Silly Sit-Downs with Rohan (SSD-05): An interview with Miles Ward, Chief Technology Officer (CTO) at SADA Systems
Hello everyone, hope you are having a good time and enjoying your life. I, Rohan Singh, welcome you again to the new edition of Silly Sit-Downs(SSD) with Rohan — an interview series with industry professionals irrespective of their genre; the purpose of these interviews is to have insightful conversations, understand people better and get rich perspicuity into career and community.
This time we have a very exclusive person Miles Ward with us, currently working as a Chief Technology Officer(CTO) at SADA, assisting companies to accelerate their business. Miles is a very popular face in the tech industry because of his immense contribution to the industry for the last 2 decades. Before Cloud, Miles used to be a part of Visible Technologies giving power to Fortune 1000 companies to harness the power of the internet and social media during a time of the dawn of social media.
He has led Solutions Teams at Google as Director, published numerous whitepapers at AWS, helped NASA live-stream the Mars Rover landing, helped Twitter adopt Google Cloud, and helped Barack Obama For America win the 2012 presidential campaign.
Miles is also contributing fervently to the tech community, mentoring people and spreading awareness about tech at a global level. He is always willing to share his knowledge and expertise with others.
Miles is a tech professional with a passion for music. He is a talented musician who plays a variety of instruments and has a deep understanding of music theory and composition.
I would like to thank Miles Ward for taking out time from his schedule for this interview :)
Rohan: Hi Miles, a big thank you for doing this interview, I really appreciate it. How are you doing?
Miles: Hi Rohan, thanks for the invitation, I’m doing good, what’s up with you?
Rohan: Not much, just working for smooth delivery of the project and excited about this interview.
Rohan: Let me start with a question that I’m very curious about.
Rohan: Facebook changed its name from Facebook to Meta, did all the rebrandings, and launched Metaverse. When it was launched there was a buzz, and numerous articles/blogs on the internet or Reddit but now I’m barely getting any news or articles or people talking about Metaverse. Is Metaverse dead now or still alive?
Miles: There’s a book called Snow Crash written by Neal Stephenson, he’s the one who coins the term Metaverse, some 30 years ago. There have been a whole bunch of different technical implementations that feel like and behave like what he was imagining at that point and I think there’s no doubt that the leaders at Facebook have studied that thinking, and have put their heads into that place. I think Meta’s incarnation of the metaverse is totally dead and it’s very likely that’s not a company that’s gonna be able to build the level of interactivity that’s required and to solve the problems around, real-person identification and make the thing fun in a way that I think maybe some of the game companies would have a better shot at. I don’t think the idea of the Metaverse is dead. I think there’s definitely going to be interfaces that work that way going forward.
Rohan: But still, Facebook or Meta had spent billions of dollars, I guess or maybe millions to launch Metaverse and it failed. Why? Is it because the implementation was not correct or Zuckerberg doesn’t have a proper vision for that, also you mentioned that if a gaming company had launched it would have been more impactful.
Miles: I mean, I think there’s a difference between recognizing the market potential, a need, and having the ability to fulfill that and having the level of sort of vision and foresight and insight about the behavior of users to be able to set what’s next. For instance, Apple has an incredible track record of taking things that just aren’t quite cool yet and making them so that they’re cool which is effectively a fashion brand in a bunch of ways. I don’t think Facebook has ever been all that successful at being fashionable, which is a huge part of the kind of social reality of these kinds of systems where you need to build a functioning network effect. And to make it so that the experience is actually something that people delight in as opposed to sort of begrudgingly adopted because it creates some kind of big utility. So, no, I don’t expect that Facebook has much of the DNA to be able to take this on and I do think that we have to go a really long way, in terms of ease of use and fun for this kind of stuff to have it feel like there’s any kind of shot at “mainstream adoption”, right? It’s going to end up feeling a lot more like a very successful video game, where Pokemon Go is probably a much closer metaverse-type experience than anything that’s been built.
Rohan: Continue talking about Meta, they have launched Threads, a competitor of Twitter and people are talking about it out there and there is a kind of serious discussion going on that Threads is a Twitter Killer on Reddit, YouTube, and Medium. We also have Bluesky, launched by ex-co-founder of Twitter Jack Dorsey and now we have Mastadon. Is this end of Twitter? What do you think?
