Podcast
Join Viktor, a proud nerd and seasoned entrepreneur, whose academic journey at Santa Clara University in Silicon Valley sparked a career marked by innovation and foresight. From his college days, Viktor embarked on an entrepreneurial path, beginning with YippieMove, a groundbreaking email migration service, and continuing with a series of bootstrapped ventures.
Links
Follow Me
Follow Me
Join Viktor, a proud nerd and seasoned entrepreneur, whose academic journey at Santa Clara University in Silicon Valley sparked a career marked by innovation and foresight. From his college days, Viktor embarked on an entrepreneurial path, beginning with YippieMove, a groundbreaking email migration service, and continuing with a series of bootstrapped ventures.
Lessons in Building Niche Hardware and Networks with Planet Computers' Marcus Taylor
Play On
02 SEP • 2025
59 mins
Share:
Marcus Taylor’s journey into technology spans over three decades, but his latest work at Planet Computers represents a fascinating convergence of everything he’s learned about hardware, networks, and resilient systems. As co-founder of Planet Computers, Marcus brings a rare operational lens to building niche devices that serve specific communities.
Marcus’ story begins with IBM labs and Logica, where he explored natural language processing and early mobile wallet prototypes in the mid-90s. Those experiments laid the groundwork for later projects at Ericsson, Digital Mobility, and the GSM Association, where he navigated the politics of telco standards and helped bring Bluetooth and Wi-Fi into the mainstream.
The conversation then turns to Erlang, the programming language behind WhatsApp and critical banking systems. Taylor recounts how his consultancy supported Ericsson, financial institutions, and global telecom providers in building fault-tolerant infrastructure that could handle real-world demands - from base stations to payment clearing systems.
From there, we dive into Planet Computers - the company reviving the keyboard-first form factor through devices like the Gemini, Cosmo, and Astro. These Planet Computers smartphones are niche by design, serving journalists, security teams, and technologists who need multi-boot Android/Linux flexibility and secure communications. Marcus explains the challenges of building hardware for specialized markets and the economics of competing with giants like Samsung and Apple.
The episode explores how Planet Computers has evolved beyond smartphones into network appliances for Industry 4.0 applications. Marcus details their work with AI-powered CNC machine monitoring in Singapore, using edge computing to give machinists real-time feedback on tool wear and machine performance. We also discuss their logistics tracking system using mesh networks for Brexit-related customs automation.
Taylor closes with reflections on the future of telcos, private 5G networks, and preventive health technology through his work with Invisa smart shoes. For builders, engineers, and founders, this is a rare perspective on how niche devices and resilient networks still shape the backbone of modern systems, and why specialized hardware continues to find its place in an increasingly software-defined world.
Transcript
Show/Hide Transcript
And then when you arrived in London it would direct you to your how to get to your hotel.
And we also had digital maps and GPS implemented.
Right.
And this is 1998.
So basically we're living in the future by far because that's, that's today, right, with Apple Wallet and all that stuff, right?
Correct.
Welcome back to another episode of Nerding out with Viktor.
Today I'm joined by Marcus Taylor from Planet Computers. Welcome to the show, Marcus.
Hi. Thank you.
We-- so we've been talking for a while and I mean I was introduced to you guys way sometime two years ago, I think it's by now.
That's right, yeah.
From the fine folks around Canonical and were kind of, they were impressed what you guys were doing.
And as I learned more about your, what you guys are doing, I've been more and more intrigued.
So maybe we can start to introduce the viewers to who's Marcus?
What do they know?
And then we can move on to Planet Computer and all the nice, lovely nerdiness that you guys are working on.
Well, who am I so.
Well, I'm sitting here in Upstala, Sweden, which is where I was born.
And at an early age my father who worked for the UN took us off to Singapore and I ended up after that in Italy.
And then I branched out on my own after graduating from high school and went to the UK and started working or.
Well, I started working and then I realized that I didn't want to be an engineer.
So I applied to Luffrey University to do ergonomics and computer science and I was there for four years and was sponsored by stc which was standard telephone and cables.
And I used to work there during the summers and winters in their relay factory and eventually also at STL Labs and had a great time, was introduced to a lot of really good tech.
Even worked on the icl, one per desk which was a precursor to if you like modern day communications.
And wrote a couple of books about, together with my manager about how to design CRT displays or cathode ray tubes.
Glass of the past.
We don't miss lugging those around.
We don't actually.
We, we.
ITT had then started making the first LCD screens.
So that was quite interesting.
And then I got a job with IBM at, in leading a, at their research center there and worked on natural language interfaces to databases.
Right.
The project was called Sylvia and we used the RISK architecture and the AS 400 and the OS and VM and all that good stuff.
And I mean it was Incredibly laborious at the time because you needed so much power to just even turn a statement like tell me who all the blue eyed employees are in the company and then turn that into an SQL statement.
But in many ways that was a precursor to large language models because we had linguists and everything working with us going through the grammar and were doing it Swedish and English and Italian and it was really interesting and seeing how people reacted to having a natural language interface to data.
