Latest / AV SuperFriends / AV SuperFriends: On Topic - That nice squishy in-between
Transcript
- 0:09This is AV Superfriends On Topic,
- 0:11a thirty-minute monthly roundtable focused
- 0:13on a single topic.
- 0:14This discussion will be opinionated,
- 0:16to the point,
- 0:16and saturated with our well-known blend of
- 0:18painfully one expertise and deeply
- 0:20ingrained maturity.
- 0:28For October,
- 0:29it's the best-sounding and most
- 0:31forward-thinking podcast in higher ed.
- 0:33This is the AV Super Friends on-topic
- 0:35episode number eighty-two.
- 0:37Broadcasting live from coast to ghost,
- 0:41from the witching hour to the happy hour,
- 0:43from haunted halls to bathroom stalls.
- 0:46rattling chains raising the dead and
- 0:47resurrecting equipment that should have
- 0:49been buried years ago i'm your fright
- 0:51frighteningly busy master of scares
- 0:53nightmares and technical affairs i'm chris
- 0:55decker you're tuned for another friday
- 0:57afternoon podcast adventure with the av
- 0:58super friends let me be the first to
- 1:00say happy halloween you magnificent
- 1:02bastards that's tomorrow
- 1:05Oh, you want to start again?
- 1:07I'll take it all back.
- 1:07We'll do it back.
- 1:08We're going to do it tomorrow.
- 1:09Happy Halloween Eve, Justin.
- 1:11How about that?
- 1:12That is the thing in some places I
- 1:13hear.
- 1:13It is.
- 1:15It's All Hallows Eve.
- 1:17I thought that was All Hallows Eve is
- 1:18tomorrow.
- 1:19Yes, that's what I'm saying.
- 1:21This is All Hallows Eve Eve.
- 1:23Eve, okay.
- 1:25Well,
- 1:25let's say hello to today's slightly
- 1:27expanded panel of AV Super Friends.
- 1:29Yes.
- 1:32Our first panelist is responsible for
- 1:33identifying unexplained noises,
- 1:35flickering displays,
- 1:36and classroom technologies that only
- 1:37malfunction when faculty are present.
- 1:40I'm sorry, when faculty are present.
- 1:43Misspoke on that one.
- 1:45He's a director of Haunted Learning
- 1:46Environments from Kansas City.
- 1:47Say hello to Mr. Jamie Reinhart.
- 1:48Although, to be fair,
- 1:50when the president's around, too,
- 1:51it always seems to pop up and have
- 1:53cool problems.
- 1:54High-level admin, all sorts of problems.
- 1:57Our next panelist maintains equipment well
- 1:58beyond manufacturer support,
- 2:00including devices that have technically
- 2:01been dead for years.
- 2:02He's a senior specialist for end-of-life
- 2:04systems and afterlife support from
- 2:06Corvallis, Oregon.
- 2:07Say hello, Mr. Mark Kolowczynski.
- 2:09What's old is new again.
- 2:10We will breathe some life back into this
- 2:11stuff.
- 2:13Why is my song so quiet?
- 2:15It's fine.
- 2:16Don't touch it.
- 2:17I'm touching everything.
- 2:17Look at this.
- 2:18Stop playing with it.
- 2:20Turn that knob a bit.
- 2:20This is getting weird.
- 2:24It's always been weird.
- 2:24Very weird.
- 2:25It's always been weird.
- 2:25Our next panelist coordinates
- 2:26investigations into trouble tickets that
- 2:28mysteriously reopen despite repeated
- 2:30assurances that the problem has been
- 2:32resolved.
- 2:33He's a manager of unresolved incidents and
- 2:35recurring apparitions from Greensboro,
- 2:36North Carolina, Salem, Mr. Larry Darling.
- 2:39Technically,
- 2:39if an incident keeps reopening,
- 2:41that becomes a problem.
- 2:43I thought that was a story.
- 2:45Nope.
- 2:46That's a problem, not a story.
- 2:47Larry, I picked that one just for you.
- 2:49Was that a chapter, though?
- 2:50Was it more close to a chapter?
- 2:52No, it's an epic problem.
- 2:55Don't put quarters in Larry.
- 2:58Our next panelist oversees the integration
- 3:00of legacy AV infrastructure with entities
- 3:02that refuse to cross over to supported
- 3:03platforms.
- 3:04He's a vice president for Paranormal
- 3:05Systems Investigations.
- 3:07I'm sorry, read that wrong, too.
- 3:08Paranormal Systems Integrations from
- 3:11Bowling Green, Kentucky.
- 3:12Say hello, Mr. Justin Rexing.
- 3:13That's actually pretty accurate.
- 3:15It's a day's day, right?
- 3:18I'm just going to go get new glasses.
- 3:19That's probably the most accurate one yet.
- 3:22And finally,
- 3:22joined by a special ghost panelist...
- 3:26He's a former systems engineer who
- 3:28continues attending project meetings from
- 3:29beyond the grave,
- 3:30primarily to explain why the original
- 3:32design was better and nobody should have
- 3:34touched it.
- 3:34He's a guest emeritus of posthumous
- 3:36technical consultation from Ottawa,
- 3:38Ontario.
- 3:38Say hello to Mr. George Burchall.
- 3:41Thank you.
- 3:41It's nice to be here.
- 3:42I see you're a very pedantic and
- 3:45opinionated group,
- 3:46so I think we're getting along really,
- 3:47really well.
- 3:50That's one way to put it.
- 3:51It's pedantic.
- 3:53So usually people just say a-holes,
- 3:55but OK, we'll take it.
- 3:56Jerks is the common parlance.
- 3:58But so on today's show,
- 3:59we're going to discuss MCP and
- 4:01interconnecting systems and all that good
- 4:03stuff as far as systems integration and
- 4:04whatnot with our with our friends here
- 4:06from Epiphan.
- 4:07But first, a little bit of housekeeping.
- 4:10And before I do that,
- 4:11just want to remind people we did publish
- 4:13our Halloween special this week.
- 4:15So if you haven't had a chance to
- 4:16listen to that whole thing,
- 4:17please go grab that.
- 4:17It's on the same feed.
- 4:18It'll be out there.
- 4:20I dare you all to listen all the
- 4:21way through the credits,
- 4:23at which point there is a song that
- 4:24will get stuck in your head and drive
- 4:26you nuts.
- 4:26So it is very special.
- 4:28Give that one a shot.
- 4:30And yes, I kept the van.
- 4:31All right.
- 4:32So if you're listening,
- 4:32this is recorded MP three.
- 4:34You're already too late.
- 4:35We've been broadcasting from beyond the
- 4:36grave every Friday at about three,
- 4:37fifteen, eight, twelve,
- 4:38fifteen Pacific or whenever Mark rises
- 4:40from his crypt.
- 4:41You can join us for that live on
- 4:42Twitter, LinkedIn and YouTube,
- 4:44where you can watch six grown adults argue
- 4:45about firmware updates,
- 4:46haunted control systems and whether or not
- 4:47the strange noise coming from the
- 4:48equipment rack is a failing fan or the
- 4:50tortured soul of a Navy technician who
- 4:52died waiting for a purchase order to be
- 4:53completed.
- 4:55If you can't join us live,
- 4:56please summon us on the socials.