Miles: In every technical system that I’ve ever seen, birth and death take a lot longer than their biological equivalence, I mean, MySpace had millions and millions of users, up until just, a couple of years ago, so rather than birth and death, let’s call it it has a clear path to growth and continued existence versus not. When I was at my social media startup, which predates the existence of Facebook and Twitter, we were describing all the different sorts of characteristics of communication media: is it public or private? Is it for an authorized pre-set list of individuals as consumers? Can everybody write or a certain number of people write? Do you have to pay to write? Is it free? Do you have to do technical work? Do you have to do hosting or infrastructure or not? Every one of the conceivable cross-sections in this grid has a product. Twitter is public, and everyone can write and you don’t have to do the infrastructure. But the length of the text is short, video is sort of a third-child difficult thing to integrate. other platforms select different options in each of those building blocks.
I think Twitter found a structure that people wanted from a sort of usefulness standpoint: short enough that you could do it while you’re on the run. Everything is consolidated into bite-sized pieces, so it’s easy to understand. I think Threads has done a good job of emulating Twitter’s behaviors which does not mean emulating Twitter’s participants. I don’t use Facebook as I know who’s on it and the people that I have connections to there aren’t the folks that I want to spend that many cycles with learning from or building, what’s next in technology or taking on new challenges, so that threads communities, maybe the wrong community to have that kind of conversation I want to have in a Twitter-style interface so that I think probably is a risk in their go-forward growth. And a hundred million users is a killer thing, but there’s a very different kind of global penetration that’s required for a tool like this to really get to Zeitgeist.
Rohan: You know I have the same thought when I read or watch something wrt this. Twitter still has the same credibility and reputation irrespective of the owner and I highly doubt that Threads will be able to achieve that credibility.
Miles: I don’t think Threads can kill Twitter, but I’m certain that Twitter can kill Twitter. It has gotten markedly worse not only in terms of operational consistency but in terms of its leader being able to actually keep it from being a hate-filled hellscape, and frankly, that’s what costs the platform the revenue from an advertising standpoint. No brand wants to be associated with that kind of vitriol, negativity, and terrible ideas.
I don’t think they’re in a position where a competitor will be so markedly better at doing exactly the same thing that they do. That it will just be obvious everybody has to switch. Content Moderation is an exceptionally hard problem, requiring really talented resources to work non-stop in a tenacious way. So, I think that there’s a lot of room for a company to do that really well, but it also takes a really long time to prove. We all have to sort of experience that it is better for months and months and months consistently before you go this place is just better. I’m gonna hang out here more often, it’s not like in order to sign up for threads I have to send my Twitter account handle through a chip or shredder and never use Twitter again, right? They are not binary.
Rohan: You mentioned that you had a social media startup. Do you want to talk about that? I’m super curious to know about it.
Miles: Sure, Visible Technologies, built in all of these words makes sense now — a social media data analysis and engagement management SaaS platform. None of those words made sense in 2006 when we built it. The only website I’d ever heard of that charged money was Salesforce. That was literally the only business that I’d heard of where you paid a subscription to have access. There was no social media, it was blogs, forums, message boards, live journals, etc and there was no real state-of-the-art in terms of content analysis. Search indexing was one way, you can try to find posts that are relevant to a given subject and there’s synonym and taxonomy mapping all the rest that but sentiment analysis seemed like a far reach. So we built an early version of a sentiment analysis system, built a data training team to be able to go through and figure out what in this context is bad news or good news, and built kind of like an Outlook inbox. Rather than Coca-Cola giving the Twitter handle, and password to Coca-Cola’s public accounts to some kid who does marketing instead, they would put it with a central leader and then have all the different social media folks; their posts go into an approval and review queue then they would get sent out and they could be scheduled. There’s now a whole bunch of social media tools that have these capabilities baked in Hootsuite and copied a bunch of those building blocks, sentiment analysis has gotten radically better. In the Google Sentiment detection system, the one that’s part of the text APIs is more accurate than my sentiment tools. And we beat Twitter so hard for content through the front door that they built the fire hose to keep us off.
Rohan: Wowwww, you guys crashed Twitter.
Miles: We crashed Twitter a bunch of times when they were getting started so those kinds of access to the data of social media have gotten markedly easier. We were hand-building crawlers that ground all this thing out of the web.