When was this just timeline wise?
So people on the.
Like this is.
This is 1993.
95.
Oh w. Yeah, that's some time ago.
And then to distract myself I joined the Stockholm Players and we used to go to.
Was part of a number of theater productions and things like that.
So that was kind of fun.
And played rugby for Stockholm, blah blah.
That was good fun.
And then I moved to Hanover and worked on a project to.
That was together with Dasso for their CATIA CAD CAM program.
And it was again to create not so much natural language interfaces, but to automatically create the bill of materials from a drawing and allocate all the part numbers and so on.
And so we did.
That's very timely because that's.
We're kind of full circle on that now again.
Yeah, yeah.
And.
And so we used to do a lot of usability testing and I designed a protocol for doing that and actually won an award from IBM for my efforts there, which was kind of fun.
And.
And then they whisked me off to Rome to work at the lab there where I was working on different communication protocols like TCP IP and Token Ring and so on.
I haven't heard that.
I haven't heard Token Ring in a while.
That's.
Y.
So yeah, I've got my gray hairs.
And again we're working on different ideas for.
So they were very keen on also making software based on usage.
So you know, if you had a huge installation of CAD stations and so on, then you would only be billed for the amount of usage and computing power that you used.
Which I suppose for IBM was really going back to its origins.
Well, that's mainframe.
Mainframe thinking, right?
Exactly, yeah.
Which is kind of what cloud thinking is today as well.
True, true, true.
So then they decided to shut the lab in and because this is when IBM was going through all its kind of, should we say navel gazing.
What are we about?
What should we doing?
Ms. DOS is overtaking OS2 and so on.
So I applied for a job in the UK and joined a company called Logica in Cambridge.
And that was really exciting because the team in Cambridge was basically very multidisciplinarian.
So we had projects going ongoing for air traffic control systems, natural language interfaces, speech synthesis, and also at the time, of course, 95, 98, which is when mobile phones would be coming to come onto the scene and Netscape and web browsers and all that good stuff.
And smart cards were also becoming very much part of the game.
And so we basically did a Venn diagram and said, well, surely the sweet spot is Internet, smart cards and mobile phones.
And so we came up with this proposition called Mobile Commerce.
And went to Barclays and Nortel.
This is like WEP days, I guess.
No, this is pre web, this is pre wap.
And so basically we had a project go though with cellnet, Nortel and Barclays, whereby basically the phone became your wallet and we proved that we could do all the transactions securely, etc using ussd.
But the outcome was nobody could decide who owned the customer.
Was it the bank customer, was it the mobile phones customer or was it Nortel's customer?
Right.
So it kind of basically developed into nothing other than being a great technical proof of concept.
And then during that time we also worked with Chris Bonnington.
We've set him up with a satellite phone and a digital camera and as he was sending K2, he was stopping every now and then and sending pictures back and a little report on how the ascent was going.
I think it was one of the first applications of combining mobile phones and satellite comms.
And then we did another project with Her Majesty's Revenue and Customs, using the Newton, and that was basically to replace all the paperwork.
Right.
And.
And so there was a really quite exciting time.
And because of that, because of the work we'd been doing with mobile telephony and so on, they moved me to the Stockholm office again and I ended up basically becoming like the Ericsson account manager.
And we ended up being creating a consortium together with AU Systems in Lund, Amadeus in Stockholm, SAS and Ericsson for ticketless travel.
Right.
So we took one of their prototype phones, which was basically a PDA with a touchscreen, and implemented smart card with it.
And so you would go to your mobile phone, there was an app there, you would book your travel based on what your commitments were and then it would download the ticket, you'd go to the gate, you'd scan the phone that would let you through the gate and then when you arrived in London, it would direct you to your how to get to your Hotel and we also had digital maps and GPS implemented.
Right.
And this is 1998.
So basically we're living in the future by far.
Because that's, that's today, right, with Apple wallets and all that stuff, right?
Correct, yeah.
So, and there's a funny story actually.
When we first got the app running, this was going to be like the keynote for a kickoff meeting for the Ericsson executive team out on Sandham.
And so were debugging and debugging.
I was sitting in my sauna office and for two days straight I kept calling me, is it ready?
No, it's not ready yet.
We still got a few bugs to iron out.
And then midnight the night before the event, we finally got everything sorted out and then I was like, how do I get from Solna to Sandham in four or five hours?
So I just took a chance and drove down to Gamla Stan and by happen chance there was a guy in a speedboat there and I jumped on his boat and said, will you take me to Sandham?
And he said yeah.
And so I, I paid him and we set off at 40 knots across the water and then we got to the pier in Sandham.
Then I reached out with this precious, grabbed it and I was having visions of it go pl into the water.
Fortunately it didn't.
Oh wow.
And then, and then because of those events, myself and NOWV and a guy called Bank Carlson bank was at Ericsson, a director in one of their mobile divisions.
