bit of im- imperial, outside the British Museum. It sort of made, made sense as a, a- an attention grabber. So, but I think music is a massive thing that people underrate, um, 'cause you just put it...
- 32:20
It's relatively quiet, but it makes a massive difference to the feeling.
- 32:23
So you just did it on CapCut, and did you think-
- 32:26
It's literally on my mobile phone. I've got-- I borrowed my wife's lapel mic. [laughs] Bluetooth la- lapel mic, which cost 200 quid from DJI, um, which makes the audio much better.
- 32:35
And then, uh, yeah, just edited on CapCut. Super, super simple.
- 32:39
So you didn't think about, like, this is the shot I want, and then, like, I, I wanna take this photo or...
- 32:45
Um, I mean, the, the, the f- the one out the front of the British Museum, yes. I... Thank you. Um, yes, I did want that one. But the... And it was actually the second time I recorded the video.
- 32:57
I, I went once and got some stuff that was, like, a bit boring, and then I went back and just took a bunch of silly photos, and that's what ended up being the video.
- 33:04
So yeah.
- 33:06
Very good.
- 33:06
Thanks.
- 33:09
Yeah.
- 33:10
Yeah. Cool. Thank you very much.
- 33:12
Thank you. [upbeat music]
Creation needs distribution primitives
Reeve’s historical template is Facebook Instant Games in Messenger: small games such as quizzes or Fruit Ninja, played competitively within group chats. He describes that experience as deprecated; this does not establish that the entire Instant Games API has ceased to exist. The useful pattern is the compact set of social and commercial primitives exposed through JavaScript’s
async/awaitstyle.Together, these let a small game participate in an existing social network instead of requiring its developer to build accounts, sharing, competition and advertising from scratch.
Reeve says he bought a Fruit Ninja clone for £15, integrated Facebook Instant Games, and woke the next day to 15 million users in Vietnam. He reports little revenue, and attributes the geography to Facebook testing in lower-value advertising regions. The anecdote illustrates why he sees social connections and immediate sharing as promising ingredients for consumer vibe coding: producing a creation is only part of the experience; putting it in front of friends gives it somewhere to go.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.