I've done a lot of stage demos. I've done pre-recorded ones, I've done live ones.
I prefer live ones, by a significant margin, even for non-deterministic AI things. In general I find live ones better, but it does depend on your circumstance. It's worth asking yourself three questions when you're doing a demo:
What is the outcome when it goes great?
What is the outcome when it goes wrong?
What percentage of the time do each of those occur?
With a pre-recorded demo, when they go great it's usually a 9/10. Things are solid, but they aren't adaptive to the situation, which means you might have lost an opportunity to insert a quip from something that happened earlier. When they go wrong though, it's a 0/10. You've told people that you don't have enough faith in your product to show it live - why should they have faith in it?. Now you might say that the demo goes right 100% of the time(ish), but there's also a human element to it. Someone might play the demo early, someone might not act in time with the live demo, any of these can give away that it's pre-recorded. I'd put their success rate at 95%.
With a live demo, when they go great they're 10/10. You can adapt them, you can show it live, things react instantly instead of slightly-out-of-sync due to scripted content. It's incredible. When they go wrong, it's a 3/10. You've got egg on your face. People make fun of you after. But they know at that point that you ran a live demo. That does count for something. We know Zuck did that demo live. They do fail a bit more, especially in the age of AI. 10% failure rates are pretty good for live demos. To me though, the higher failure rate is worth it for the better PR if you do fail.
100% agree. It's not a BINARY choice though! I prefer to do the demo live, and if it fails, have a one-push switch over to a recording and quickly declare that live projects are a work in progress. I think this is ultimately how you show both strength of live demo preparation, but can humbly recover the show if the live version fails by switching to recorded.
But, NEVER TELL ME IT'S A LIVE DEMO WHEN ITS PRE-RECORDED. You will lose my business forever if I catch it.
I don't agree that a bad live demo is a 3/10. The recent demo showed me conclusively that the Meta Ray Bans are not at the quality level where I would buy them, especially not for $800. That's pretty much a 0/10 in my book. Since it was live usage, it's indicative of the real quality of the product.
That live demo, despite the failures, unexpectedly made Meta much more likable to me than (say) the sterile and cringy prerecorded Apple events, which are just completely off-putting.
Furthermore, look at how the prerecorded Apple Intelligence demo at WWDC 2024 turned out, where more than a year later we still don’t know if and when its most salient features will see the light of day. Whereas with the Meta demo, we can assume that they were reasonably confident it would work on stage, and the failure may really just be due to the flaky wifi at the venue.
IMO a perfect prerecorded demo would have to be shockingly good to get better than like a… B- for me. It’s just too easy to cut or reshoot the interaction to be perfect.
I'm a sales engineer and demoing is one of the core parts of the job. I think this is applicable if it's an internal demo or a demo to a large audience, i.e a conference where you have a very fixed window of time to present.
I would never recommend doing this in a sales context though, where you're selling to prospects. Mainly because it gives you no flexibility to pivot and adjust the content based on what they're interested in, and also because it gives off the impression you don't have a lot of faith in your product to perform live.
I'd still 100% recommend having pre-recorded backups however. I was once doing a demo to our largest customer in region and the second we joined the call our core product had a SEV0. I was very relieved to have something to show while the engineers fixed the issue.
If you pre-record your live demo then it isn't a live demo...
What I dislike is presenting something as if it is live when it isn't. Demos are fantastic things and it is even perfectly fine to have imaginative aspects to them. But there's a division line between demoing and lying. Take Rabbit R1 as an example[0]. This whole thing was faked yet it was presented as a live demonstration, not an illustration. In other words, fraud.
The difference is if you're trying to show your customers (or potential customers) what you're imagining the device works like vs what the device actually works like. The difference is how you're communicating the difference to them.
I'm not sure why this is a contentious opinion, but just don't lie to people?
Absolutely. I don't mind a recording, as long as it's presented as a recording and anything misleading is disclosed. If you say you're going to demo something, demo it. We have enough erosion of truth in this world already -- don't contribute to it!
I do live demos at work all the time and the best part is that when someone asks "can it frobulate with encapulation?" I can actually press the frobulate button with encapulation enabled and show them. Can't do that with a recording.
I had a professor who had taken courses from Paul Halmos, who authored one of my favorite math books, Naive Set Theory. I regarded the book so highly that I was somewhat star struck to learn my professor had taken courses from him.