NOWV was at Imperial College, but it also started a company called Dynamical Systems Research and myself, we sat down one day and said, well, it's the dot com era, why don't we just, you know, get a move on.
So we founded Digital Mobility and set up in London and we got things moving quite quickly, raised quite a bit of money and we had an alliance with Guernsey Telecom because at the time in the UK you couldn't roam and you couldn't send SMS, for example, between Vodafone and Cellnet.
So but if you had a Guernsey Telecom sim you could, so there was a big advantage that, so we put together a package, the Digital Mobility package, which was aimed at high net worth individuals, frequent travelers and so on, which combined travel, banking, telecoms and so on.
So, and so I being in charge of partnerships and business development then had to go out and form partnerships with the banks and the travel companies and so on.
And they were all very receptive until we hit the usual snag who owns the customer?
So nobody seemed to be willing to kind of divvy up the transactions in an equitable way.
Everybody wanted more than was satisfactory, but we sold it on and were beginning to get results.
And then were about to get an investment of $7 million from Lehman Brothers and at the 11th hour and 59th minute and 59th second, they withdrew from the.
And so were left, you know, what do we do now?
And so we ended up having to basically sell off the assets of the company.
We sold our communication and Java Applet distribution system to Tencent in China.
And.
And then what was left we sold to Norway to Shipstead.
And fortunately I still had my shirt on my back.
And.
And shortly after that I joined the GSM association as a technical director and was responsible for mobile phones.
And that was really interesting, understanding the politics behind how all these operators work together.
So I certainly had an inside track on that.
And so one of the things I was responsible for in my new division was WI FI and Bluetooth technology.
So the operators were like, is WI fi going to cannibalize our business or can we find a way to collaborate with these new mushrooms of WI FI activity that are popping up everywhere?
Right.
And so through 3GPP we came up with a standard which enabled roaming between WI fi and 3G at the time.
And then the same thing with Bluetooth.
The Bluetooth SIG was very keen to work with the operators to find application areas and so on.
And with all the various implementations.
And again, mobile operators were kind of wary again because of how it might affect their business, rather than seeing it as an enabler, which we are benefiting from right now by having excellent voice communication.
Yeah.
And then I, I was asked to do a press release about Bluetooth technology and Bluetooth Stig, which I did.
And then Vodafone got really upset, so I had to leave the GSM association for some bizarre reason.
I hadn't got all the right ticks in the boxes.
But that was no big, big deal.
Landed a great job with a Swiss company after that called sickba, based in PRI or near Geneva in Switzerland.
And SIKBA are the world's largest producers of pigments and inks that are used in banknotes and they'd spun off a separate division called Product Security, where they use their know how in pigments and fluorescent pigments to create a technology for preventing counterfeiting and smuggling.
And this is used actually even today by big brands like Gucci and all those fashion Brands.
But also we had a.
They just landed a project with the Malaysian government to prevent counterfeiting and gray imports of cigarette packets.
So I was made the general manager of the new division in Malaysia and went out there, recruited a team of about 40 guys and girls.
And the system was based on a, what were they called now?
Anyway, an inkjet printer that we put in line on the packaging line of the cigarettes.
So you've got these packets of cigarettes flying by hundreds per minute.
And this inkjet printer had to print the dot exactly in the right spot on the cigarette packet before it was wrapped in cellophane.
So it's a watermark essentially that's unique to each.
Yeah, so it was a multi level security.
So you had your geometric shape, then you had your fluorescent ink, and then you also had a, if you like, a micromolecule with a very specific 3D shape that Sigma made.
So this was all part of the ink that was used.
And so the idea, so the, basically it worked like this.
The cigarette packets went out into the market, then the customs officers had a special little gun which would illuminate this dot and then if it pinged back the right frequency that the reader would give a green light and you would know that it was a locally produced cigarette packet.
But if it was red, obviously it had been imported from without duty paid.
And so, but of course, like anything, eventually some people will figure out how to analyze the ink and come up with something similar.
But this, if you like, secret molecule was always the, if you like, the forensic test which would determine whether it was genuine or not.
And then, so we had seven packaging factories in Malaysia, scattered all over the country, jti, Philip Morris, et cetera.
And we had to distribute the ink.
So the whole business model was based on, we supplied the equipment, we gave it to them a bit like Tetra Pak business model.
And then the consumption of ink would determine, so how many dots they printed per month would determine the.
In.
In the revenue that SIGVA would generate from that.
Okay, interesting.
And then, so then I telemetrized, if you like, all the printers.
So I put a little GSM modem in each one.
So we would be then centrally recording all the dots so we didn't have to go out and take all the numbers off the printers every week or month.
Yeah, and that was a bit.
And that also enabled us to determine or predict when the printers needed cleaning or maintenance.
So that went quite well.
And then during that time I got married and my wife was pregnant and Sickba was so impressed with what we'd achieved that they then landed a project in Brazil and they wanted me to go to Brazil to do alcohol and beer bottles.
But my wife put her foot down and said, no way, Joseph.