- 4:58You can follow us on Twitter, LinkedIn,
- 4:59YouTube, Blue Sky, all that stuff.
- 5:01No candles, pentagrams,
- 5:03or blood sacrifices required,
- 5:05although we don't rule anything out if it
- 5:06helps our engagement numbers.
- 5:08Please don't forget to spread the curse.
- 5:09Share this podcast with your friends,
- 5:11colleagues, enemies,
- 5:11unsuspecting trick-or-treaters,
- 5:12and that one faculty member who insists
- 5:14the projector worked perfectly in in.
- 5:18Sneak our episodes into their playlist.
- 5:19You have QR codes in haunted houses.
- 5:20Whisper AV superfans three times in a
- 5:22darkened classroom.
- 5:23to see what happens to the touch panel.
- 5:25If the system suddenly responds,
- 5:27please just submit a ticket.
- 5:30Finally,
- 5:30if you want to keep this shambling corpse
- 5:33of a podcast alive,
- 5:35consider throwing a few dollars into our
- 5:37donation coffin.
- 5:37Your contributions help us to pay for
- 5:39hosting equipment,
- 5:40the occasional exorcism of a possessed
- 5:41matrix switcher,
- 5:42and whatever else we happen to be spending
- 5:43that on.
- 5:44You can find the donation link for that
- 5:45in the show notes or on the website
- 5:46at avsuperfriends.com.
- 5:49And all that to be said,
- 5:50happy Halloween tomorrow from the AV Super
- 5:52Friends.
- 5:53And thank you for your support because we
- 5:54couldn't do this without you.
- 5:56Justin, we have to check that account.
- 5:58See if anything showed up in there.
- 5:59Justin never logs into that.
- 6:01It's a little low, guys.
- 6:03Not going to lie.
- 6:03All right, that's fine.
- 6:05We need more donations.
- 6:08More candy in our jack-o'-lantern.
- 6:10You want some more candy in your
- 6:11trick-or-treat bag there?
- 6:12Yep.
- 6:14That's what I'm giving out this year at
- 6:15my house.
- 6:15You're just getting flash drives with
- 6:18recorded episodes.
- 6:20Is it out of your new van?
- 6:22Do you have a sign that says free
- 6:23candy?
- 6:23Yes.
- 6:26It's his old van.
- 6:26Thank you very much.
- 6:27The same one he's been using for years.
- 6:29With no windows?
- 6:30Or make sure there's duct tape on the
- 6:31windows.
- 6:32Said puppies for a while.
- 6:35All righty.
- 6:35Well,
- 6:36now that we've set that stage correctly,
- 6:38so see what you're in for, George.
- 6:39It's going to happen.
- 6:40All right.
- 6:43So, yeah, joined by a special guest,
- 6:46George Birchall from Epiphan here.
- 6:48George,
- 6:48let's give a little bit of background on
- 6:49Epiphan and then talk about you guys just
- 6:52announced something really cool last week.
- 6:54I saw a demo of this the week
- 6:55before, blew my mind,
- 6:57and we immediately said we need to have
- 6:58you guys on to talk about this because
- 6:59this is just...
- 7:01This is the way things are going,
- 7:03whether or not people know it yet or
- 7:04not.
- 7:04And I love seeing what you're doing with
- 7:07this.
- 7:08Oh, thanks, Chris.
- 7:08It is a really fun time for all
- 7:11of the people in your shoes and your
- 7:14peers' shoes.
- 7:15And to give a bit of background,
- 7:17we're going to get into the weeds on
- 7:18what we've announced.
- 7:19We're going to get way into the weeds.
- 7:21It is going to go deep.
- 7:22But to start off,
- 7:24so we're known as a company building
- 7:27streaming and recording hardware and
- 7:29classrooms.
- 7:30So a lot of people will have us
- 7:31in, you know,
- 7:32ten classrooms or a hundred classrooms.
- 7:34And the one thing I see when I
- 7:36speak to people in your shoes is that
- 7:38everybody seems to be overworked to some
- 7:40capacity where they have
- 7:41way more rooms and spaces and disparate
- 7:44bits of technology that they have to
- 7:46figure out how to manage,
- 7:48which means you get all of these people
- 7:50who have been in roles for twenty years
- 7:52and have to know like how much knowledge
- 7:54you have to know to be able to
- 7:55do your job about all the different
- 7:56systems and technologies that have changed
- 7:59constantly.
- 8:00So the bar to be good at what
- 8:02you guys do is so hard, so high,
- 8:04right?
- 8:04Like you have to just know so much
- 8:07and
- 8:08This means it's hard to break into this
- 8:12field.
- 8:12It's hard for new people to come along
- 8:13and learn how to do it because you
- 8:14have to learn all these new systems and
- 8:17new technologies.
- 8:18So what we've done is we put out
- 8:20something called an MCP.
- 8:23And that is basically the glue that lets
- 8:24you use AI tools such as Clod or
- 8:28ChatGPT to be able to talk to all
- 8:31of this equipment in that room.
- 8:34So our MCP lets you speak to Epiphan
- 8:37equipment.
- 8:38But we're envisioning this world where you
- 8:40have a translation tool that can help you
- 8:43speak to all of the equipment in that
- 8:45room.
- 8:46So we're thinking beyond just our own
- 8:48streaming and recording side of things.
- 8:50But imagine if you can go into ChatGPT
- 8:53and say, how are my rooms doing?
- 8:55and understand how your displays are
- 8:58working out,
- 8:58how the lighting systems are working,
- 9:00how the streaming and recording systems
- 9:01are working, the control systems,
- 9:03all of that together,
- 9:04and not have to navigate twenty different
- 9:06UIs to understand what's going in that
- 9:09room.
- 9:10So I know we were chatting, you know,
- 9:12getting ready for this call about some of
- 9:13the applications,
- 9:14and that's probably where we'll spend the
- 9:15bulk of our call today.
- 9:16But the nuts and bolts of what we
- 9:17announced is an MCP,
- 9:19which allows you to speak to
- 9:21our technology in the room using ChatGPT
- 9:24or Cloud or LLM.
- 9:26Yeah, I think for
- 9:30Listeners who may have no idea what MCP
- 9:33is,
- 9:33I think we do want to at least
- 9:35give some little nugget as to what's going
- 9:37on here.
- 9:38Model context protocol.
- 9:41Put it in really simple AV terms.
- 9:42The way I explain it,
- 9:43it's kind of the USB-C of AI.
- 9:45It's the connector.
- 9:46It allows things to bolt in and expose
- 9:48the capabilities of a service or a
- 9:50product.
- 9:50You say it's the USB-C of AI?
- 9:53That's how it's often actually described.
- 9:54I would say...
- 9:55It makes things discoverable.
- 9:57It does all these things.
- 9:58It...
- 9:59Yeah, but USB-C sucks.
- 10:00It does.
- 10:02When you say all these different things
- 10:04that happen,
- 10:05you can do a lot of different things.
- 10:06It's not just one thing.
- 10:07It's all of the underlying capabilities
- 10:10can be exposed by way of this small
- 10:12server that allows AI services to bolt in
- 10:16to then do other services after the fact.
- 10:19It's kind of like the new API.
- 10:23That's how I think of it, Mark.
- 10:26Yeah, Mark.