I asked my professor what it was like, was he an incredible professor? He told me the lectures were useless! I was shocked and he said they were so well prepared they were effectively a recording of the material - and no advantage over reading the book. At that point I reflected and realized that, by far, my favorite professor was one who constantly made mistakes while solving problems on the board. He'd have to notice something wrong and track down the error, and it was wonderful and highly instructional to watch the process.
I can even pause the replay mid-demo, take over, and resume again later. People are always impressed at how flawless my live demos are, and with some judicious audience participation, the illusion is perfect.
I love this idea. I remember giving a live demonstration of a configuration management tool (SaltStack circa 2012) and some unknown networking conflict between my Virtual Machines and the local wifi subnet caused my commands to timeout. I was too junior to resolve it live and was left with boring slides explaining what ought to be happening.
Luckily it was a small group of devs and not a large venue. Still embarrassing...
Your comment only makes sense if you think that the purpose of the demo is that the presenter gets street cred for doing something hard. If, like me, you think that the purpose of the demo is to show or teach the audience something, then it doesn't matter whether it's live or not, all that matters is that it's done well.
It's like expecting reality shows to be a 100% unfiltered feed of real life. Their purpose is to entertain, not to document.
> Your comment only makes sense if you think that the purpose of the demo
Don't contort my words to justify your lying.
The purpose of a demo is to demonstrate things.
The purpose of a live demo is to demonstrate how things will work in the user's actual experience.
You can have pre-recorded demos and there's nothing wrong with that. The only issue is calling a pre-recorded demo "live". That's when you cross the line into deception.
Or as I said before
>>> Seriously... just call it a demo or pre-recorded. As soon as you call it "live" you've crossed into the territory of fraud.
You don't have to make believe I don't understand what a demonstration is to get there.
> Their purpose is to entertain, not to document.
And now I believe it is YOU who do not understand what a demo is. A demonstration is not entertainment nor documentation, although it can incorporate these elements. A demo is an illustrative example. Do we need to go to Merriam Webster here?
Oh if we're going to be pedantic, I can be pedantic. A "demo" is showing something, and "live" is showing it while both the demonstrator and the audience is in a room together. Therefore, this is a live demo, QED.
I'm not 100% sure, but I think this tool just types the commands. So it avoids typos and you can just focus on what you're saying as you type nonsense. Sounds like it actually does run your pre-written commands. From the readme:
> What's more, It's a Live is actually running the commands you're typing, so you have full interoperability with other
programs.
It seems like you're misunderstanding what parent meant by "prerecorded" - not a screen recording, just pre-writing of commands to be executed during the demo. Would you consider it "deception" to hit up arrow a few times in a terminal during your demo to execute a command from shell history? This is effectively the same. Take a look at the linked repo, it's very clever.
The original link for itsalive says it works by making it not live and the op says just as much.
This is very different from practicing a script and you actually doing all the actions in real time. You're allowed to rehearse for your live performance but unless you're doing it live then it's not live.
People also hate lip syncing because it's a deception of being live.
And yet, ~all singers lipsync, because the point isn't to show off what the artist can do vocally every single night, but to entertain the audience, which lip syncing achieves.
Are you assuming or do you have evidence to back up that claim. Because I actually go outside and watch musicians perform. Are you telling me they fake the errors too? LOL
> Every time you press a key, It's a Live will write one character from the file into the terminal, making it look like you're typing every single command with the practiced ease of a consummate professional.
It's like pressing next on a slide. That's not live. That's lip syncing.
Good god, this isn't that hard man. You're fine to lip sync if you call it lip syncing. Just don't claim you're actually singing when you're not and you're all good. Why is this so difficult to grasp?
The linked software only seems to type the 'correct' preknown letter anytime you press any key on your keyboard.
So the actual 'demo' part is actually live. The only thing that's fake is the typing, which could be replaced honestly with copy & paste without changing anything except how exciting it is to see the speaker pretend to type at 100wpm.
So it's lip syncing. That's not live. People aren't that fond of lip syncing. At least when they find out that it's happening. They get upset because they've been deceived
I’d say it’s more like reading off a teleprompter. You didn’t come up with the speech on the fly, and you didn’t memorize it, but you are filmed as if both things might be true.