So, so then I was back in the uk.
I resigned from my work there.
They'd given me a fair settlement and I was then looking, trying to think, what do I do now?
And then I went to an alumni meeting for my old school in Rome, St. Stephens, and I bumped into a guy who was about 10 years younger than me, Francesco Cesarini.
And we got talking and he said, I've just been made redundant from Ericsson.
I, I used to be part of the Erlang team and I'm starting up a consultancy here and I was like, well, why don't we join forces?
I mean, I've got a lot of experience working as a management consultant, system integration projects, blah, blah.
So he said, yeah, let's do it.
So we got an office in Spitalfields and one of the first projects we did actually took me back to Malaysia because I got to know quite a bit of people and had a bit of a network in Kuala Lumpur.
And one of the guys I knew, Shukri, he was responsible for something called mihas, which was Malaysia's attempt to become, if you like, the center of halal food in the Far East.
And he said, we really need to find a way to create traceability for halal chicken and halal food.
So based on, if you like, the work I'd done with mobile telephony and Ericsson and with sigpa, I said, well, I don't see why we can't come up with a system which would enable us to put markings on the chickens and the chicken produce using, if you like, using something similar to what we've done in the past.
So we put together a solution and we presented it to the Malaysian government and so on.
But then of course, as usual, lots of people want a bit of the pie and so it become an impossible business opportunity to actually execute on.
But that was kind of fun.
And then so the consultancy started growing, we started getting more.
So Erlang.
Erlang was the language that Ericsson used for their system, sorry, Axe systems and their base stations, the GSM base stations.
And why did they do that?
Because Erlang is, if you like, a virtual machine that is super resilient, very scalable and fault tolerant.
So, you know, even if base station goes down, it will then reboot and restart.
And because there was so Many out in the field, Axe exchanges, base stations.
Despite the fact that basically shut down the Erlang team to its spare skeleton crew, they still needed people to carry on development, coming up with bug fixes and so on.
So we ended up supplying them with consultants to do what they had originally invented.
And, and so we did a number of different projects and then because of the features of Erlang, we got involved with WhatsApp.
I don't know if you know WhatsApp is based on Erlang.
Yeah, we got projects with.
What do.
You call them now, Vocalink.
Vocalink is the exchange center for all the clearing banks in the uk.
So when money is sent from one bank to another, it goes through the vocal link.
And so they had a great big IBM system to do this and they asked us to reverse engineer this to run on Linux boxes using Erlang, which we did successfully and then they sold that to the bank of Singapore.
So.
And that went well.
What other projects did we do?
We did ticketing systems, gaming systems.
Yeah, it is very versatile language.
And then I was getting to the point where I thought maybe there's something else I want to do in life.
And we'd been approached by a Danish company called Trifork and Trifork is a Danish system integration company in Copenhagen, but they're actually in Zouk in Switzerland and they're listed on the stock exchange.
So I met the owner of Trifork and he said well, would you be interested in selling your shares to Trifor?
And so we can really helpang expand and so on.
And so yeah, he made me a good offer which I couldn't refuse.
And so I took that on and moved on from there.
And then after that I invested some money in a company called Silent Sensors which I helped co found which was how to make tires smart using, if you like, the beginnings of edge computing.
So this was because of the legislation requiring TPMS systems in cars and vehicles to monitor tire pressures which are fundamental to basic car safety.
And then I.
So we built up silent sensors over time.
Originally were using, if you like, an RFID solution which required quite hefty antennas and infrastructure gates and so on, which was a bit of a non starter.
So then went back to the drawing board and said well, we need to provide our own power source because we at this point we'd managed to interest some major tire companies like Bridgestone and Continental and so were looking at energy harvesting, so we found a piezoelectric solution and then we found a solution for Solid state batteries.
So we created, if you like, a plaster or a flexible bit of electronics.
And were based up in Sedgefield as part of the cpi, the center for.
What was cpi.
I can't remember what it stood for.
But basically they had all the equipment necessary for flexible electronics production.
Right.
So we started making prototypes.
We also interested MRF india and the project went very well.
But because of the, if you like, the conservative nature of tires, it was necessary for them then to run basically millions of miles with the prototype tires to evaluate and determine that they were safe.
Not very startup friendly.
No, no.
There's a sort of.
The chasm of doom and destruction became insurmountable in some respects in terms of being able to get to a point where the investors felt they were going to get a return on their investment in a reasonable time frame.
But silent census is still going.
I kind of browned out.
I had enough of this and so I sort of said bye.
Still got shares in the company and they're still going through the motions and perfecting it.
And maybe in the next decade it will come to light.
But, but in the course of this, I, I again made a number of very interesting contacts.
One of them is Invisa, which is a smart shoe solution, which is basically an activity tracker in your shoe.
Again using piezoelectrics and solid state batteries for the power source.
And also because of the solid state battery requirement, I found a company in Poland called the Batteries that can manufacture both the battery and the energy harvester in scale.