- 10:26Let's get it right.
- 10:28I will stay right.
- 10:29You guys can stay wrong.
- 10:32I'm still lost.
- 10:33With an API,
- 10:34you'd have to go and do a whole
- 10:35lot of work to connect it, right?
- 10:37You don't just say, hey, APIs,
- 10:38talk to each other.
- 10:39You've got to map all of that.
- 10:41This does all of that for you.
- 10:43It's like an API that has intelligence
- 10:45that just says,
- 10:45you can figure out how to map all
- 10:46of these commands between these different
- 10:48services.
- 10:49Right.
- 10:50Well,
- 10:50and as part of leveraging MCP through
- 10:53different models,
- 10:54even if those commands aren't necessarily
- 10:56there,
- 10:57it can get a lot of that information
- 10:58back and forth.
- 10:58So sometimes you're not even relying upon
- 11:01what the developers of the other hardware
- 11:03or software have provided.
- 11:04Indeed.
- 11:06Yeah.
- 11:07And so you all developed this MCP.
- 11:08Yes.
- 11:08On your hardware.
- 11:11And we got...
- 11:13so far down the rabbit hole on our
- 11:14planning call.
- 11:15But what spurred you to kind of,
- 11:17you kind of gave a little brief on
- 11:18it.
- 11:18Was there,
- 11:19were there specific elements of the
- 11:21hardware in your offerings that you're
- 11:23like,
- 11:24this is why we want to do this.
- 11:25Like, what are we,
- 11:27Why did you embark down this road?
- 11:28What was it about this?
- 11:29You know,
- 11:29it actually came from working outside of
- 11:33the audio video world.
- 11:34So on a day-to-day basis, you know,
- 11:36we make AV equipment,
- 11:38but we use all kinds of different
- 11:39services.
- 11:40And I'll give you one example.
- 11:41We use a product called HubSpot.
- 11:42HubSpot is how we learn about all of
- 11:44our customer accounts and all of our sales
- 11:46and marketing aspects.
- 11:47And if anybody here has used a product
- 11:49like HubSpot or Salesforce or Atlassian
- 11:52tools or some of these big systems where
- 11:54they're very,
- 11:55very powerful to go learn how to use
- 11:57HubSpot,
- 11:58it's not going to be like in a
- 11:59minute.
- 12:00You're going to it's going to take you
- 12:01a while.
- 12:02Guess what?
- 12:03I don't use HubSpot anymore.
- 12:04I just use Claude.
- 12:05And I say, hey, Claude,
- 12:07tell me what you know about Justin and
- 12:09his organization.
- 12:10Don't do that.
- 12:11All over you.
- 12:15And then I don't have to go and
- 12:16learn how to use HubSpot.
- 12:17So it's like, wow,
- 12:19I can be so much more powerful just
- 12:20by using Claude instead of having to learn
- 12:22how to use this complicated piece of
- 12:23software.
- 12:24And our equipment is not nearly as
- 12:26complicated as a software like Salesforce
- 12:29or HubSpot,
- 12:30but it still requires a certain amount of
- 12:31knowledge to go use it.
- 12:32So I'm like, huh,
- 12:33could I just use Claude and say, hey,
- 12:35go set up my Pearl,
- 12:37my Epiphan device in a certain way.
- 12:39And then let me know if something goes
- 12:41funny with it.
- 12:42So that's the most basic version of it
- 12:44that we came to mind.
- 12:45And it was it was a parallel to
- 12:46what we were doing in our work life
- 12:48with all of the software that we're using.
- 12:51It just hasn't really evolved to the point
- 12:53where it's part of every hardware stack.
- 12:54So I think that's why this is sort
- 12:56of new to the AVU community.
- 12:57But in the software land, we're all using,
- 12:59you know, MCPs to control, maybe not MCPs,
- 13:02but we're all using at least LLMs to
- 13:05control our software.
- 13:06Yeah,
- 13:07I think that the agentic workflows that we
- 13:09were starting to do
- 13:12bolting things to models and different
- 13:13capabilities and different agents and
- 13:14different harnesses.
- 13:15Like we need some common connector type.
- 13:17This MCP gives us the ability for whatever
- 13:19that service you want to provide can look
- 13:22in and take action and see what is
- 13:24available to it.
- 13:24I think that's kind of why I believe
- 13:27everybody here saw the value in this
- 13:30posture and this offer because I think
- 13:31we're all like, yes,
- 13:32we see these agent workflows
- 13:33coming from every different angle.
- 13:35And it was good to see a company
- 13:36like Epifancy.
- 13:37You know what?
- 13:37We see it.
- 13:38Here it is.
- 13:39Love seeing that posture and kind of you
- 13:41rolling that out there.
- 13:43Well,
- 13:43and we see it in the community as
- 13:45well in that I think if you speak
- 13:47to most of your peers,
- 13:48they have some kind of like daily check
- 13:50on their rooms.
- 13:51And I was like,
- 13:51how do you get ready for a classroom?
- 13:52And some people will check like literally
- 13:54they're listening at the first five
- 13:56minutes of every class just to make sure
- 13:58the sound is good and that the video
- 14:00is good.
- 14:00Or it's on a daily level.
- 14:02They check every morning,
- 14:03make sure all the classes are right,
- 14:04make sure the chairs are set up.
- 14:06Boy, AI does that really, really well.
- 14:09It may not do it as well as
- 14:10you being physically in those spaces,
- 14:12but to have AI and say,
- 14:14go check these five rooms and tell me
- 14:17these five things.
- 14:17Is the audio working?
- 14:18Is the lighting on?
- 14:20You can list those things.
- 14:21Those are relatively easy to instruct the
- 14:24technology that we have in those rooms to
- 14:26report back and say, yeah,
- 14:28I can see the cameras pointed at where
- 14:29it's supposed to be.
- 14:30I can see,
- 14:31I can hear sound coming from somewhere.
- 14:33So just at a basic health check level,
- 14:36we thought, wow,
- 14:37when we talk to people about that aspect
- 14:39of it,
- 14:39that seems to be universally people go,
- 14:41yeah, I kind of get into that.
- 14:43like it's a anything that's repeatable
- 14:45that has like a very clear definition of
- 14:47what success can be yeah yeah right so
- 14:49george to that point real quick uh you
- 14:51mentioned the kind of like the daily check
- 14:53of things or the weekly check that
- 14:55everyone would love to do more often jamie
- 14:57for years you guys would have people there
- 14:59at five in the morning going through near
- 15:01every room to ensure anything odd was
- 15:03caught you know before um before classes
- 15:06started and all that but
- 15:09George,
- 15:10I'm curious as that as you heard that
- 15:12comment, you know,
- 15:13how did what was that that
- 15:16I'm curious what that process was like
- 15:17where you tumbled to that, like, hey,
- 15:19well, we've got this device in there.
- 15:21It can see and hear most of the
- 15:22stuff going on in there.
- 15:24Why don't we use this as kind of
- 15:26the person who's going around checking
- 15:27stuff?
- 15:28I mean, what was that process like?
- 15:30To me, it's something that's like,
- 15:31this is so simple.
- 15:32Why didn't I think about this?
- 15:33And then when Ron showed me kind of
- 15:36the demo of this a few weeks ago,
- 15:37I was like,
- 15:38this is amazing for exactly what you said.