With a teleprompter (or queue cards or notes) you are still speaking the words.
But with this system you are not. So it is more like a slideshow where your voice is pre-recorded per slide and you are pretending to speak.
I would still call a teleprompter a live setting. I wouldn't be surprised if Zuck had one. People can still stumble and maneuver and adapt in real time. Just like in a play. Rehearsal does not make the act not live. But if we want to get more fine grained there is a difference between a live demo and an on-the-spot demonstration. Think of it this way, people don't knock politicians for using teleprompters because they are lip syncing the words. They knock them for being overly rehearsed and not off the cusp. The complaint is actually beyond the teleprompter, the teleprompter is just a visual indication of this. But then imagine how people would respond if a politician went up on stage and pretend to actually speak but the words coming out were pre-recorded.
Or maybe we can meet more in the middle with something like autotune? I still don't think it is there because you'll still need to say the right words for autotune to work but even still people generally don't like that and feel like it is deceptive. So I might be being strict here, but we need to at least recognize that all these concerns arise from some form of deception. Where the expectation of what's happening doesn't meet reality.
No one will be upset if they never find out, but the real question is how they respond when they do. Does the person you're demoing to feel cheated, even a little, if you used a tool like this? I'd wager they would because they thought you were doing all the actions there in front of them.
I actually really appreciate that. Thanks. And I do value your perspective. I know I'm passionate here but you did make me think about the issue differently and pushback is always good.
If you're launching some fragile and finicky AI thing before a crowd like you're Steve Jobs, but you're not Steve Jobs, then, sure, pre-record it.
But when the standards are lower, people appreciate the live demo. Videos are much less compelling, and they could've watched that at home in a few minutes, rather than fly out to you.
In grad school -- in a lab where, instead of "publish or perish", our mantra was "demo or die" -- we were giving demos most weekdays, usually to only a few people at a time. Always live. I think once I had to play a video of someone else's stuff, and it fell flat.
My best demo story is when I was showing a new thing (just a class project), to a few people crowded around a computer in my lab area at an open house event, typing into my software agent in a chat window (pre-LLMs), ad-libbing and speaking aloud what I was typing, "I'm interested in... thinks Microsoft and... thinks Oracle..." Then I turned my head around to look at the people behind me, to see that they were following, and I swear to gosh, Larry Ellison himself was standing in the doorway of my lab area, giving me this fiercely aggressive stare. Whether he was glaring because I said Oracle or Microsoft, I do not know.
My worst demo story was at the very start of my demo career. I had just arrived on campus, and I was borrowing a computer in an open area. A research scientist comes up, and sees on my screen this Web-scraping personalized newspaper thing I had made. So I'm showing it to them, and they're interested, then all of a sudden they say, sadly, "Mother Theresa died..." Which was what my "demo" was showing.
So maybe don't include an uncontrolled live data feed of depressing or distracting news in your demo.
Also, implementing early software agents in Java apparently summons Larry Ellison.
I agree that you should always have a pre-recorded version as backup, but live demos communicate a confidence in your product, and that can be worth something. Whenever I see a pre-recorded demo I wonder how many takes it took them, or how many pauses were taken out in editing, etc.
It helps if you have a master of ceremonies capable of running three rings simultaneously. Jobs could do it. Looking at what Apple is doing, they feel Cook is not. You lean into your strengths and avoid your weaknesses. Meta should realize that theZuck is not Jobs and is much closer to Cook
This article reminded me of a talk I gave a couple times that went very smoothly because I pre-recorded the typing parts. Something like this:
- record the screen as you type out all of your code as you would do live
- don't need the audio, just the video
- edit it down to remove typos and pauses
- speed it up to a nice fast typing pace
Now you can drop the video into Keynote or whatever, and Play/Pause the video as you speak.
A crucial part of this is that the video has effectively no pauses. Maybe just a few short ones between sections, to leave time to hit Pause before the next keystrokes appear.
Instead of a 30 minute video that I had to perfectly keep pace with, it was like 5 minutes of rapid typing that I could speak along with and pause when it was time to explain something.
At this point in time, it is absolutely trivial to pre-record whatever you'd like, and it will look glorious, and it doesn't need any basis in reality.
As a customer, when you show me a video, I default-assume you're lying.