So that's currently what I'm working on.
At the same time I'm working with Planet Computers making mobile phones like this one with a keyboard reminiscent of the Scion.
Yes.
And if you give me one second, I will bring another little prop.
Hang on.
Yeah, right back.
No worries.
So this is what we call our xr.
It's not powered up now, but basically this is a color touchscreen which you can program and customize to customers requirements or basically as a out of the box solution.
And we also prototyped some RFID rings which you could use for making transactions and so on.
Before we dive into that, I want to take a step back and like early days, Planet Computer, like you started with Gemini, maybe like take me back to the early days there.
So people are not familiar with your kind of product portfolio and your backstory of how you came to be.
Really like, what was the starting point and kind of what did you set out to do in the first place?
Right.
Well, Basically the whole Planet Computers proposition is based on, if you like, the ergonomics of a keyboard.
The keyboard is a very efficient way to communicate compared touchscreens or voice communication.
So.
And the reason for this was also because the, if you like Imperial College.
So the founder of Scion was a professor at Imperial College and I didn't know that.
Oh yeah, yeah, South African got my name.
I, I can see his face but I can't remember his name, but it'll come to me.
So NOWV and he became friends and then when Scion basically shuttered because it was overtaken by events and I mean then we can go into the whole story about Symbian and so on and so forth, but they weren't able to, if you like, keep pace with what happened with smartphones and integrate 5G, 4, 3G, 2G into their phone as their PDAs.
But the basic mechanics and engineering of that PDA was still very valid if one wanted a pocketable PC, if you like.
So NOWV did a deal and Scion then gave him the rights to continue developing this.
So in order to get it to market, Planet Computers then did a crowdfunding using Indiegogo, I think, if I remember rightly, and that was successful.
And then they, so we had the Gemini and then we had the Cosmo and then we had the Astro.
I mean, yeah, and this, what Planet Computers never really achieved was if you like, enough investment to do not just the design, development and production, because as it turns out, mobile operators, they don't want to pay for the marketing, they want to do the distribution and so on, but they don't want to do the marketing.
So if you want to sell a million phones, you need to have a marketing budget, probably a million pounds to get yourself in front of everybody else.
And then of course we've got the big players like Samsung and Apple and so on, that so huge that they basically the, the effort, if you like, for a mobile operator then to sell these sort of niche phones is the benefit to them financially is insignificant really.
So they don't really give you much time.
So yes, you can sell them over the Internet and so on, but you're constantly hitting that crystal ceiling, if you like that glass ceiling in order to get the volumes.
Because then, because the people that are producing it, they don't want to produce 10,000, they want to produce 100,000 or a million a year.
Yeah, they call it Moq Battle, right?
Correct, yeah.
So we're still Struggling on if you like.
And we're developing a new version with a better processor, 5G.
We're also looking at NTNs to make it satellite compatible, et cetera.
Right.
But that's still.
If you like.
But talk to me about.
Because the reason why I think it's interesting is a.
It's kind of carries on scions kind of like initial vision and obviously the form factor obviously which is I guess makes me nostalgic.
But beyond that there are some interesting use cases.
I mean we spoke about last time with Janko in London.
Like we spoke about some interesting use cases where these devices are being used.
Right.
Like they are not necessarily your consumer use case because I think that's obviously a very difficult market to get into and yeah, least from the pricing point of view with MOQ problems.
But talk to me, what about like the use case guy?
He manages.
I'm not sure how much what you can mention on there, but he mentioned some real interesting use cases for like these applications or application of these phones in the world that you've seen already today and your customers.
Well, I mean I don't know the most.
The typical use case is if you like your business user that uses Excel and Word and so on and doesn't want to lug a laptop around with them.
He wants something you can stick on the inside pocket of his jacket.
And from there you've also got a few examples of journalists use them quite extensively.
There's some authors that use them quite extensively because it enables them to, you know, when something pops into their head they can just whip it out and write a paragraph or a chapter on the train or wherever.
So those are the, if you like the typical use cases.
I mean I would say 95 of anybody that's bought a planet smartphone is.
Is a male between I guess 40 to 60 years old because of the fact that they either had a BlackBerry or a Scion.
The, I mean the other use cases are if you like for homeland security and so on because it's multi sim we can customize it with different types of encryption so we have customers in the Far east that use it for their police and so on.
But that's about, you know, as wild as it gets.
No, but the reason why it is interesting I think is.
Is the multi boot thing is an interesting.
Oh, you mean.
Okay, sorry, yeah.
You mean being able to switch between Android and Linux?
No, no, I meant more like the use case for that.
Right.
Because that in turn means that there's a very small set of users that would require one such feature.
Right.
And that's a very unique proposition.
I think.
I haven't seen anything out there that has similar functionalities in the same form factor at least.
No, no, yeah.
I mean that uniqueness is perhaps something that should be leveraged more.
But on the other hand, when customers do require it, they want it in quantity and at this point in time we're just not able to do that.
Yeah, yeah, yeah.