- 15:40This is a routine,
- 15:41I don't wanna say low level,
- 15:42but rudimentary thing that we do on a
- 15:44regular basis.
- 15:46why why can't i automate that and get
- 15:47a report when i show up on campus
- 15:49at eight in the morning hey i looked
- 15:51at these sixty spaces there's two you need
- 15:52to go check but otherwise the rest are
- 15:54good to go well it's it's doing what
- 15:57you guys are doing now it's not inventing
- 15:59new processes it's just finding another
- 16:01way of doing those processes so we had
- 16:04all of the data from people like you
- 16:06saying boy i really wish i knew if
- 16:07the audio was good in these rooms before
- 16:09my classroom started
- 16:10So people were building homemade solutions
- 16:12to do it.
- 16:13I knew people who would do a test
- 16:14recording and then that recording you get
- 16:16pushed to some server and then they would
- 16:18download that and try and analyze it with
- 16:19some kind of tooling.
- 16:21So there's homemade solutions galore and
- 16:23this just basically makes all of those
- 16:25homemade solutions.
- 16:27Well,
- 16:27it's a better version of all the work
- 16:29that you're doing now, so it's.
- 16:31It's not necessarily net new in terms of
- 16:33the behaviors.
- 16:33It's just a different way to do it
- 16:35that's more efficient.
- 16:35Well,
- 16:36I love the idea that it could automate
- 16:37this stuff because we have those spaces on
- 16:39campus too where somebody wants us to go
- 16:40check, hey, we're doing exams this week.
- 16:41Can you go check one every day in
- 16:42the morning sort of thing?
- 16:44I love the idea that I can take
- 16:45that.
- 16:45I can have that report created for me.
- 16:48I can have it automatically emailed to
- 16:49somebody.
- 16:50So when they come in,
- 16:51they get the report, hey,
- 16:51you're five rooms.
- 16:52They were checked an hour ago.
- 16:54They're all in good shape.
- 16:54Good to go.
- 16:56Yeah,
- 16:56and you can check things using this kind
- 16:58of approach that you couldn't necessarily
- 17:01check using other methods.
- 17:03You can't necessarily ask an API,
- 17:06tell me if the lights are on in
- 17:08this room.
- 17:10It doesn't have the genetic capabilities
- 17:12of doing those things.
- 17:12The intelligence on top of it.
- 17:13It's able to look at a frame from
- 17:15a video,
- 17:15capture a still image from this video,
- 17:17analyze that frame,
- 17:18and tell me what's happening in that room.
- 17:20Are there students in the room?
- 17:22Is there a faculty member in that room?
- 17:24That doesn't show up in SignalPath.
- 17:28That's where it gets pretty exciting,
- 17:29actually.
- 17:30Go ahead, Justin.
- 17:32Well,
- 17:32I was going to say or ask a
- 17:34question here, and I don't know,
- 17:35this might be getting too into the weeds,
- 17:37but what is, like a company like yours,
- 17:40what is the process of deciding what
- 17:42capabilities to expose to that MCP versus
- 17:45what not to?
- 17:46How do you all receive feedback on
- 17:50what if your mcp did this and it's
- 17:53not doing that now how does that process
- 17:55look like to keep developing that mcp and
- 17:58the capabilities or just overall thought
- 18:01on how that even started to get developed
- 18:04and what's even exposed sure i mean we
- 18:06had a pretty good uh starting point and
- 18:10that we have cloud
- 18:12management,
- 18:12which allows you to see into all of
- 18:14these rooms so we can tell what people
- 18:16are typically doing.
- 18:17So that's a good starting point.
- 18:18We can see people care about these twenty
- 18:20seven things.
- 18:21So we started there.
- 18:22Let's let's replicate what they can view
- 18:24in a dashboard.
- 18:25And we get feedback.
- 18:26That's the easy start.
- 18:28Yeah, that's easy.
- 18:28They tell us what they want to see.
- 18:30And we're always iterating on that side of
- 18:31things.
- 18:32But then there's the mystery side of
- 18:33things where when you give it the ability
- 18:36to see into those rooms, well,
- 18:38that's when you get into the visual
- 18:40intelligence side of things.
- 18:41And it can say there's thirteen,
- 18:44thirteen people in that room.
- 18:46And so your room utilization suddenly
- 18:48exposes itself.
- 18:49And that's more because we're giving
- 18:51access to cameras and audio in these
- 18:53spaces.
- 18:54Yeah,
- 18:55I'm just thinking of like odd capabilities
- 18:57that, you know,
- 18:58if you had one of these devices in
- 19:01every classroom in a building and you've
- 19:04standardized on that,
- 19:05and let's say you have good integration
- 19:07with other departments on campus to where
- 19:13you're leveraging that for security
- 19:15applications, because I know we don't,
- 19:16you know,
- 19:17we put a lot of security cameras
- 19:18everywhere,
- 19:19but let's say there's a lockdown.
- 19:22How easy would that be for the police
- 19:25department or somebody to run a quick
- 19:27report from an agentic AI to just see
- 19:31what the classrooms are doing right now
- 19:33with all the students?
- 19:34There's, you know, eighty percent of them,
- 19:36ninety percent of them,
- 19:37all the students are kind of sitting there
- 19:38wondering what's going on.
- 19:39And these two over here,
- 19:40there's no students there.
- 19:41So that may be where we need,
- 19:43you know,
- 19:44have you thought or talked about any odd
- 19:47cases like that of using your MCP for
- 19:50edge cases that aren't even related to AV?
- 19:53I've had those conversations with people
- 19:55in your shoes.
- 19:56You could tell me individually what your
- 19:59school is like,
- 19:59but it seems to be the security systems
- 20:01tend to be a disparate set of technology
- 20:04and cameras and spaces.
- 20:06If you're a CIO and you can make
- 20:08that level of decision, yeah,
- 20:09let's merge these two under one umbrella.
- 20:11than that reality that you just painted,
- 20:13where you get not only more data exposed
- 20:16to those security systems,
- 20:17you get efficiencies.
- 20:19Insight to the classrooms,
- 20:20insight to what's going on at the moment.
- 20:22Because I mean,
- 20:22that's kind of where that data-like
- 20:24problem exists is where agentic AI comes
- 20:26in.
- 20:26So if you have an agentic AI living
- 20:28with your security system and then
- 20:31yeah but that same thing agentic ai
- 20:33layered on top of your mcp and you're
- 20:35looking at all that together that's
- 20:36correct that's an incredible data set from
- 20:38just a security application that's just
- 20:40one example of yeah it requires no no
- 20:44additional lift it's just policy right
- 20:46it's just all there you're just going to
- 20:47tie it in and it takes ten minutes
- 20:49to tie these things together so yeah i
- 20:51think there's a couple different pieces
- 20:52that i want to explore a little bit
- 20:54is
- 20:55One, culture of Epiphan, right?
- 20:58Do you have like pervasive AI usage within
- 21:00the organization?
- 21:02Yeah, yeah, we do.
- 21:04So you guys are just iterating,
- 21:07teaming with...
- 21:08We're typical of a technology company in
- 21:10that we tend to be early adopters of
- 21:12a lot of technology.
- 21:14And we quickly learned that AI was going
- 21:16to be a big part of our business.