We’re for a while between 2006. And when we really got the whole system up and going and about 2010, we had a complete copy of social media. All of it. We had all of it.
Rohan: Wowww, I’m flabbergasted. Four years.
Miles: Yeah. On hard drives that I stuck in racks, just like there aren’t any posts that we were missing if we missed them, then people yelled at us and we didn’t get renewals on our subscriptions.
Rohan: That’s crazy on-premises.
Miles: Totally on-prem, you want to learn why Cloud is amazing? Try to build a SaaS application without the cloud which I did, it sucks.
Rohan: It would be a nightmare to manage all those hard drives and data. I seriously cannot presume building anything without a cloud. It’s beyond my thinking capacity.
Rohan: Moving forward to AI which is profoundly booming in the industry. Every person out there on the street is talking about it but with rapidly increasing AI adoption and usage the danger of cybercrime also comes into the equation. If uncontrolled an AI model can be a evil, ruckus and can be a nest of intrigue for cybercriminals.
Miles: I’ll encourage folks to take a look at an application called Worm GPT, which is an auto-populating GPT-based social exploit system built on top of the GPT API. Every time you make tools better in any industry, those tools can be used for harm and used for defense protection and acceleration of the creation of value, as opposed to the destruction of it. I had a good talk with Phil Venables, he’s the CISO for Alphabet and his viewpoint was that while those tools are certainly very powerful and as a result, if bad guys have access to them, they can do quite a lot. There is an asymmetric capability where the construction of machine learning models is absolutely dependent on far more infrastructure like radically more dollars to get started than kind of ever before and it seems sort of thing is like if servers cost a million bucks, you might get a lot less server-based attacks on stuff because kind of one of the prerequisites for crime is that you’re a little short on bucks, so you’re trying to be cost efficient in attack modalities. My expectation is that this accretes power to huge businesses and those huge businesses are the ones that can afford to build the foundational models and frankly are the places where most of those models get vended. So they can watch and see bad actors making requests to APIs to use foundational models. If there are only four or five companies, there’s a far narrower choke point to reject bad actors from accessing these kinds of systems.
Meta’s announcement of Llama2 today shows that we’re going to continue to work to make machine learning models more and more price efficient and will make them able to run in arbitrary environments so maybe that barrier is coming down a bit but it’s certainly higher than what it costs to send a piece of spam, so you’ve got to be an attacker with pretty incredible resources to say the most efficient way for me to carry out an attack is to use a multi-million dollar large language model infrastructure rather than just spamming somebody with cheap labor from the middle of nowhere. There’s kind of a balancing act there.
Rohan: [laughs] It’s like in order to get money, you have to spend some money.
Rohan: There is a blog on Medium by Harvard Kennedy School which I read before, it was about how AI can be used to take over a democracy, how voters can be induced, and how election results can be manipulated using an AI. They mentioned an AI machine named Clogger — a political campaign in a black box; with one objective — to change people’s voting behavior. How would you evaluate or form an opinion on this?
Miles: Tongue-in-cheek example: I think the construction of the Gutenberg printing press which allows the acceleration of communication by central parties is a threat to democracy, where you have the ability to amplify your speech and make it persistent in a way that’s beyond the current legal structures of the time. I mean, every generation has had new tools to help new participants, push the messages and the ideas that they’re interested in. I think that we have done a bad job of working out government regulations on hate speech in a way that allows participants to do a more aggressive job right. Now, I think the companies are hesitant in their use of a sort of structured censorship of extreme hate speech because of the risks of exposure under Section 230 and other kinds of protections that they’re afforded for all the other things that happened on the platform. I think there are some steps that the government needs to take so that companies can actually take the actions which I think the vast majority of folks would find valuable. I think that we have to think about rulings like Citizen United, where you can use money as a direct facsimile for political speech, where I don’t think the two are equivalent. I think there’s a certain set of limitations and I don’t know what those thresholds or standards should look like, but saying that technology will deform political discourse is the same as saying money will deform the political discourse. They are literally interchangeable so if we don’t fix that money can be used then it doesn’t really matter what the technologies are or not.