Because that's, I mean from a use case perspective that opens new doors.
In particular, I'm thinking I just came back from defcon and like, obviously that's a space where I have a lot of security people there and like being able to just whip out a tablet and boot up tally or whatnot.
Right.
And some Linux distro and like do pen testing on the go on a small device is very covert.
I guess that's a cool use case for these devices, right?
Absolutely.
Yeah, yeah.
But yeah, I mean the whole cyber security thing is again something that we get pulled into.
But, and that's where the smartphones that we have are part of that solution.
But also the network appliance that we've developed is also very relevant there in terms of creating a distributed network that is able to, if you like, distribute the computational requirements over a large area.
And because it's got a lot of media interfaces, camera, audio and so on, we can.
And so the current projects that we have at the moment, we have one in Singapore which is called AI Connect, and that's about using artificial intelligence with CNC machines to improve productivity by monitoring parameters like sound, vibration, heat, power consumption of a CNC machine to determine when the tool is going to wear out.
Because if you can extend the life of that and reduce the number of tool changes, then your productivity goes up.
But it also addresses the fact that there is a, a growing shortage of skilled machinists in the world.
Blue collar work is not kind of like top of everybody's list when they're 18 and looking for a job.
So the idea with the system is that it becomes, if you like, your assistant that tracks your skill level, provides you with feedback, tells you when to change the tool, tells you how to do the configuration and setup.
And so we have the, this appliance that I just showed you and that's hooked up to the PLC of the CNC machine.
So we're collecting all the data and that's also shared with the rest of the factory.
So the idea is that all the CNC machines in the factory are then connected and then upper management can see what they need to order in terms of just in time flow of parts and consumables.
But also the idea is if you like a lot of the application of AI and machine learning at the moment means that the time between something happening and something being done about it is actually the latency is quite long because it's the manager that makes the decision, not the machinist.
Whereas with our solution here, it's the machine machinist that makes the solution on the spot at the time.
So talk to me more about that.
So you connect to I mean I guess your industrial control system of sorts.
Right.
So maybe a CNC machine, this goes into whole like industry 4.0 kind of movement with preventive maintenance and all that stuff.
But you have this device that connects one way or another to machinery and then you how what do you go from there?
Do you have like a mesh network that you.
Because these are oftentimes not very connected environments.
Right.
So because life cycle these devices, machines tend to be meshed in tens or twenties of years rather than software where we have very quick churn.
Right.
So these are my machinery from quite old while that you plug into.
Right.
So talk to me more about what you've seen there, the use cases there because I think that's an interesting one.
Well yeah, you raised two interesting topics.
One is the communication.
So In Singapore certainly 5G is very prevalent and private 5G networks are growing in popularity.
So we can either use wi fi or 5G.
5G offers a lot of features if you like, that WI fi doesn't in terms of security and allocating bandwidth and so on to different assets.
And the second thing you raised about is if you like second hand machinery can be made smarter with this solution.
And there's a lot of secondhand machinery that then has a second life.
And then you've got new machines of course which also are a target, but less so if you like for where we are currently with the proposition.
So the second hand machinery, I mean you could even apply this to machines from 20 years ago because the precision is pretty much the same as machines that are on the market now.
It's just that they've got more servos and more features and so on.
But yeah, so I mean there's a lot of trade going on.
And so one of the markets we're exploring at the moment is the companies that refurbish these machines and then resell, sell Them and by also offering these add on services they you create if you like value added services which in terms of maintenance and support and so on and making sure they have the right consumables at the right time.
Yeah, I mean what have you, I mean that's the whole industrial system space I find fascinating because it's obviously there's just so much money tied to it.
Right.
Because there's obviously entire industries depending on obviously that's the underpinning of these industries at large.
But have you found like when you plug in these systems I would imagine a lot of his proprietary systems I'd imagine and there probably a lot of reverse engineering of protocols.
I mean diversity of protocols can bus is probably involved and so on support.
Like what have you seen in terms of there to actually make this work for all the nerds out in the audience watching this?
Like what if you want to get into this stuff, like what have you found?
Well, I mean basically what we do is we ingest all the manuals for the machine.
Yeah.
And then we plug it into the PLC system and as you say quite rightly, we then integrate that so that we're taking the signals from the machines and feeding them into our box along with all the data that we're collecting from our sensors.
And then in order to calibrate the machine we have a component that we make and that component, if you like, stresses out the machine.
We make it do stuff that is really at the edge of the envelope of what it wants to do.
And that generates noise, vibration, heat and power consumption.
And then we can correlate that.
So then we take the machine part off and we measure it and then we'll push it and push it until we reach a failure point.
And then we correlate the sound or the vibration or the heat signature to that point in time.
So then because this curve is sort of usually it's not kind of linear, it kind of pops along and then suddenly it kind of pops off the scale.
So we just need to focus in on that, that inflection point.
And that's if you like where the AI comes into it, the LLM that is T specific use case.