Also, as a presenter: If your live demo fails and you can't turn that into a win, maybe you're not good at presenting. Your audience wants you to succeed, it takes a minimal amount of charm to make it work. (Heck, even Mark turned this into a win, and he's not exactly a charm bomb)
Thought at first this was maybe just being applied to external product demos, but this reads like dev demos, too:
> Since most of the time you're demoing something you previously worked on, this prevents you from working on current stuff right up to the demo, because you're worried you might break your local setup for the demo by installing different node modules, switching to different Git branches, etc.
This isn't the case for me generally... probably more of problem for web-apps things I guess.
I keep multiple workspaces locally for the multiple things I'm working on, but that hasn't been my pain point. Usually it's that I'm demoing a work in progress (not my choice generally) that has rough edges.
I enjoy the internal demos where questions can be asked about doing XYZ, and that can be shown if it's quick. Recordings tend to take away from that, and be less memorable/engaging I guess.
First time I recorded a demo was when it became too hard to return things to a known good state after every demo. In this case that meant multiple RAID arrays, 3 servers, an application with DB, and monitoring software.
All so that I could demonstrate that when I turned off a RAID controller a red light came on and the application just…continued.
So much work for something largely unspectacular meant that a recorded demo that could be run anytime from a laptop was far more appealing. My junior colleagues were now doing demos they otherwise wouldn’t, and the seniors turned off the sound and edited the video into snippets they could introduce and talk around
A company I worked for switched over to pre-recorded demos and everyone talked about how clever it was for the first few larger audiences. Then they made a mistake and replayed a clip during the same session and the audience chat blew up. You could see a dip in new users for days after the demo.
Apple can pull it off, but if you’re at a mid market tech company you just sunk like weeks people’s time into choreographing a live movie. Super stressful time sync.
IMO live demo is better in some cases, but if there are complicated or time-consuming steps where the prerequisite can not be met quickly, a recording would be better
> So here’s the story behind why yesterdays live #metaconnect demo failed - when the chef said “Hey Meta start Live AI” it activated everyone’s Meta AI in the room at once and effectively DDOS’d our servers
> That’s what we get for doing it live!
And if you believe that, I've got a bridge to sell you...
It makes me question the rest of this article when the author takes that at face value. A DDOS made your LLM say the same thing over and over? Yeah, I don't think so. Also, this new tech, how many device that support “Hey Meta start Live AI” even exist or were "listening" to the stream.
This is about a valid as Zuck saying that it was "Wifi" issues. How often do you have Wifi issue where you can talk to remote servers just fine?
or not like the recent FB AR glasses. if your live demo doesn't work in rehearsal, there's a good chance it's not going to work during the actual presentation. there's a reason the phrase "demo hell" is well understood.
But the problem is that the product is buggy. The question then is "can you fix those bugs before products ship."
I'm not sure what the problem is here. Customers don't want buggy products. Live demos are more informative to users who can't tell what the actual experience will be like vs what the envisioned experience is.
I'm sorry, but as a customer I don't care what's in a developer's dreams, I care what I can buy.
I always do half and half, a mostly live demo with some cooking show style start and results. It does depend on your audience, and there is charm in a live demo.
No one wants to see a terminal, that's for sure, so don't display code or start pecking away at a terminal unless you really think you'll need it.
I've done a lot of stage demos. I've done pre-recorded ones, I've done live ones.
I prefer live ones, by a significant margin, even for non-deterministic AI things. In general I find live ones better, but it does depend on your circumstance. It's worth asking yourself three questions when you're doing a demo:
What is the outcome when it goes great?
What is the outcome when it goes wrong?
What percentage of the time do each of those occur?
With a pre-recorded demo, when they go great it's usually a 9/10. Things are solid, but they aren't adaptive to the situation, which means you might have lost an opportunity to insert a quip from something that happened earlier. When they go wrong though, it's a 0/10. You've told people that you don't have enough faith in your product to show it live - why should they have faith in it?. Now you might say that the demo goes right 100% of the time(ish), but there's also a human element to it. Someone might play the demo early, someone might not act in time with the live demo, any of these can give away that it's pre-recorded. I'd put their success rate at 95%.