- 21:17And we've been running on full steam.
- 21:19And then the other kind of pieces that
- 21:22I think...
- 21:23I really like this kind of composable,
- 21:26flexible architecture that...
- 21:28It's not that we're hearing like,
- 21:30Epiphan's not building out this whole
- 21:31shingler law of AI downstream automation,
- 21:35single pane of glass.
- 21:35That's not what it is.
- 21:36They're posturing their solutions to be
- 21:39able to work with whatever you want to
- 21:40architect this within two, right?
- 21:42And so,
- 21:44Make sure listeners don't think that when
- 21:45we say MCP,
- 21:46that doesn't equate to you automatically
- 21:47have all of this AI service automatically
- 21:50happening.
- 21:50It's connective tissue.
- 21:51It's between those things.
- 21:52You still do need to build out that
- 21:55whole agentic kind of UI and what you
- 21:57want that to be.
- 21:58But I think the shift we're going to
- 22:00see in the industry is how do these
- 22:02tools now work into maybe a...
- 22:06a built ecosystem that you as an
- 22:07organization have put together you're no
- 22:09longer having to buy a full ecosystem from
- 22:12manufacturing anymore right so now we're
- 22:13having kind of compatibility through
- 22:16standards that we've always been asking
- 22:18for to now unlock these new types of
- 22:20agentic workflows that people are
- 22:22screaming that they want but we have to
- 22:24start somewhere and i think you took the
- 22:26right step is making your hardware able to
- 22:28work with whatever is going to come
- 22:30downstream
- 22:31Yeah,
- 22:31that's a big part of our DNA is
- 22:33to be agnostic and let you move video
- 22:35here, anywhere you want, the way you want.
- 22:39So that was a fairly easy decision for
- 22:42us.
- 22:42But I got to ask you guys,
- 22:43are you seeing people move away from these
- 22:46custom built control systems that allow
- 22:48you to manage the technology in your
- 22:51spaces to your own built ones?
- 22:53I don't mean custom built,
- 22:55like a vendor based system to your own
- 22:57dashboards.
- 22:58i would say the answer to that is
- 23:00we are looking at doing that ourselves and
- 23:04to answer your question have we seen that
- 23:06done yet not really very sparingly but i
- 23:08think that's where things are going it's
- 23:10yeah there's edge cases out there people
- 23:12who have done that but but that
- 23:14conversation comes up more and more and
- 23:15more that like well could i just get
- 23:17some generic hardware and do this myself
- 23:19you can so i'm not locked yes i
- 23:20know i'm speaking rhetorically justin yeah
- 23:23um
- 23:24So I'm not, to your point,
- 23:24not locked into this thing where I have
- 23:26to buy from vendor ABC and only their
- 23:28stuff works with whatever, you know,
- 23:30drivers or plugins or modules.
- 23:32Or even just having to manage multiple
- 23:34different,
- 23:35like everybody's got their own cloud and
- 23:37device management tooling.
- 23:38So then you have to log into that
- 23:40tool.
- 23:40Right.
- 23:41By the way, if you're curious,
- 23:43that number for me is,
- 23:44as best I can tell so far, nineteen.
- 23:47Nineteen to one.
- 23:50Our AVP asked that question.
- 23:52He was like, well,
- 23:53how many of these systems do we have
- 23:55out there and how many do we actually
- 23:56need?
- 23:57Just keeping my running list as it comes
- 23:58up, I'm at nineteen.
- 24:01Can you imagine it, Chris?
- 24:02Being able to take those nineteen and fold
- 24:04them into one of your own
- 24:07I've already started down that process.
- 24:09We're working on it.
- 24:12We're working with that guy and that guy.
- 24:15Call CRX Systems for more information.
- 24:19There was a piece you made, Chris,
- 24:22between George,
- 24:23that I think there's a lot of folks
- 24:26that will see this capability as...
- 24:29Maybe not the value add.
- 24:30This might be the value that they really,
- 24:33like, yes,
- 24:35I want this thing for this identity in
- 24:37my stack just to do this automation and
- 24:39remediation and eyes in and kind of that
- 24:41whole piece.
- 24:42And that I think there's an amazing
- 24:45opportunity for I think all manufacturers
- 24:47now.
- 24:47Like you put out a thing and you
- 24:49start adding these agentic capabilities
- 24:50and all of a sudden the thing changes
- 24:52its shape.
- 24:53right before your eyes of like what it
- 24:54can be.
- 24:55And I think this is the beginning of
- 24:56what we're going to see much more of
- 24:58this over the coming months and years
- 24:59that,
- 25:01To me,
- 25:01I love the capability this unlocks as far
- 25:04as eyes, ears,
- 25:05and bridging for AI with any learning
- 25:07space.
- 25:08Well,
- 25:08what I've heard from a couple of folks,
- 25:10Mark,
- 25:10on that is not only does it give
- 25:12them new value around the AV thing,
- 25:14but it gives you more data to speak
- 25:18to your CIOs about what's happening in
- 25:20these spaces.
- 25:21Because we've all been in these beautiful
- 25:23campuses with all of these incredible
- 25:25classrooms that are empty.
- 25:27underused, poorly used.
- 25:29And you guys see that.
- 25:31You see all the money that goes into
- 25:32spaces.
- 25:32And sometimes they get used brilliantly,
- 25:34and sometimes they just don't.
- 25:36Well,
- 25:36now you're going to have a report that
- 25:37says, well,
- 25:39this new wing that you put together,
- 25:41we're averaging nine students per class.
- 25:44So you get room utilization stats.
- 25:46And that's all just from a simple,
- 25:48relatively simple, I should say,
- 25:50agentic workflow, where you say,
- 25:51tell me who's using these spaces.
- 25:54So that, to me,
- 25:55gets me excited thinking about all the AV
- 25:56and IT people having a bigger voice in
- 25:58the room because you have all those new
- 26:00data.
- 26:01That is one question that comes up in
- 26:03my head a lot on campus is you
- 26:05look through the scheduling system.
- 26:07It's like, oh,
- 26:08there's supposed to be thirty kids in this
- 26:09class and you go by and there's like
- 26:11five going on here.
- 26:13But you don't know that data unless you're
- 26:15actually seeing into the room.
- 26:16But even beyond that, like in the next,
- 26:20not this replacement cycle,
- 26:21but the next replacement cycle,
- 26:23I've got to redo five active learning
- 26:25spaces across campus.
- 26:28They're all up for,
- 26:30and when you look at the cost of
- 26:32a room with too many monitors,
- 26:33as I like to say,
- 26:34versus a standard room,
- 26:36like if I had a way of actually
- 26:38looking into the space and using AI to
- 26:42kind of
- 26:44describe and understand what teaching and
- 26:46learning practices are happening in the
- 26:48room,
- 26:49I would have a much better case of
- 26:51saying, yeah,
- 26:51I can invest the money with these colleges
- 26:54because they're putting it into practice,
- 26:56but these other spaces, it's lecture.
- 26:58That's what they're doing.
- 27:00Larry, for that capability,
- 27:01you ought to call it flock.
- 27:04It might be,
- 27:05but at least that would be useful in
- 27:08that scenario because we're looking at a
- 27:10substantial investment difference.