And that then enables us to give feedback to the operator saying refer to page blah blah on the operator manual to do this setting or to change this etc.
Etc.
So in a nutshell, we're bringing together all the data because if you think back 20, 30 years ago, the Machinist would be using his eyes and ears and touch to say the machine's running fine now.
And then, I don't know, it put his hand on the machine table and you'd feel a vibration.
You say, oh, need to turn it down a notch now.
Oh, that feels better.
You know, a little bit like when you're tuning a car and you want it to just tick over smoothly so you fiddle about with the carburetor until it's doing its thing.
Right.
So if you like, we're giving naive users the eyes and ears of an expert.
I mean, this is a long leap from where you started with communication.
Right.
But I mean, it's an interesting pivot, I would say, in terms of that.
So that's cool.
That's.
Yeah, I mean.
But I mean, I suppose at the heart of all this is DSP technology, digital signal processing, because that's what we're doing.
We're processing, we're turning analog information into digital information, which were then, in a way, turning into analog information for human consumption.
Right, that's fair enough.
That's fair enough.
Enough.
Very cool.
So that's where you're.
Is that where the current focus is mostly now, or is there a new.
It's a new iteration of the Astro coming out or how.
What's the, what's the roadmap looking for?
I think at the moment the Astro is kind of on the shelf, if I like, if you like.
And our focus is really on the network appliances because we see there it's an easier sell.
We sell, if you like, to corporates or cloud users and so on.
And we've got.
Got a lot of interest in, in Asia, thanks to what we're doing in Singapore and Japan.
I mean, just looking back at the Astro and the range of phones that we did in the past, Japan was always one of our best markets because really they.
Yeah, because they like being able to type in kanji fast on the keyboard.
So that's just an aside but interesting from an ergonomics point of view and usability point of view.
The other thing we're doing is we've also got a project in logistics using a radio technology from a Danish company called Neocortec.
And this is for tracking assets like the containers and pallets and packages.
And the use case is driven by Brexit.
Because of Brexit, actually red tape has become even bigger for the UK market than it ever was before.
So the basic concept that we're trialing at the moment is we have our Neocortec mesh network tags in all the palettes and when the pallets are loaded with the goods and products, they're all scanned and associated with this palette.
So when this palette moves from, I don't know, Istanbul to London, when it's loaded, that bill of lading is sent to HMRC into the airline, the pallet is loaded into the plane.
The other cool thing about Neocortec is that if there's no gateway present, all the mobile nodes go silent.
They're just in listening mode at this point and they wake up.
You can set the frequency at which they wake up and that will determine this lifespan of the battery.
But we've tuned it so the lifespan is about seven years.
Oh wow.
And, and so the pilot then arrives in London, it's loaded off the plane and there's a gateway at the airport and the all the little nodes wake up and HMRC then gets notification that this has happened.
And this initiates digitally all the transactions that are required in order for the goods to be imported into the uk.
Payment of taxes, documentation, if it's, I don't know, animals or whatever it might be.
So that means that effectively as soon as it's taken off the plane, the goods are cleared for distribution to wherever they're supposed to go in the uk.
That's very cool.
That's very cool.
So we're still using the distributed network concept here.
So you've got the nodes then you've got the gateways.
The gateways are connected to an XR and that is if like storing and aggregating the data and then redistributing it.
So it also means it's resilient and fault tolerant.
So if the 5G network goes down or whatever, we've still got all the data locally as soon as the network comes back up.
So this can be applied then to sea freight, air freight, railways and so on.
Yeah, very cool.
I mean one thing I want to kind of look back into as we kind of wrapping up here is 5G and you've been in the telco industry for a long time and I, at least I've just started to pay more and more attention to the telco world and kind of the commoditization I guess of the telco market in the sense of how things are moving to software.
I've done episode before on 5G and Magma and all these like open 5G networks and open source kind of et its way into telco world.
I'm curious about your vantage point there, how you see that and how you see that changing.
I know there are plenty of big telcos are doing pilots out there for this stuff.
I'm not sure that's something you've been paying attention to.
I have to say.
I'm not as a, as you are perhaps, but I mean as far as networks are concerned, with the thing that were concerned seeing as a, a unique proposition, was to implement NTN capability to enable the phones to work over the satellite networks.
But I mean, because the, the Astro is basically an open platform, I see no kind of obstacles as to enabling it for other networks.
But that said, it's always been problematic to make the Astro a global phone because of the so many flavors of 5G and all the different frequencies and so on.
And that has also been a bit of a, if you like a bugbear is how to keep abreast of all the various changes that are being made.
So, yeah, I mean, I suppose it's great in some ways that there's all these different flavors of communication, but it kind of also moves you away from the spirit of what three GPP and GSM was all about from the beginning, was to create this universal ubiquitous network to facilitate communication.
Yeah.
Do you see a world where 5G or for the, whatever, 6G maybe, but the replacement of this will be as kind of a replacement for WI fi where like you will see much more of the private instances of these.