With a live demo, when they go great they're 10/10. You can adapt them, you can show it live, things react instantly instead of slightly-out-of-sync due to scripted content. It's incredible. When they go wrong, it's a 3/10. You've got egg on your face. People make fun of you after. But they know at that point that you ran a live demo. That does count for something. We know Zuck did that demo live. They do fail a bit more, especially in the age of AI. 10% failure rates are pretty good for live demos. To me though, the higher failure rate is worth it for the better PR if you do fail.
100% agree. It's not a BINARY choice though! I prefer to do the demo live, and if it fails, have a one-push switch over to a recording and quickly declare that live projects are a work in progress. I think this is ultimately how you show both strength of live demo preparation, but can humbly recover the show if the live version fails by switching to recorded.
But, NEVER TELL ME IT'S A LIVE DEMO WHEN ITS PRE-RECORDED. You will lose my business forever if I catch it.
I don't agree that a bad live demo is a 3/10. The recent demo showed me conclusively that the Meta Ray Bans are not at the quality level where I would buy them, especially not for $800. That's pretty much a 0/10 in my book. Since it was live usage, it's indicative of the real quality of the product.
That live demo, despite the failures, unexpectedly made Meta much more likable to me than (say) the sterile and cringy prerecorded Apple events, which are just completely off-putting.
Furthermore, look at how the prerecorded Apple Intelligence demo at WWDC 2024 turned out, where more than a year later we still don’t know if and when its most salient features will see the light of day. Whereas with the Meta demo, we can assume that they were reasonably confident it would work on stage, and the failure may really just be due to the flaky wifi at the venue.
Yeah I agree. But I’ll split it slightly.
From a trust perspective I want a real demo. Technical team, deep dive, concept sold now talk me through it type stuff.
From a “show non-tech executives the art of the possible as part of 2026/2027 planning” I want a recording.
Failure in the latter isn’t a 3/10, it’s a -10/10.
IMO a perfect prerecorded demo would have to be shockingly good to get better than like a… B- for me. It’s just too easy to cut or reshoot the interaction to be perfect.
Given Tesla’s track record of completing faking their pre recorded demos, I don’t trust them at all. As a consumer live demos all the way
In another world we would call that fraud. But here it's business as usual.
I'm a sales engineer and demoing is one of the core parts of the job. I think this is applicable if it's an internal demo or a demo to a large audience, i.e a conference where you have a very fixed window of time to present.
I would never recommend doing this in a sales context though, where you're selling to prospects. Mainly because it gives you no flexibility to pivot and adjust the content based on what they're interested in, and also because it gives off the impression you don't have a lot of faith in your product to perform live.
I'd still 100% recommend having pre-recorded backups however. I was once doing a demo to our largest customer in region and the second we joined the call our core product had a SEV0. I was very relieved to have something to show while the engineers fixed the issue.
If you pre-record your live demo then it isn't a live demo...
What I dislike is presenting something as if it is live when it isn't. Demos are fantastic things and it is even perfectly fine to have imaginative aspects to them. But there's a division line between demoing and lying. Take Rabbit R1 as an example[0]. This whole thing was faked yet it was presented as a live demonstration, not an illustration. In other words, fraud.
The difference is if you're trying to show your customers (or potential customers) what you're imagining the device works like vs what the device actually works like. The difference is how you're communicating the difference to them.
I'm not sure why this is a contentious opinion, but just don't lie to people?
[0] https://www.youtube.com/watch?v=22wlLy7hKP4
Absolutely. I don't mind a recording, as long as it's presented as a recording and anything misleading is disclosed. If you say you're going to demo something, demo it. We have enough erosion of truth in this world already -- don't contribute to it!
I do live demos at work all the time and the best part is that when someone asks "can it frobulate with encapulation?" I can actually press the frobulate button with encapulation enabled and show them. Can't do that with a recording.
I had a professor who had taken courses from Paul Halmos, who authored one of my favorite math books, Naive Set Theory. I regarded the book so highly that I was somewhat star struck to learn my professor had taken courses from him.
I asked my professor what it was like, was he an incredible professor? He told me the lectures were useless! I was shocked and he said they were so well prepared they were effectively a recording of the material - and no advantage over reading the book. At that point I reflected and realized that, by far, my favorite professor was one who constantly made mistakes while solving problems on the board. He'd have to notice something wrong and track down the error, and it was wonderful and highly instructional to watch the process.