- 27:14I think it's great that we take it
- 27:17from AV being a thing that was the
- 27:18output,
- 27:19like here's your data output on a screen,
- 27:21interact with it,
- 27:22to now leveraging that same stack for the
- 27:25data aggregation piece of this.
- 27:28So now we are part of that data
- 27:30capture and storytelling,
- 27:32and everything we're doing in that IT
- 27:33stack should be a bit
- 27:35drawing meaning how are we bringing
- 27:37meaning and context to what we're spending
- 27:39the business process and I think this tool
- 27:40is perfect for that whole thing which to
- 27:43me like that is so far like not
- 27:46on like the recording piece of the
- 27:47spectrum but now they're sitting here in
- 27:49the same conversation right they were
- 27:51never in a conversation now all of a
- 27:52sudden they're bolted onto one another
- 27:56Well, Mark, to that point,
- 27:57I think it's interesting that AV is,
- 28:00you know,
- 28:01we talk about how we present or we
- 28:03produce video and audio for experiences
- 28:05and teaching spaces and however you want
- 28:07to define that.
- 28:08but we can tie to aux sensors and
- 28:09we can tie to cameras and we can
- 28:10tie to scheduling.
- 28:12And so leveraging something like this,
- 28:13leveraging MCP with, you know,
- 28:15so I've got a hardware box in there
- 28:16that can do the translation.
- 28:17So I can,
- 28:18I can buy one of these Epifans,
- 28:19I can put it in a room,
- 28:21I use it for recording and streaming,
- 28:22but because it connects to all this other
- 28:24stuff,
- 28:24it gives me that context of what else
- 28:27is going on there.
- 28:27And kind of is that connective tissue in
- 28:29a physical sense,
- 28:31leveraging then MCP on the software side
- 28:33to kind of drag all that in.
- 28:34And now
- 28:35Suddenly,
- 28:35AV no longer like those guys downstairs in
- 28:38the basement of the library,
- 28:40but this is a critical part of our
- 28:43infrastructure and critical part of our
- 28:46efficacy of how well are we deploying
- 28:49teaching and learning spaces on campus.
- 28:53To Larry's point of cost,
- 28:54but even beyond that,
- 28:55are they even using these spaces?
- 28:57Are they in the right spot?
- 28:58Does this even make sense?
- 28:59And we can start to gather some data
- 29:02on that and some numbers on that that
- 29:04previously was maybe someone digs in and
- 29:07they work up a real complex report in
- 29:09the ticketing system, whatever,
- 29:10and they dump that out and that gets
- 29:11sent out once a year or twice a
- 29:13year to leadership.
- 29:15But
- 29:16This is functionally real time.
- 29:17I know it's not real, real time,
- 29:18but darn near real time snapshots of
- 29:21what's going on on a regular basis in
- 29:22these spaces from which you can build a
- 29:24lot of that kind of pull those trends
- 29:27and build that intelligence on top of it.
- 29:29It's a very cool concept.
- 29:30I love it.
- 29:33Just having our spaces become more aware
- 29:37in a way that does not take gobs
- 29:40and gobs of bespoke programming to do.
- 29:41I mean, because people back to learn,
- 29:43I could do this for years.
- 29:44I was doing reports.
- 29:44I built out.
- 29:45There's schools out there who have done
- 29:47this.
- 29:47They have built these systems,
- 29:48but they have very smart people who spend
- 29:50a huge amount of time to do it.
- 29:52But being able to now take and say,
- 29:54okay,
- 29:54we're going to partner with AI and
- 29:56We're going to bring in this layer that's
- 29:58that nice squishy in between and,
- 30:01and start to drive meaning so that our
- 30:03spaces can become aware of what's going on
- 30:05in there.
- 30:06So we can more accurately tell that story.
- 30:08Like,
- 30:08are we meeting our metrics of success?
- 30:12Yes or no.
- 30:12Do you think this wonder, go ahead,
- 30:14George.
- 30:15Well,
- 30:15to your point about how you have all
- 30:17these new powers available to you,
- 30:20do you think your peers are threatened by,
- 30:24you know, they spent their life,
- 30:25their value was a Crestron programmer,
- 30:27let's say.
- 30:29Now,
- 30:31a lot of people can become average
- 30:33Crestron programmers using AI.
- 30:34Yeah.
- 30:38Am I right?
- 30:38Oh, yeah.
- 30:39So, equally to those other ones.
- 30:40Best average you can be.
- 30:42Above average.
- 30:44How is that talked about in your world?
- 30:47Are people threatened by this or do they
- 30:49mostly see the upside?
- 30:51So, what I've seen directly around me,
- 30:53the people that I work around,
- 30:57they have slowly adopted AI.
- 31:00They thought at first it was, you know,
- 31:03going to be a problem oh that's going
- 31:04to take my job but not really the
- 31:07case what I'm seeing is it expands
- 31:09capabilities and so you can argue well if
- 31:14you if you did need if you have
- 31:16one programmer and you actually need two
- 31:18or three you're actually taking people's
- 31:19jobs because you're supposed to hire those
- 31:21people well if you work at a university
- 31:23creating a new position that's not reality
- 31:27and so we have to use new tools
- 31:29to
- 31:29to magically create these positions,
- 31:32which is just expanding capabilities of
- 31:34what our programmers,
- 31:35configurators can do.
- 31:37And I think once people understand that
- 31:39and adopt that philosophy,
- 31:43That's when everything just starts hitting
- 31:44the road and people start really
- 31:46leveraging AI.
- 31:47It's that flywheel, right?
- 31:48Once you get that flywheel moving and
- 31:50people start to see the value,
- 31:51and I think it's easy to get some
- 31:53efficiency use,
- 31:54but you've got to get past the efficiency
- 31:56into the opportunity use.
- 31:57I think that's where everybody wants to
- 31:58be.
- 31:59It's easy to just do the same things
- 32:00you've done before, but faster.
- 32:02But are you doing wholly new things with
- 32:05the capability?
- 32:05And I think that's where we're trying to
- 32:07push the envelope.
- 32:07I will make one interesting point.
- 32:09We've actually seen some resistance on the
- 32:11student side of AI.
- 32:14And this is a risky kind of paradox
- 32:16to get around here is that if we
- 32:18had students that are working in a
- 32:20capacity where their job is to do
- 32:21readiness checks and the like,
- 32:23there may be pushbacks if this thing were
- 32:24to drop in there and say, oh,
- 32:26that's going to feel like to them that
- 32:28this is definitely taking some of their
- 32:29job, right?
- 32:30And so if they say that's part of
- 32:32their identity as a student,
- 32:33so you have to,
- 32:33you can't remove an identity,
- 32:35you have to build on it, right?
- 32:36So it's very much that change management
- 32:38piece.
- 32:40as someone who employs a whole bunch of
- 32:41students to do exactly that,
- 32:43I don't see it that way.
- 32:44I would,
- 32:44cause they're not getting to every space
- 32:45every day.
- 32:46Now what Jamie was doing at UMKC for
- 32:47years is very unique because they,
- 32:49they saw the value of that.
- 32:50They put the dollars into having people
- 32:52there,
- 32:52even some FTEs there to back up the
- 32:54students.