So instead of you spinning up your own 5G route, instead of spinning up your own WI fi router at home, you're on your own private 5G network because obviously you need.
Telecos do not always have the best coverage wherever you live.
Right.
So where do you see the world changing in terms of that?
Yeah, it's, I mean, why can't we have just one communication protocol?
I mean, do we need all these different communication protocols?
I suppose if we look at it from the perspective of you've got your fiber network, then you've got the WI Fi that enables you to connect so you can go globally if we go 5G, 6G.
I mean, even when we're talking 4G and 3G, there were many propositions for nano cells and microcells and so on, and they've not really made much headway.
I mean, to some extent they have, but not what you would have expected.
That's mostly because of regulation more so than tech, I would say.
Right.
Because it's very difficult to get permission for these deployments and the cost Very expensive.
Yeah.
If you want, if you want to buy a small cell, like your average consumer will not buy a small cell.
Right.
And run it at home because they're.
Way too expensive and complicated.
And I think the same applies to small to medium companies.
I mean, does the CTO or CEO have, or CTO or CIO have the experience to manage a network like this because there's so many parameters to set up and so on.
Maybe this is again where AI comes into it.
The whole process of setting things up can be offloaded onto a smart AI solution.
I mean Amazon is playing this game already in the US right?
For private 5G.
So there's a scenario where the clouds will be playing in this space.
Right.
And the cloud instance becomes just another control plane for your local 5G essentially.
Right?
Yeah.
And then on the other hand, if we look at that from the AI perspective, people or companies want to own their data.
So they're becoming more and more reluctant to have that running in the cloud and want to have it on premises and maybe even air gapped for some applications.
So you've got some interesting kind of problems.
In fact, last night I was watching a remake of War of the Worlds where the aliens basically are these data monsters that invade and infiltrate all the data centers and take over humanity by controlling all the data and take them back to the stone age.
Fortunately they just needed to put a DNA virus in and that destroys the aliens.
But, but yeah, you, you know, as soon as you start looking at this, you hit on privacy.
Just basic employment.
What, what jobs are going to be out there in the future?
Where, where should our next generation be focusing their energies at school?
There's a whole, I guess that's, that.
That'S the million dooll question, right, that I think a lot of us struggle with at the moment and I, I, I sure as hell do not have answer to that.
I have thesis and I have ideas but these, at the rate of change we are seeing, it's incredibly difficult to make any long term thinking on this.
Right?
Yeah.
Two episodes back where we talked about this with one of my friends about those.
What's the name?
Post labor economics is the term for itself.
Right?
It's, it's a fascinating space of like obviously we have a, we have concepts like ubi Universal Basic Income and stuff like that around supporting the economy.
But good point.
I mean kind of looping back to where you're saying before it's hard to get young people into what's traditional been blue collar labor.
But they are probably more future proof than being a computer.
Computer software engineer these days.
Right.
In many ways.
I mean look at electricians, plumbers, you name it.
I mean they can charge a.
A serious buck when they come to.
Oh yeah, yes.
Just try to get anything done, like anything in your house or for work.
Right, yeah, absolutely.
They.
They can come in very good day rates.
So absolutely.
There's.
There's a shortage, massive shortage of that.
So absolutely.
There's, there are.
There are far more people that probably should be spending cycles on looking into those space because they're harder to replace.
It's going to be a lot harder to build the machine to replace a toilet than there's going to be to build a machine that's going to write your backend code.
Yeah, exactly.
Or be a lawyer or whatever it is.
Right.
So absolutely.
Marcus, this has been very interesting.
Any closing thoughts, Any closing words before we wrap up?
Things people should be paying attention to.
To people you are.
Things you are paying attention to that other people should pay attention to.
Anything else you want to add?
No, I enjoyed.
I hope I haven't rabbited on too much about my career, but yeah, I mean, personally I'm at the moment very interested in health and the technology underlying that and edge computing, machine learning and so on with the project we have with Invisa.
So yeah, I think health is probably one of the areas which I feel all this technology kind of brings together in a way that gives immediate benefit and feedback to the individual and also to society.
100%.
I'm a big fan of connected smart devices.
He's got a new smart water bottle that discounts how much water I'm drinking.
If you can't measure it, you can't manage it, as the old saying goes.
Right, Correct.
I'm a big fan of that.
So absolutely.
I'm 100% on the same page.
Measure everything and then maybe you throw away 90 of the thing you record, but the remaining 10 might be life saving in 10 years.
Yeah, so absolutely.
Yeah, yeah.
Particularly with your health.
Yeah.
Shout out to another Swedish company.
Well, not it, but Neko Health is another one.
That's very cool in that space.
I did, I did their stuff a few months ago and that's super cool.
Give you a full body scan and so shout out to Neko for being very good on preventive main.
Preventive health.
So correct.
Yeah, perfect.
Marcus, thanks for coming on the show and have a good one.
Thank you so much.
Thanks, Victor.
Take care.
Bye.
Found an error or typo? File PR against this file or the transcript.