I prerecord all my live demos, and wouldn't have it any other way:
https://gitlab.com/stavros/itsalive
I can even pause the replay mid-demo, take over, and resume again later. People are always impressed at how flawless my live demos are, and with some judicious audience participation, the illusion is perfect.
I love this idea. I remember giving a live demonstration of a configuration management tool (SaltStack circa 2012) and some unknown networking conflict between my Virtual Machines and the local wifi subnet caused my commands to timeout. I was too junior to resolve it live and was left with boring slides explaining what ought to be happening.
Luckily it was a small group of devs and not a large venue. Still embarrassing...
That's literally 100% of all live demos. Something always happens.
Seriously... just call it a demo or pre-recorded. As soon as you call it "live" you've crossed into the territory of fraud.
Your comment only makes sense if you think that the purpose of the demo is that the presenter gets street cred for doing something hard. If, like me, you think that the purpose of the demo is to show or teach the audience something, then it doesn't matter whether it's live or not, all that matters is that it's done well.
It's like expecting reality shows to be a 100% unfiltered feed of real life. Their purpose is to entertain, not to document.
> purpose of the demo is to teach
Then that's documentation with extra steps.
Yep.
The purpose of a demo is to demonstrate things.
The purpose of a live demo is to demonstrate how things will work in the user's actual experience.
You can have pre-recorded demos and there's nothing wrong with that. The only issue is calling a pre-recorded demo "live". That's when you cross the line into deception.
Or as I said before
You don't have to make believe I don't understand what a demonstration is to get there. And now I believe it is YOU who do not understand what a demo is. A demonstration is not entertainment nor documentation, although it can incorporate these elements. A demo is an illustrative example. Do we need to go to Merriam Webster here?Oh if we're going to be pedantic, I can be pedantic. A "demo" is showing something, and "live" is showing it while both the demonstrator and the audience is in a room together. Therefore, this is a live demo, QED.
We aren't being pedantic here.
Right, because watching a video of me doing a magic trick is the same thing as me performing it for you. You're incredulousI'm not 100% sure, but I think this tool just types the commands. So it avoids typos and you can just focus on what you're saying as you type nonsense. Sounds like it actually does run your pre-written commands. From the readme:
> What's more, It's a Live is actually running the commands you're typing, so you have full interoperability with other programs.
If it is live, it is being done in real time.
If it is pre-recorded it was done previously.
If you call a pre-recorded demo "live" then it is fraud.
What more is there to get?
It seems like you're misunderstanding what parent meant by "prerecorded" - not a screen recording, just pre-writing of commands to be executed during the demo. Would you consider it "deception" to hit up arrow a few times in a terminal during your demo to execute a command from shell history? This is effectively the same. Take a look at the linked repo, it's very clever.
The original link for itsalive says it works by making it not live and the op says just as much.
This is very different from practicing a script and you actually doing all the actions in real time. You're allowed to rehearse for your live performance but unless you're doing it live then it's not live.
People also hate lip syncing because it's a deception of being live.
And yet, ~all singers lipsync, because the point isn't to show off what the artist can do vocally every single night, but to entertain the audience, which lip syncing achieves.
Live music of professional, popular singers is mostly lipsynced.
Are you assuming or do you have evidence to back up that claim. Because I actually go outside and watch musicians perform. Are you telling me they fake the errors too? LOL
The inputs are pre-recorded but the outputs aren't.
What more is there to get?
They aren't.
RTFM
It's like pressing next on a slide. That's not live. That's lip syncing.Good god, this isn't that hard man. You're fine to lip sync if you call it lip syncing. Just don't claim you're actually singing when you're not and you're all good. Why is this so difficult to grasp?
I know how it works, I wrote the fucking thing.
The problem is you not knowing what "live demo" means
The linked software only seems to type the 'correct' preknown letter anytime you press any key on your keyboard.
So the actual 'demo' part is actually live. The only thing that's fake is the typing, which could be replaced honestly with copy & paste without changing anything except how exciting it is to see the speaker pretend to type at 100wpm.
So it's lip syncing. That's not live. People aren't that fond of lip syncing. At least when they find out that it's happening. They get upset because they've been deceived
I’d say it’s more like reading off a teleprompter. You didn’t come up with the speech on the fly, and you didn’t memorize it, but you are filmed as if both things might be true.
I disagree.
With a teleprompter (or queue cards or notes) you are still speaking the words.