I also think there’s an incredible opportunity in AI to help democracy by reducing the barrier of entry to education for every participant alive. If you think YouTube is a big deal for helping people learn, trust me, that depends on the hard work of the YouTube team to work content moderation and suppress hate speech and other kinds of destructive and corrosive concepts. I think that there’s the same kind of work needs to be done in what I expect to be effective like personalized tutors for every learner. That’s what Bard, GPT, and AI tools end up behaving as if I can have every student be able to interact with an utterly patient, attentive knows everything you’ve been taught, has tested everything that you have learned, and can walk you through what are the analyzes and thinking of our best minds and help all of us develop more quickly that is in enormous boon democracy flourishes in an educated populace. So my hope is that the help that generative AI and other AI tools can provide around education will be a far bigger boon to democracy than more ridiculous hate speech as spam in our public channels and communication.
Rohan: Yes, even at Google I/O 23 Sundar Pichai mentioned that AI is one of the most profound inventions that we have in this century. I got to know recently there are several LAWGPTs out there that let people understand their country’s law and legal rights, all you need to type the sentence, and it will give the answer. It has been built on top of ChatGPT. Then I also read about Welsh Congressmen who used ChatGPT to form his speech at parliament in June.
Miles: I think there has been for a really long time a huge tactical advantage, associated with being able to articulate yourself clearly and communication skills are persistently valuable. Here’s your statistics gag for the day — “50% of people are below average intelligence”. If all of a sudden, I can sound twice as smart and that just happens automatically. What a huge leap forward in my ability to communicate. If you’re below average, it’s awesome because it can pull you back up to the mean, it takes the average of all the writing that’s ever been on anywhere and tries to produce that as the responses to our prompts. For the very best communicators in the world. It can be a helpful collaborator and assistant.
Rohan: Recently Apple launched Apple Vision Pro, Google also had a similar thing which we know as Google Glass but they discontinued it. Was it because Glass was ahead of the time or didn’t meet the expectations that Google had or there were some other reasons?
Miles: The Glass program like many of Google’s experiments is held against the same standard as every department in the business, which is the Google Ads business, which commands exactly the maximum amount of that market that the Federal Trade Commission will allow at some 80% gross margins. If Google could make Ads cheaper would completely eliminate all digital competitors overnight and then be sued for antitrust in my view. They just can’t get bigger and they have ground the costs out of that system by way of radical innovation at the infrastructure and software layers to the point where the thing produces $80 billion of profit annually.
Rohan: What $80 billion, Holy Moly!!! I don’t know how to digest it.
Miles: If you’re a new startup, you’re at area 120 before they turn it off or you’re the engineers on the reader or you’re all fired up about Wave or Google Glass. Your business needs to either increase the market that the Ads business can participate in at 80% gross margins. That’s a cool idea. So maps is a whole huge system and it gets a bunch of incremental traffic, and they can place Ads on there, and yay you’re worth the price of entry. If you don’t, then you have to have a bigger or at least an equivalent total addressable market to Ads and be able to operate it with roughly the same kinds of profit margins otherwise you’re a terrible company for Alphabet to own. Glass is a super path-breaking technology and there are very valuable use cases. Can you produce glass devices and subscriptions to the hardware at 80% gross margins and is the total market for that larger than a hundred billion dollars a year? Since the answer is no, then it’s an experiment and you got to figure out what makes sense. So for example, Google Cloud has a larger total addressable market than Ads, the technical infrastructure market is between three and four trillion dollars, not three or four hundred billion dollars like the Ads market. So okay, there’s more money to be got at you could get bigger and not get squashed and over time software components that are a part of that can easily operate at 80% gross margins. Now, the total profitability of that business, we literally just crossed over into profitability, but it is on an increasing trajectory of profitability over time. So the expectation is, you could earn more profit total in cloud, than in ads because you can get to a bigger total market. You don’t have to sell just a hundred billion we’re already at 30 billion, Google could sell 400 billion or 500 billion, and if it’s slightly lower margins buying they still are profitable and they could produce more profit than Ads produces, that is an absolutely critical mission for the business. I like, every business to grow and be profitable, it’s the most obvious thing in the history of capitalism. You can take any of the different products that they interact with and use that guide post, whether is it like Ads or not and have a pretty clear view as to you as to likely longevity of that experiment.