- 32:54But,
- 32:55Most of us,
- 32:56if somebody's regularly checking a space,
- 32:58maybe it's once a week,
- 32:59maybe it's once every other week or one
- 33:01month.
- 33:02If I'm able to, as somebody who like,
- 33:04let's say I've got five students and this
- 33:06is kind of their job,
- 33:07but at the same time,
- 33:08they can leverage this as another tool.
- 33:11So yes, my zone is these fifty spaces.
- 33:15I can only get to ten a week,
- 33:16ten a day, whatever that number is.
- 33:18But now through this,
- 33:18I can get the other fifty or other
- 33:20forty get this way.
- 33:21And then I check the other ones.
- 33:22And it's just kind of it's one more
- 33:25tool in their tool belt to be even
- 33:27more comprehensive in their in their
- 33:29process.
- 33:30So you have to have intentional
- 33:31conversation around that, right?
- 33:32Because to some people,
- 33:33like when you just know,
- 33:34I just freak out and run around screaming.
- 33:37And there's some anxiety around it, right?
- 33:40They're like, oh, are they, like,
- 33:41if you don't walk them through that
- 33:42process,
- 33:43some of the students are very skeptical of
- 33:45what's going on because they are already
- 33:47being dropped into a world in which
- 33:48they're being told every day in the media
- 33:50that AI is coming to the job.
- 33:51So they're very,
- 33:52very heightened around that.
- 33:53And so when we say,
- 33:54here's a tool that's AI,
- 33:56some of them are going to be a
- 33:57little bit nervous around that.
- 33:58So you have to walk that line and
- 34:00give them the new power with that tool,
- 34:03not just dump it on them.
- 34:04Even if...
- 34:05Even if they're not against AI,
- 34:09they're going to have to learn how to
- 34:10use AI correctly.
- 34:11A lot of the problems that I see
- 34:13is people like, oh,
- 34:14I'll try to use it,
- 34:15and they use it without understanding how
- 34:17to use it properly.
- 34:19It's like, oh,
- 34:19it's just giving me the wrong answer,
- 34:20so I gave up.
- 34:23That's on you.
- 34:23That's not how that works.
- 34:25And this goes back to George's question a
- 34:27few minutes ago.
- 34:28Those of our colleagues out there who are
- 34:29traditional programmers,
- 34:31and this is kind of what they grew
- 34:32up with,
- 34:33is there pushback?
- 34:34There absolutely is pushback from some of
- 34:35them.
- 34:35No, I can do this better.
- 34:37I need to be much more hands-on on
- 34:39this.
- 34:40I get it.
- 34:40There are those,
- 34:41those unique cases out there where yes,
- 34:43you need to have someone in there really
- 34:45living in that code.
- 34:47But I think we're also seeing a lot
- 34:48more who are willing to embrace it and
- 34:50simply leverage it as a tool to continue
- 34:52doing their job more effectively,
- 34:55not simply replacing them wholesale,
- 34:57but leveraging it as something to,
- 34:59you know, uh, uh,
- 35:01they can do more or do faster or
- 35:03whatever especially there's cookie cutter
- 35:04spaces or something like that there's a
- 35:06lot of mistakes yeah or maybe just more
- 35:10mistakes that get rectified faster sort of
- 35:12thing but there is a lot of kind
- 35:13of interest in that for in certain uh
- 35:16certain corners of the of the industry and
- 35:18i think we'll see more of that as
- 35:20it kind of continues to grow and be
- 35:22more pervasive as a
- 35:23if nothing else, a topic of discussion.
- 35:25But as some of these tools like the
- 35:27ones you guys have developed for this sort
- 35:29of, you know,
- 35:30call it an environmental system check or
- 35:31whatever that may be,
- 35:32but
- 35:34Why can't I leverage that as part of
- 35:35the bespoke program that I'm going to run
- 35:38or I'm going to build to run the
- 35:40larger system?
- 35:41Can it also then go talk via MCP
- 35:44to my Epiphan device to get more room
- 35:48status of what's going on?
- 35:50Yeah, yeah.
- 35:50It favors the curious and the people
- 35:53looking for efficiencies.
- 35:54They'll find a way to this.
- 35:56It's very natural.
- 35:58There was a question from our driver.
- 36:00Yeah.
- 36:00He was asking if it's able to communicate
- 36:04with other devices on the network or is
- 36:05it just things that are physically
- 36:06connected to that?
- 36:07Well,
- 36:08the nice thing about a device like an
- 36:11Epiphan Pearl is we can see a lot
- 36:13of signals coming from the network.
- 36:14So we can tell all of the
- 36:16the IO that's happening.
- 36:17So we won't tell when an HDMI signal
- 36:19is there or not.
- 36:20We can also see what's happening in a
- 36:21camera image.
- 36:23So we can tell you if a camera
- 36:24is black,
- 36:24maybe somebody took the camera like
- 36:26faculty like to do and pointed at the
- 36:27ceiling.
- 36:28If that has happened,
- 36:29we can tell that and we can tell
- 36:31when an audio signal drops.
- 36:32So we can see inherently in terms of
- 36:34audio video signals,
- 36:35we could also see into the network.
- 36:37If there's a network performance issue,
- 36:39we can see it.
- 36:39We can't go in and control that network
- 36:41necessarily unless it's our configuration
- 36:43of it.
- 36:44And then the other thing that we can
- 36:45see
- 36:46which would be helpful,
- 36:47is to see the schedule coming in from
- 36:49some scheduling tool.
- 36:50So if our device is not seeing a
- 36:52schedule, we're going to say, well,
- 36:53you booked this room here,
- 36:54but there's no schedule on the device for
- 36:56recording.
- 36:57So we see peripheral services and
- 37:00equipment.
- 37:02But the vision we're sort of more excited
- 37:06about is when you have an MCP for
- 37:07Epiphan and an MCP for everybody else in
- 37:10that room.
- 37:10And then you have full control.
- 37:13And you can just stay home.
- 37:15Somebody had to go first, so congrats.
- 37:17And that was going to be my response
- 37:19to Cam about that is that, yes,
- 37:23maybe the Epiphan via MCP can see a
- 37:26lot of other devices,
- 37:27but once those other devices are also
- 37:28exposed to the network,
- 37:29then in theory they can all talk,
- 37:31so why not?
- 37:31I mean, I think the bigger thing is,
- 37:34are we seeing the beginning days of the
- 37:37death of the big centralized control
- 37:38system?
- 37:39If everything's kind of equally on the
- 37:41same level, exposing its capability,
- 37:42and you just have these agentic...
- 37:44capability,
- 37:45like must we have this ring leader in
- 37:47there doing all that stuff anymore?
- 37:48I don't know the answer to that.
- 37:49Yeah.
- 37:50If the ring leader does a really good
- 37:51job,
- 37:51the thing is like building homemade
- 37:52tooling, you know,
- 37:54that's sometimes really, really great.
- 37:56Other times you build just spaghetti,
- 37:58like it just looks awful and you realize,
- 38:00oh, I should,
- 38:00I should actually buy a proper service
- 38:02that has figured this out.
- 38:02Who's actually does UX and knows how to,
- 38:05how to make good tooling.
- 38:06So I think the world for software
- 38:08controllers will exist if they just
- 38:10continue to do a good job of it.