But with this system you are not. So it is more like a slideshow where your voice is pre-recorded per slide and you are pretending to speak.
I would still call a teleprompter a live setting. I wouldn't be surprised if Zuck had one. People can still stumble and maneuver and adapt in real time. Just like in a play. Rehearsal does not make the act not live. But if we want to get more fine grained there is a difference between a live demo and an on-the-spot demonstration. Think of it this way, people don't knock politicians for using teleprompters because they are lip syncing the words. They knock them for being overly rehearsed and not off the cusp. The complaint is actually beyond the teleprompter, the teleprompter is just a visual indication of this. But then imagine how people would respond if a politician went up on stage and pretend to actually speak but the words coming out were pre-recorded.
Or maybe we can meet more in the middle with something like autotune? I still don't think it is there because you'll still need to say the right words for autotune to work but even still people generally don't like that and feel like it is deceptive. So I might be being strict here, but we need to at least recognize that all these concerns arise from some form of deception. Where the expectation of what's happening doesn't meet reality.
No one will be upset if they never find out, but the real question is how they respond when they do. Does the person you're demoing to feel cheated, even a little, if you used a tool like this? I'd wager they would because they thought you were doing all the actions there in front of them.
After reading your perspective, I'm now inclined to agree. Not much more to say, but I didn't want you to think the words were wasted.
I actually really appreciate that. Thanks. And I do value your perspective. I know I'm passionate here but you did make me think about the issue differently and pushback is always good.
If you're launching some fragile and finicky AI thing before a crowd like you're Steve Jobs, but you're not Steve Jobs, then, sure, pre-record it.
But when the standards are lower, people appreciate the live demo. Videos are much less compelling, and they could've watched that at home in a few minutes, rather than fly out to you.
In grad school -- in a lab where, instead of "publish or perish", our mantra was "demo or die" -- we were giving demos most weekdays, usually to only a few people at a time. Always live. I think once I had to play a video of someone else's stuff, and it fell flat.
My best demo story is when I was showing a new thing (just a class project), to a few people crowded around a computer in my lab area at an open house event, typing into my software agent in a chat window (pre-LLMs), ad-libbing and speaking aloud what I was typing, "I'm interested in... thinks Microsoft and... thinks Oracle..." Then I turned my head around to look at the people behind me, to see that they were following, and I swear to gosh, Larry Ellison himself was standing in the doorway of my lab area, giving me this fiercely aggressive stare. Whether he was glaring because I said Oracle or Microsoft, I do not know.
My worst demo story was at the very start of my demo career. I had just arrived on campus, and I was borrowing a computer in an open area. A research scientist comes up, and sees on my screen this Web-scraping personalized newspaper thing I had made. So I'm showing it to them, and they're interested, then all of a sudden they say, sadly, "Mother Theresa died..." Which was what my "demo" was showing.
So maybe don't include an uncontrolled live data feed of depressing or distracting news in your demo.
Also, implementing early software agents in Java apparently summons Larry Ellison.
I agree that you should always have a pre-recorded version as backup, but live demos communicate a confidence in your product, and that can be worth something. Whenever I see a pre-recorded demo I wonder how many takes it took them, or how many pauses were taken out in editing, etc.
Meta took that risk and failed.
It helps if you have a master of ceremonies capable of running three rings simultaneously. Jobs could do it. Looking at what Apple is doing, they feel Cook is not. You lean into your strengths and avoid your weaknesses. Meta should realize that theZuck is not Jobs and is much closer to Cook
This article reminded me of a talk I gave a couple times that went very smoothly because I pre-recorded the typing parts. Something like this:
- record the screen as you type out all of your code as you would do live
- don't need the audio, just the video
- edit it down to remove typos and pauses
- speed it up to a nice fast typing pace
Now you can drop the video into Keynote or whatever, and Play/Pause the video as you speak.
A crucial part of this is that the video has effectively no pauses. Maybe just a few short ones between sections, to leave time to hit Pause before the next keystrokes appear.
Instead of a 30 minute video that I had to perfectly keep pace with, it was like 5 minutes of rapid typing that I could speak along with and pause when it was time to explain something.
At this point in time, it is absolutely trivial to pre-record whatever you'd like, and it will look glorious, and it doesn't need any basis in reality.
As a customer, when you show me a video, I default-assume you're lying.