I have every expectation that it’s a lot of the design ideas and the thinking that went into it and a whole bunch of the technologies that were invented for glass will certainly live on all the pixel phones that use the lens and the camera technology that was invented for glass. There’s a whole bunch of innovation that happened there and I think augmented reality as a whole discipline is certainly a place where all of us want to have a little birdie over our shoulder helping us figure out what we’re supposed to handle next or the context we need to answer the question that’s in front of us today or just support as we move through organizing what is an increasingly complicated world but I can’t look goofy can’t make it so you can’t make friends at a bar like the social realities of these kinds of tools are super-super important. That’s why, how much innovation Apple put into the screen of the Apple vision, not the monitors shooting at your eyeballs that’s for you. The screen makes it so you can see somebody’s eyes so it is even remotely socially acceptable to have this giant thing on your head. Innovation and thoughtfulness about the human condition are just as important as the technical breakthroughs to make the utility for the user.
Rohan: You are in the industry for the last 2 decades. Do you miss the old days of engineering — less Internet, non-AI world, and less algorithmically manicured websites?
Miles: No hell, no, everything is so much faster, and you can get to the funds so much more quickly, like some of the demonstrations that we’ve been building on Generative AI, I will be in front of our, executive customers or working with collaborators at Google and literally spit out an idea. What if it would do this and I’ll say a second to write that down and share it with our internal folks and it’s not next year, it’s not after multiple funding rounds and a whole huge, it’s like tomorrow there’s a demo that does that and my hold on a minute, I’m an imaginative guy and I can come up with all sorts of silly things that I want for these things to do if we can start working at the speed of thought in an increasing number of disciplines. So no I have no longing for the old days.
Rohan: Your thoughts on the community?
Miles: I think every human is on a journey. We’re all growing up, we’re all building the relationships that we are the sum of right, there is a little bit of me and then there’s also all my buddies and all the people that I learned from and learned from me, and help me get better at teaching and the sort of whole interaction of broader participants and if you do the multiplication on that every person and all of their relationships to every other person is what creates the opportunities, excitement, and fun in what we all do for a living. If I just received instructions from a single system and I did the things that were in the instructions and then when I was done after eight hours, I didn’t talk to anybody and I didn’t talk to anybody while I was working either, that’s a pretty horrible world.
Since we all should be able to empathize not only with the people that are younger than us and earlier in their careers but with the challenges that people that are older than us and kind of further along that they experience or the process of learning that all of us have to go through, we should be able to find a point of commonality with everybody else and if you don’t see that as an opportunity to be a service to help others to give and to pay it forward or however, you want to think about it then I don’t know, something’s broken and I don’t want to hang out with you. The community that I see in technology one is hugely varied, exists at different levels of sort of activity and participation, and is easily co-opted by corporate interests and profit motives and folks that are off trying to leverage something like that for a specific outcome or benefit. I think the highest benefit of the community is the improvement in the capabilities of the people involved. If that’s not the first order bit for everybody’s involvement, I think you’re sort of in the wrong spot.
I also think that figuring out how to help in frying those cycles of opportunistic exposure to other people’s smartness and making sure that you’re connected to what is actually working for others, what resources have been made it most efficient for people to learn new things, what are the values that sit at the far side as some of these opportunities, kind of know what to prioritize and how can you help others identify the same? those are really high-value activities. It’s evident that none of us are sufficient to solve the scale of the problems that sit in front of us, humans get radically more powerful when we work in teams and so this ad-hoc social global team of the interconnected community of participants is capable of way more than the individual people registered with a central database and something that’s more structured. My hope is that given how big the problems we have are; global poverty, climate change, the stuff we have to sort out around, sort of equal rights and equal access to opportunity, those challenges are gonna take billions of people to solve so we should work together and do that.
Rohan: That’s exactly the goal of community, people come together and offer solutions to a problem to the world, the more people you have the more efficient you are able to offer. In my perspective, every human out there is a part of the community anyhow directly or indirectly, whether you are helping out someone with a tiny problem over StackOverFlow or delivering a session on a global stage; solving a problem, or enhancing the existing solution.
Rohan: Thanks Miles for doing this interview. Thank you so much for your time today. I really enjoyed it and am more educated now. Have a good day!
Miles: Sure, thanks Rohan, have a good day you too, bye.