- 38:11But I understand the temptation to just
- 38:13like
- 38:13I think it's just more open right more
- 38:16options on the table it's not the same
- 38:18three cast of characters that it once was
- 38:20exactly
- 38:22I love it.
- 38:23Um, so George,
- 38:23you guys have an event coming up,
- 38:24I think, is it next week,
- 38:26two weeks where you're going to dive into
- 38:28kind of how you're,
- 38:29how you're leveraging MCP through your
- 38:30services.
- 38:31There's a webinar and all that coming up.
- 38:32Do you not sure if you have that
- 38:34information in front of you?
- 38:36I should have that date coming up because
- 38:37we will be standing an hour talking about
- 38:40this.
- 38:40Um, so we run these webinars, um,
- 38:44and on November,
- 38:45that's how long it took me to look
- 38:46that up.
- 38:47We'll be talking about our Perl systems
- 38:49and how AI can complement them.
- 38:51And the title is Learn How to Use
- 38:54Your Fleet.
- 38:55So that'll be at two o'clock Eastern time.
- 38:59And if you visit our website, epiphan.com,
- 39:01you can find it there.
- 39:03Awesome.
- 39:04There you go.
- 39:05And I'm assuming this is kind of,
- 39:07I'm hopeful that I'm going to put you
- 39:08on the spot here,
- 39:09that this is kind of step one in
- 39:12many things like this to come,
- 39:13kind of building on this capability.
- 39:15You guys aren't one and done and you're
- 39:16out.
- 39:17Absolutely, yeah.
- 39:18This is kind of the first version.
- 39:19So what you'll see from us next is
- 39:21you'll see the ability to use,
- 39:24so right now we're saying go use,
- 39:26you know,
- 39:27Clod or ChatGBT and then connect it into
- 39:29Epiphan.
- 39:30So that's sort of a two-step process.
- 39:32there's another simpler version where you
- 39:35can just log into our cloud and have
- 39:36AI capabilities within the cloud.
- 39:38So that's what you'll see around the time
- 39:40of ISE in February next year,
- 39:42where you won't need to use ChatGPT.
- 39:46You don't have to have that licensing for
- 39:47that software.
- 39:48You can just rely on our own built-in
- 39:50AI within our own cloud.
- 39:52That's the next step on the roadmap.
- 39:53And then there's some bigger,
- 39:55wilder things coming hopefully by next
- 39:58June.
- 39:59Awesome.
- 40:00Bigger, wilder things.
- 40:01Looking forward to that.
- 40:02Yeah.
- 40:02Halloween.
- 40:03Yeah.
- 40:06Well, I guess all that to be said,
- 40:06if you have Epiphan pearls,
- 40:08get them updated and start messing around
- 40:11with this.
- 40:11Sounds like a very cool capability.
- 40:12If you don't have Epiphan pearls,
- 40:13it sounds like you need to go out
- 40:14and buy a whole bunch of Epiphan pearls.
- 40:16And to be clear,
- 40:17that capability is available today.
- 40:20Correct.
- 40:20Yeah.
- 40:20Yeah.
- 40:22Yeah, you guys announced it last week,
- 40:23but you've been working on this for a
- 40:24while.
- 40:24This is available right now.
- 40:26This has been available for about a month,
- 40:28actually.
- 40:29Yeah.
- 40:29But we're talking about it right now.
- 40:31Yep.
- 40:31Breaking news.
- 40:31Way behind.
- 40:32Wow.
- 40:33Yeah.
- 40:34Well, we were kind of sneaky about it.
- 40:35We put it out into the world.
- 40:37Didn't even tell people what was going on.
- 40:39Yeah.
- 40:40Cool.
- 40:40Well,
- 40:40I'm glad we got to talk about it
- 40:41because I think it is definitely...
- 40:43Well,
- 40:43I'm excited to hear how people can use
- 40:45this, and not just us,
- 40:46like how people can embrace this kind of
- 40:49approach with whatever stack they've got
- 40:52going on, because it's early days,
- 40:53but it's pretty exciting.
- 40:56Yeah, it's a very cool approach.
- 40:59Even simply leveraging your hardware as
- 41:01that in-between layer that can get us so
- 41:05much data about what's going on in our
- 41:06spaces already without having us to go in
- 41:09and rewire a whole bunch of things.
- 41:10It's already there.
- 41:11It's already doing its thing.
- 41:12So why not?
- 41:12Yeah, yeah.
- 41:14Very cool.
- 41:15All right, George,
- 41:16anything else you want to bring up?
- 41:17Any last items?
- 41:18Anything else we want to make sure we
- 41:19point people towards?
- 41:21No,
- 41:21I just we'd love feedback on this because
- 41:23it's new and because we're learning about
- 41:25it.
- 41:26We need to hear what kind of applications
- 41:27people actually want to solve.
- 41:28And I hear some funny ones.
- 41:29Someone's like,
- 41:30I want to be able to know when
- 41:31people are cheating in my classrooms.
- 41:33You help me with that.
- 41:34I'm like, kind of.
- 41:36So these are the weird things that we
- 41:38like to learn about.
- 41:39So if you have any of those,
- 41:40let us know.
- 41:41And we'll either tell you you can solve
- 41:43them already or maybe we'll figure out how
- 41:45to solve it.
- 41:46So they're asking AI if students are
- 41:49cheating with AI.
- 41:50Yes.
- 41:51Yes.
- 41:54You got it.
- 41:55And they are.
- 41:56The answer is yes, they are cheating.
- 41:58But are they cheating or are they simply
- 42:00learning more efficiently?
- 42:01They're just being resourceful.
- 42:04Cheating.
- 42:04Yeah.
- 42:05That's all it is.
- 42:07Okay.
- 42:08Cool.
- 42:08All right.
- 42:08Well, George, appreciate the time.
- 42:10This is very cool stuff.
- 42:11Love to see where this is going in
- 42:12the coming months.
- 42:13I think it's a very clever use.
- 42:14And I just love seeing the innovation here
- 42:17of leveraging new tools and new
- 42:19new resources to,
- 42:20especially on stuff you may already have,
- 42:23to just build better learning spaces.
- 42:25This is fantastic.
- 42:25Fun time.
- 42:26Yeah, yeah.
- 42:27It's great.
- 42:27Thank you for having me.
- 42:28I've enjoyed it.
- 42:29All right.
- 42:30Well, we're going to roll out.
- 42:30We'll be back next week with more nonsense
- 42:32on whatever we're talking about then.
- 42:36I don't know.
- 42:37Probably more AI MCP stuff.
- 42:38Mark will talk about AI for ten minutes
- 42:39regardless.
- 42:39Come on.
- 42:40Great.
- 42:40My job.
- 42:42Thanks for putting up with yet another
- 42:43episode of AV Superfriends On Topic.
- 42:46You can contact us with questions,
- 42:48topic ideas,
- 42:48or general complaints at
- 42:50avsuperfriends.com.
- 42:59Nice work, everyone.
- 43:00Sharp broadcast.
- 43:01Really good.
- 43:02Everyone on the floor as well.
- 43:04Really a lot of hustle.
- 43:05I liked it.
- 43:06The opinions expressed by the AV
- 43:07Superfriends are solely those of the
- 43:09individuals and do not represent their
- 43:10respective institutions, organizations,
- 43:12companies, or clients.