Also, as a presenter: If your live demo fails and you can't turn that into a win, maybe you're not good at presenting. Your audience wants you to succeed, it takes a minimal amount of charm to make it work. (Heck, even Mark turned this into a win, and he's not exactly a charm bomb)
When I'm shown a pre-recorded demo, I suspect the product isn't very robust.
Thought at first this was maybe just being applied to external product demos, but this reads like dev demos, too:
> Since most of the time you're demoing something you previously worked on, this prevents you from working on current stuff right up to the demo, because you're worried you might break your local setup for the demo by installing different node modules, switching to different Git branches, etc.
This isn't the case for me generally... probably more of problem for web-apps things I guess.
I keep multiple workspaces locally for the multiple things I'm working on, but that hasn't been my pain point. Usually it's that I'm demoing a work in progress (not my choice generally) that has rough edges.
I enjoy the internal demos where questions can be asked about doing XYZ, and that can be shown if it's quick. Recordings tend to take away from that, and be less memorable/engaging I guess.
First time I recorded a demo was when it became too hard to return things to a known good state after every demo. In this case that meant multiple RAID arrays, 3 servers, an application with DB, and monitoring software.
All so that I could demonstrate that when I turned off a RAID controller a red light came on and the application just…continued.
So much work for something largely unspectacular meant that a recorded demo that could be run anytime from a laptop was far more appealing. My junior colleagues were now doing demos they otherwise wouldn’t, and the seniors turned off the sound and edited the video into snippets they could introduce and talk around
A company I worked for switched over to pre-recorded demos and everyone talked about how clever it was for the first few larger audiences. Then they made a mistake and replayed a clip during the same session and the audience chat blew up. You could see a dip in new users for days after the demo.
Have to mention this middle ground, I've been using it for all CLI based technical demos since I discovered it:
https://github.com/sloria/doitlive
Live is also a massive time waster/cost.
Apple can pull it off, but if you’re at a mid market tech company you just sunk like weeks people’s time into choreographing a live movie. Super stressful time sync.
IMO live demo is better in some cases, but if there are complicated or time-consuming steps where the prerequisite can not be met quickly, a recording would be better
Audience knows when something is live
> So here’s the story behind why yesterdays live #metaconnect demo failed - when the chef said “Hey Meta start Live AI” it activated everyone’s Meta AI in the room at once and effectively DDOS’d our servers
> That’s what we get for doing it live!
And if you believe that, I've got a bridge to sell you...
It makes me question the rest of this article when the author takes that at face value. A DDOS made your LLM say the same thing over and over? Yeah, I don't think so. Also, this new tech, how many device that support “Hey Meta start Live AI” even exist or were "listening" to the stream.
This is about a valid as Zuck saying that it was "Wifi" issues. How often do you have Wifi issue where you can talk to remote servers just fine?
Can't pre-record all demos.
But anywhere the demo gods can show up, for example in live coding during a demo, can be partially or fully recorded, but narrated in person.
A live demo is better for selling. A live demo says "this is real and it works".
I find this overly credulous of FB’s stated reason why it failed. Maybe it’s just bad?
If you aren’t doing a live demo, it’s not a demo, it’s just a sizzle reel.
I think the problem is, so many demos are cooked these days, that it’s so much more trustworthy to see it work live.
or not like the recent FB AR glasses. if your live demo doesn't work in rehearsal, there's a good chance it's not going to work during the actual presentation. there's a reason the phrase "demo hell" is well understood.
But the problem is that the product is buggy. The question then is "can you fix those bugs before products ship."
I'm not sure what the problem is here. Customers don't want buggy products. Live demos are more informative to users who can't tell what the actual experience will be like vs what the envisioned experience is.
I'm sorry, but as a customer I don't care what's in a developer's dreams, I care what I can buy.
At the same time, no one will be talking about the new Meta if not for the demo failures.
The lion does not concern himself with the possibility the live demo might fail.
It builds trust if its live
I always do half and half, a mostly live demo with some cooking show style start and results. It does depend on your audience, and there is charm in a live demo.
No one wants to see a terminal, that's for sure, so don't display code or start pecking away at a terminal unless you really think you'll need it.
Yes - I call them "the other cake" demos
I for one love seeing a terminal, as long as the presenter is a competent typist and has properly thought the demo through.